Malaysia - Assessing the effectiveness and impact of fiscal incentives

Report No. 35934-MY Malaysia Assessing the Effectiveness and Impact of Fiscal Incentives in Malaysia April 19, 2006 Poverty Reduction, Economic Management, and Financial Sector Unit East Asia and Pacific Region FOR OFFICIAL USE ONLY Document of the World Bank This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.   ACKNOWLEDGEMENTS This report was prepared by Albert G. Zeufack (Task Team Leader), Magdi Amin, Nick Bloom, Duanjie Chen, Dinesh Cheryan, Kenneth J. McKenzie, Anwar Shah, and Charles C. Udomsaph. The peer reviewers were Jack Mintz, Jacques Morisset, and Neil Roger. Comments were received from Kazi Matin, Eric Sidgwick and Olaf Unteroberdoester (IMF). Extensive comments were received from the Malaysian Industrial Development Authority, and the Ministry of Science, Technology, and Innovation. RachelWeaving provided editorial assistance. The production of the report was supported by Hedwig Abbey and Anita Kartadjoemena. The preparation of this report benefited greatly from the assistance and cooperation provided by the Government of Malaysia. In particular, the authors would like to thank Dato' Halipah Esa, Deputy Secretary General of the Ministry of Finance, Dato' Aziyah bt Bahaudin, and Mrs.Sarojini Deviinthe Tax DivisionoftheMinistryofFinance, and Staffinthe Department of Statistics for their generous help in providing data and explanations to the missions. Also, the authors would like to thank the Malaysian Institute of Taxation, the Ministry of International Trade and Industry,the Malaysian Industrial Development Authority, the Inland Revenue Board, the Federation of Malaysian Manufacturers, the Malaysian International Chamber of Commerce and Industry, the Malaysian Institute of Accountants, the Malaysian Institute of Certified Public Accountants, the American Chamber . of Commerce, PriceWaterhouseCoopers and International University College of Technology Twintech for organizing meetings and interviews with the team. This is the final draft report of the first phase of a multi-year work program on fiscal incentives program, an analytical and advisory activity between the Ministry of Finance of Malaysia and the World Bank. Draft chapters of this report were prepared between April and July 2004. World Bank missions visited Malaysia from April 26 to April 30 and May 24 to May 28 to collect information and conduct interviews with relevant government and private sector institutions. Preliminary results were discussed on July 16, 2004 during a World Bank mission and via video-conference. Earlier Drafts were shared with the Tax Division of the Ministry of Finance of Malaysia in September 2004 and November 2005. In June 2005, findings of this study were presented to the Minister of Finance 11, H.E. Tan Sri Nor Mohamed Yakcop, who authorized the roll-out of phase 11. Phase I1work, an input to the budget 2007, will include World Bank assistance in enhancing in-house capacity for tax analysis in the Ministry of Finance. Emphasis will be placed on building and maintaining a Marginal Effective Tax Rate model and a Micro Simulation model to inform tax policy, and assessing micro-level implications of tax measures through econometric analysis. This report is a follow-up to the "Firm Competitiveness, Investment Climate, and Growth" study conducted incollaboration with the Economic PlanningUnit.The FirmCompetitiveness report was deliveredto the Government inApril 2004 and published inJune 2005.   CONTENTS EXECUTIVE SUMMARY ............................................................................................... i 1 . ASSESSINGFISCALINCENTIVES:INTERNATIONALEVIDENCE ........1 B EvidenceFromDeveloping AndTransition Economies.,......................................... A. Introduction............................................................................................................... . 1 2 2 . INVESTMENTIMPACT OFTAX AND TAX INCENTIVESIN MANUFACTURING ........................................................................................................ 5 A. The Current Tax Incentive Program: An Overview .................................................. Tax Incentives Grantedunder PIA ........................................................................... 5 Incentives Grantedunder the IncomeTax Act (ITA).................... .......................... 6 10 Problems in the ExistingInvestment Incentive Program ........................................ 10 B The Impactof Taxes andIncentives on the Cost of Capital ................................... . 11 11 11 C Tax Impact on Foreign Investment: A Cross-country Comparison........................ Effective Tax Rateson Capital Investment by Domestic Firms............................. MarginalEffective Tax Rate Analysis: the Concept .............................................. . 16 Incentive Programs inCompetitor Countries ......................................................... D Conclusions............................................................................................................. Malaysia's Tax Competitiveness inthe Region ..................................................... 16 19 . 3. FISCALINCENTIVESFORRESEARCHAND DEVELOPMENT .............2324 A. Conceptual Framework........................................................................................... 24 Why Should R&D be subsidized?.......................................................................... 24 The Nature of Spillovers ......................................................................................... 26 B Overview Of Malaysian R&D Tax Incentives........................................................ . 27 27 Oversight. Monitoring. andEvaluation .................................................................. Structural Assessment............................................................................................. 33 36 Interaction betweenDepartments ........................................................................... Costing andManagement of Malaysia's Current Policies...................................... 37 C. Sample Impact Evaluation of an R&D Tax Incentive............................................. 38 Costing Based on Census Data............................................................................... 39 41 41 METRCalculations ................................................................................................ Comparison of the Costs from Different Data Sources.......................................... Costs Basedon Information from MIDA/IRB ....................................................... 42 D The Operational Context: Non-Tax andRegional Systems.................................... . 48 Non-Tax Incentives.......................................... 48 Benchmarking Malaysia's FiscalIncentives .......................................................... ...................................................... E. Conclusions AndRecommendations....................................................................... 51 53 4. IMPACT OFFISCALINCENTIVESONINVESTMENTIN MANUFACTURING ...................................................................................................... 55 ............................ B. LinearRegressionEstimate of the Impactof Incentives on Investment Rates.......62 A. Incentives Usedby ManufacturingEstablishmentsinMalaysia 55 C. CausalInference Through Propensity Score Matching........................................... 66 The Evaluation Problem ......................................................................................... 66  The Matching Method ............................................................................................ 67 MatchingEstimates of the Impact of Reinvestment Allowance/Accelerated Capital Allowance ............................................................................................................... 68 D. Conclusions............................................................................................................. 69 5 . THE COSTSOFFISCALINCENTIVESFORINVESTMENT .................... 70 A. Accounting for the Cost of ForgoneTax Revenues................................................ 70 Opportunity Costs of Tax Incentives...................................................................... Sensitivity of Investment to Taxation..................................................................... 72 74 B. Optimal Tax Treatment of ResearchandDevelopment.......................................... 76 C Conclusions . ............................................................................................................. 79 6. FINDINGSAND RECOMMENDATIONS ....................................................... 80 A. Findings................................................................................................................... 80 80 Incentives for ResearchandDevelopment ............................................................. Current Incentives................................................................................................... 81 Use of Incentives and Impact on Plant-level Investment Rates.............................. 83 84 B. Next Steps ............................................................................................................... Cost of Incentives ................................................................................................... 84 References ........................................................................................................................ 86 ANNEXES ....................................................................................................................... 89 FIGURES Figure 3.2: Conceptual Framework showing Lower Cost of R&D.................................. Figure 3.1: Conceptual Framework Showing Social Returnvs.Private Return ..............25 26 Figure3.3: Malaysia's Tax Incentives for R&D: Tip of the Iceberg?.............................. 50 Figure 4.2: Types of Incentives ........................................................................................ Figure 4.1: Use of Incentives ............................................................................................ 56 57 Figure 4.3: Firms' Awareness of R&D Incentives ........................................................... 57 Figure 4.4: Value of R&D Incentives ............................................................................... Figure 4.5: R&D IncentiveUse over Time....................................................................... 58 59 Figure 4.6: Awareness of Incentives for Small- and Medium-Scale Firms..................... Figure 4.8: Use of Incentives by Small- and Medium-scale Firms over Time.................61 Figure 4.7: Value of Incentives for Small- andMedium-scale Firms.............................. 60 61 62 Figure4.10: Distributionof the Propensity Scores........................................................... Figure4.9: Investment Rate andIncentive Use................................................................ 68 Figure5.1: Optimal Tax Treatment of R&D when Marginal Cost of Public Funds=1.429 ................................................................................................................................... 78 Figure 5.2: Optimal Tax Treatment of R&D when Marginal Cost of Public Funds=1.819 ................................................................................................................................... 78  TABLES Table 1.1:Additional InvestmentperUnit of Forgone Revenuesby the Treasury............2 Table 2.1:Business Tax Provisions Applicable to Foreign Invested Manufacturing Firms Table 2.2: Investment Incentives for Manufacturing, Agriculture, andTourism...............5 ..................................................................................................................................... 9 Table 2.3: Marginal Effective Tax Rates(%) for Manufacturing, by Type of Assets .....13 Table 2.4: MarginalEffectiveTax Rates (%) for Manufacturing, Agriculture, and Tourism..................................................................................................................... 14 Table 2.5: Effective Tax Rate (%) on Foreign Direct Investment inManufacturing: TableComparison of Malaysiawith China, Indonesia, Singapore, andThailand ..............21 3.1: Summary of Company Types andTax Allowances They Receive .................27 Table 3.2: R&D Expenditure and Forgone Revenue by Industry..................................... Table 3.3: MIDA's Investment Allowance andPioneer Schemes ................................... 40 41 Table 3.4: Breakdownof Claims on DD.......................................................................... Table 3.5: Variation inR&D Expenditure Due to DifferingRates of Tax Incentive......41 44 Table 3.6: B y Industry (Short Run) .................................................................................. 45 Table 3.7: B y Sales (Short Run)....................................................................................... Table 3.8: B y Region (Short Run).................................................................................... 46 46 Table 3.9: B y Industry (Long Run)................................................................................... 47 Table 3.10: B y Sales (Long Run) ..................................................................................... 47 Table 3.11: B y Region (Long Run).................................................................................. Table 3.12: R&D tax incentives in selectedcountries...................................................... 48 51 Table 3.13: Science and Technology: Summary Indicators for SelectedEast Asian Table 3.14: TIMMS-R Scores........................................................................................... Economies................................................................................................................. 52 Table 4.1: Investment Equations: LargeEstablishments.................................................. 53 Table 4.2: Investment Equations: Small- andMedium-scale Establishments..................65 66 Table 4.3: Propensity ScoreEstimates of the Effect of Pioneer Status and/or Investment Tax Allowance Use................................................................................................... 69 Table 5.1: Marginal Cost of Public Fundsfor Taxes on the Returnto Capital inMalaysian Manufacturing ........................................................................................................... 74 Table 5.2: PercentageIncreaseinTax Sensitivity Relative to the General BaseCase Requiredto Justify the Lower METR....................................................................... 76   EXECUTIVESUMMARY 1. Tax systems influence the volume and location of firms' investments. Tax policy instruments are widely used by developing and industrial countries alike to stimulate investments in industry and technology. Virtually all countries use one or a combination of incentives including tax holidays, accelerated capital consumption allowances, and investment allowances through the corporate tax code. Many countries that have been particularly successful in attracting FDI and achieving high levels of domestic private investment have used tax incentives extensively. These countries include Ireland, Hong Kong (China), Malaysia, and Singapore. However, international evidence strongly suggests that only certain types of tax incentives, especially those that apply to new capital investment or to incremental investment over and above the investment usually undertaken, and are payable when the investment i s undertaken, are likely to be more effective and to significantly affect investment decisions of firms. The debate around tax-incentives i s not if they should exist or not, but which ones are the most cost effective. 2. What are the most effective, or cost-effective, trade-offs within a tax system? There i s growing body of evidence indicatingthat an efficient tax incentive system is one that i s simple, transparent, and automatic, and thus relatively easy to enforce. Effective incentive systems tend to rely less on tax holidays such as Malaysia's Pioneer Status and more on instruments such as lower corporate income tax (CIT) rates, investment allowances, andor investment tax credits, which have proven more successful in promoting investment. This basic approach meets the many and sometimes contradictory goals of a business tax system. It has worked well for a number of successful countries, including Chile, Hong Kong (China), Hungary, and Mauritius, and more recently has been adopted by Eastern European countries such as Estonia, Latvia, Ireland, Romania, andPoland. 3. The objective of this report is to review the structure of Malaysia tax incentives in light of international experience with a view to provide input to Malaysia's budget formulation process. Fiscal incentives play a central role in Malaysia's industrial policy, the subject of lively controversies. It i s therefore necessary to derive our recommendations from strong analytical andempirical evidence. This report uses marginal effective tax rate (METR) calculations, econometric and matching techniques to assess the system's effectiveness and impact for the manufacturing sector. Also, recognizing that incentives come at a cost, the report uses the marginal cost of funds (MCF) approach to provide a preliminary assessment of the cost of fiscal incentives in terms of tax revenue. Extensive Annexes provide technical details on these methodologies. i  Diagnosis 4. Malaysia's tax and incentive structure is overly generous and complex, and may leave roomfor tax abuse. By comparison with those of neighboring and competitor countries, including China, Indonesia, Singapore and Thailand, Malaysia's tax and incentive system for manufacturing i s overly generous and complicated. Several factors contribute to this situation. First, there are too many incentives, yet some with unclear legislation, creating significant uncertainty as to eligible expenses for potential investors. Second, there is a slow roll out. Significant lags in enacting enabling legislation and regulations undermine the excitement created by budget announcements. Third, major incentives such as pioneer status, investment tax allowances require prior approval through applications. While such applications are almost always granted, the process works as a psychological barrier for some potential, especially domestic investors, who simply do not claim them. Fourth, there are no real incentives for investors with home country tax regime of worldwide income with foreign tax credits. In addition, since the Malaysian tax system makes available loss carry-forward and accelerated depreciation allowances to regular business taxpayers, tax holidays become less attractive to serious capital investors and leave room for tax abuse. Further, where tax holidays are combined with the extremely generous and deferrable write-offs, Malaysia's system not only increases the after-tax profits of profitable companies but in many cases can transform an investment that i s not profitable before taxes into one that i s profitable after taxes. For most companies that get these incentives, the corporate income tax (CIT) i s effectively at or below zero. 5. Malaysian tax holidays are ineffective. We use marginal effective tax rate (METR) analysis to quantify the impact of the current tax and incentive provisions on capital investment in Malaysia. Simply put, the METR i s the net amount of taxes paid as a percentage of income earned from an investment project on an annualized basis. Granting a tax incentive in the form of accelerated depreciation deductions, for example, lowers the METR, and as a consequence, the overall METR under the general system i s higher than it otherwise would be. Therefore, the cost of offering this incentive exceeds the revenue generated from the higher tax rate it induces. Malaysia's Pioneer Status incentives raise the lifetime tax burden for firms that invest inlong-livedassets and are in a loss position in their formative years. Conversely, these incentives lower the burden for firms that invest in short-lived assets and are profitable from the outset. 6. Malaysia's overly generous R&D incentive schemes have not substantially raised R&D spending in the private sector. Malaysia has at least 22 grant, loan, or equity-based schemes operating at the firm, institutional, or individual level for this purpose. But these schemes have had little impact on R&D spending inthe private sector, There i s a possibility that government funding has crowded out private demand for the limited number of available researchers. A rough estimate of the available funding exceeds RM300 million per year, which i s a fairly robust amount of fundingconsidering that there are about 6,000 - 8,000 researchers in Malaysia. In addition, the structural nature of the schemes provides scope for widespread tax abuse, which may be difficult for MIDA and IRB to control without extensive costly and invasive monitoring. Firms' aversion to heavy-handed monitoring may explain the very low uptake of these schemes. 11 ..  7. In Malaysia, as in many other countries, R&D is quite sensitive to tax subsidies, but this does not mean that these subsidiesare economicallyefficient. The current system of tax incentives for R&D i s effective in the sense that a ringgit of incentives generates roughly a ringgit of spending on R&D, on average, but this does not mean that it i s economically efficient. Some of the increased expenditure reflects the rising costs of increasingly scarce R&D personnel, rather than more R&D activity. Moreover, the opportunity cost of the tax subsidy in terms of forgone revenue i s often ignored but can be very important; our investigations suggest that Malaysia's existing subsidy may come at too high a cost. 8. Malaysian tax incentives may be very costly to the economy. In a preliminary cost assessment, we use the marginal cost of public funds approach to evaluate the tax revenue that could have otherwise been raised if the incentives were not inplace. We find that raising one more ringgit in tax revenue, by increasing the tax on capital in the general regime by a small amount, results in a reduction inincome inthe private sector of 1.4 ringgit. Of this, one ringgit reflects the higher taxes paid and 0.4 ringgit i s the marginalefficiency cost associatedwith the higher tax rate. 9. Malaysian fiscal incentives schemes benefit mainly large firms. Descriptive statistics from the Productivity and Investment Climate Survey (P1CS)-a survey of 1,200 firms conducted by the Economic Planning Unit in collaboration with the World Bank in early 2003-show a striking difference in the use of incentives between large firms and SMEs: 60 percent of large firms have taken advantage of incentives for their most recent investment, whereas only 38 percent of SMEs have done so. Possibly this reflects lack of awareness of the schemes andor the costs of complying with their requirements. 10. The reinvestment allowance has a positive impacton large firms' investment. Using the PICS data, we conducted impact analysis for selected fiscal incentives. Regression-basedand propensity score matching estimates suggest that the reinvestment allowance/accelerated capital allowance positively affects investment, but only among large firms. Large-scale establishments that have benefited from the reinvestment allowance/accelerated capital allowance have investment rates 9.4 percent higher than those firms that have not. Among small- and medium-scale establishments, the allowance has no significant effect. 11. Pressuring the tax system to compensate for deficiencies in the investment climate may have a large opportunity cost for Malaysia. Foreign and domestic firms' decisions to invest are complex and fiscal incentives may be just a small piece in the . puzzle. Reducing the regulatory burden and improving the overall investment climate might have more effect than a generous system of tax incentives. Regression-based estimates suggest that regulatory constraints have a strong negative impact on the investment rates of Malaysian firms. Ownership regulations for large-scale establishments and import regulations for SMEs are the greatest deterrents. Among large establishments, those that indicate that ownership regulation i s the biggest obstacle to doing business in Malaysia have 20.3 percent lower investment rates than others. Similarly, among SMEs, those that indicate import regulation i s the biggest obstacle to iii  doing business inMalaysiahave 13.1 percent lower investment rates than the others. This result is statistically more robust than the finding on the impact of the reinvestment allowance/accelerated capital allowance-tax incentives that have been identified as cost effective. Whatneeds to be done? 12. Scaling back tax incentives in favor of a more general decline in corporate income taxes will enhance efficiency. The general strategy we recommend i s to move towards a regime that i s simpler and less discretionary, with a lower corporate income tax (CIT) rate. This i s consistent with international evidence. Most studies have identified a low tax rate, accompanied by loss carry-forwards for tax purposes and accelerated depreciation, as key components of an effective tax system and one that i s highly attractive to foreign investors. In this regard, Hong Kong (China) and Ireland provide persuasive examples. Both countries have adopted a low flat company tax rate (17.5 percent in Hong Kong and 12.5 percent in Ireland) for all business sectors with no discretionary tax incentives. InMalaysia, the general CIT rate i s 28 percent, with a lower CIT rate for small businesses (20 percent). The announcement in the 2005 Malaysian budget that the government i s planning to introduce a value-added type consumption tax on January 1, 2007, unifying the sales and service components of the current consumption tax, could provide an opportunity to lower the CIT. 13. Sequencing the reform of the tax incentives system. To move in the direction of lower tax incentives and a general decline in corporate income taxes, the priorities are to: 0 Cost all corporate tax andnon-tax schemes including those for R&D. 0 Maintain incentives that are directly linked to the amount of investment undertaken, such as investment tax allowances, but to simplify and rigorously enforce their application. 0 Progressively eliminate tax holidays such as Pioneer Status, which may be popular but appear to be ineffective, while reducing the overall regulatory burden. 0 Reduce the corporate income tax, at a later stage, buildingon the savings realized from eliminating non-effective incentives. 14. Improving the overall investment climate by reducing the regulatory burden. The reform of the Malaysiantax incentives system should be combined with a sustained effort to improve the investment climate. International evidence points to the fact that no tax incentive has ever attracted investors to areas where fundamental conditions for profitability are not satisfied. Such conditions include the availability of natural resources, skilled workers, adequate infrastructure, political and economic stability, and the size of the domestic market, as well as the tax regime. In an unfavorable overall investment climate, tax incentives are a less effective way to attract investment than are efforts to reduce the regulatory burden on firms. iv  The road ahead 15. Malaysia has not been assessing the cost of its incentives. Unlike other countries at its level of development, Malaysia does not systematically monitor the overall cost of incentives. As aconsequence,the following questions remain unanswered. 0 B y how much should the government reduce CIT? 0 What will be the cost of reducing CIT, interms of revenues? 0 What should be the balancebetween incentives and corporate taxation? 16. We recommend: 0 An appropriate costing of all corporate tax schemes including those for R&D as the immediate next step to this study. This would provide an essential check on the operation and uptake of schemes, but it i s also a vital component of the proper cost-benefit analysis that i s necessary to answer these questions. 0 More work on the revenueimplications of reform proposals. 0 As an important part of any comprehensive reformpackage, broader examination of how business incentives fit into the tax system as a whole. Other taxes, such as excise taxes, import duties, personal income tax (and its integration with the corporate tax), sales taxes, interact with business taxes to create an overall tax environment that can be an important consideration for foreign investors. V   1. ASSESSINGFISCALINCENTIVES: INTERNATIONAL EVIDENCE A. INTRODUCTION 1.1 A rationale for investment incentives put forth long ago is to encourage capital investment that i s inadequately providedby markets. Designers of various tax incentives around the world have interpreted this rationale as indicating that tax incentives can induce investors to follow directions favored by government, and as a result will produce wide benefits for the overall economy. Unfortunately, the directions favored by governments have often been arbitrarily defined, for example in terms of capital allocation by industry, type of investor, location, size of investment, or some combination of these criteria. Since governments have rarely foreseen economic structural change accurately, or succeededinpicking the "winners" chosen by the market, industry-specific incentives have often become an obstacleto fostering infant industries. Tax incentives for promoting regional development have generally failed: none has ever attracted investors to areas that lack fundamental conditions for profitability-such as availability of natural resources, skilled workers, adequate infrastructure, or even the population density required to support markets (Shankar and Shah 2003). Further, regardless of how carefully targeted tax incentives are, their immediate fiscal impact has always been to reduce tax revenue, whether directly or indirectly, because they narrow the tax base. When this revenue loss cannot be offset by the expected trickle-down or spillover effect of "induced" capital investment, a government must raise revenue elsewhere or cut its spending-potentially with harmfuleffects on the economy. 1.2 This report is structured as follows. This chapter presents international experience with different types of corporate tax incentives. Chapter 2 examines Malaysia's current tax regime for business and its attractiveness to foreign investors, relative to those of competitor countries in the region. Chapter 3 examines the potential fiscal incentives that can be used to stimulate research and development activity in Malaysia. Chapter 4 uses plant-level survey data to investigate how current incentives are actually affecting the investment rates in Malaysia'. Chapter 5 uses the results obtained in preceding chapters to analyzethe cost of Malaysia's fiscal incentives. Chapter 6 summarizes the findings and recommendsnext steps. 'Because METRs are computed on incremental investments and may not shed proper light on the actual impact to determine cost-effectiveness of alternative instruments, the report uses on micro-econometric analysis to complement results on effectiveness and impact. Also, given data limitations, a Benefit-Cost Analysis i s undertakenonly for R&D incentives.  B. EVIDENCE FROM DEVELOPING TRANSITION AND ECONOMIES 1.3 Despite the popularity of tax policy instruments, there i s very little evidence on the actual impact of these instruments on business activity and government revenues. For developing countries, some examples of cost effectiveness are reported in Shah (1995). Using available evidence, that study ranks alternative tax incentive instruments for their relative efficacy, measuredas the investment they generateper dollar of revenues forgone bythe treasury (Table 1.1). Table 1.1: Additional Investmentper Unitof Forgone Revenues by the Treasury (Benefit-cost ratiofor various tax instruments) Tax incentive Country Short Run Intermediate Long Run Run CorporateTax Mexico: Detergents 0.03 0.04 0.05 Rate Reductions Other chemicals 0.01 0.01 Pakistan:Apparel 0.00 0.00 Leather 0.00 0.00 0.71 Cumulative 0.21 1 0.28 Turkey: Electricalmachinery 0.06 0.10 ~ Transport equipment Accelerated Mexico: Detergents 0.32 0.38 0.40 Depreciation Other chemicals 0.19 0.24 0.25 Pakistan:Apparel 0.23 0.59 2.10 Leather 0.09 0.23 2.13 Turkey: Electricalmachinery 0.38 0.68 1.45 Transport equipment 0.94 1.44 2.25 InvestmentTax Mexico: Detergents 0.44 0.51 0.54 Credit Other chemicals 0.26 0.32 0.34 Pakistan:Apparel 0.28 0.71 2.50 Leather 0.11 0.28 2.54 0.95 Cumulative InvestmentTax Turkey :Electricalmachinery 0.40 0.72 1.54 Allowance Transport equipment 1.oo 1.54 2.40 Industrialchemicals 0.25 Petroleumderivatives 0.95 Canada 1.80 Pakistan:Chemicals 1.38 Allowance Pharmaceuticals 0.24 Turkey: Industrialchemicals 1 0.13 I Petroleumderivatives 0.26 IExpensingof Pakistan 1.49 investment I I Source: Shah, Anwar (1995). Fiscal Incentivesfor Investment and Innovation. OxfordUniversityPress. 1.4 The international evidence strongly suggests that tax incentives that apply only to new capital investment (and preferably to incremental investment over and above the investment usually undertaken anyway), and are payable when the investment i s undertaken, are likely to be more effective than others. Thus full expensing of investment (without interest deductibility) would be most effective. Refundable investment tax 2  credits and allowances for acceleratedcapital consumption are likely to be less effective, and tax holidays andtax rate reductions are even less so. 1.5 Investment tax credits and allowances and fast write-offs are cost-effective tax incentives, with benefit-cost ratios usually exceeding one: a Investment tax credits and allowances. Under an investment tax credit, companies in a specific industry, or more generally, are allowedto deduct from their tax liabilities a fraction of their spending on new additions to physical or research and development (R&D) capital stock, or employment. Tax credits can be granted for specific activities, and, by providing a direct subsidy, can be more effective than tax rate reductions. An investment allowance i s similar in effect to a tax credit except that it is a deduction from the income subject to corporate income tax. The effectiveness of tax credits and allowances depends upon whether they are refundable and therefore provide cash upfront to the firm, and, if they are not, on the extent to which they can be carried forward. Tax credits can be targeted to specific types of investments, both tangible and intangible, and they may be discretionary or automatic. Automatic tax credits are more desirable. For foreign subsidiaries, a relevant consideration i s whether or not the credits are offset by the system of foreign tax crediting. To the extent that they are, they may represent at least partly a transfer of tax revenues to foreign treasuries, as noted above. Overall, however, as shown in Table 1.1, tax credits and tax allowances for physical and R&D capital can be cost-effective tax instruments for stimulating investment. a Fast write-o@. Certain types of costs are eligible for a fast write-off for tax purposes. Most commonly, these are depreciation costs (capital consumption allowances) that can be acceleratedor even fully expensed. Such acceleration provides an incentive for investingin assets that are eligible for faster write-offs. Where equity markets do not function well, allowing a firm to fully expense its investments (without interest deductibility) under the tax system makes the government an equity partner and promotes investment. Such atax incentive i s also generally seen to be cost effective. 1.6 Preferential tax rates and tax holidays are costly tax incentives, with benefit-cost ratios usually less than one. Preferential tax rates. Certain types of firms may receive lower tax rates than others, on either a temporary or a permanent basis. The use of preferential tax rates is a blunt instrument for providing investment incentives, because the incentive does not vary by the amount of investment undertaken. In the absence of full loss-offsetting provisions and when marginal tax rates are already low, the incentive effect they offer i s minimal. Preferential tax rates are difficult to target in practice and they can provide a windfall gain on investments that would have beenundertaken even without these measures. Further, past experience suggests that incentives given to a selected few priority industries create pressure from other industries for similar treatment. Over time, these incentives proliferate, as happened in Brazil, Indonesia, and Mexico 3  inthe early 1980s. As a result, a tax systembecomescomplex, and its ability to raise revenues in an equitable and less distorting manner becomes impaired. This outcome induces tax avoidance andtax evasion. 0 Taxholiday representsan extreme case of preferential tax rates, whereby a firm is tax free for a specified period. Tax holidays may be awarded on a discretionary basis to firms in designatedindustriesor areas. Firms awardedtax holidays are typically those just starting up, and are referred to as "pioneer firms". Tax holiday is a widely used incentive in developing countries including Bangladesh, Brazil, Cote d'Ivoire, Lesotho, Malaysia, Morocco, Pakistan, and Thailand, and among these, Morocco and Lesotho have extended their tax holiday provisions to foreign investors as well. Just like those of a change in tax structure, the effects of a tax holiday policy on the user's cost of capital are of three possible types - one positive and two negative. A zero tax rate during the holiday period has a positive effect on investment. By contrast, reductions in the present value of depreciation allowances and the rising value of these allowances towards the end of the holiday period discourage current investment. The overall effect of a tax holiday on investment depends on the extent to which capital consumption allowances and tax losses can be carried forward to the post- holiday period. If there are no carry-forward provisions, some firms actually benefit by being taxable in their formative years, when they are in a loss position. If a full carry-forward i s allowed, effective subsidies in certain instances would be too generous. In general, a tax holiday i s generous to firms that use non-depreciable factors of production. In any event, tax holidays open tax arbitrage opportunities for investors to shift taxable income into activities in which they can take advantage of the tax holiday. Tax holidays may thus encourage fly-by-night or short-term operations at the expense of long-terminvestments. Mintz (in Shah 1995) shows that in Malaysia, with tax holidays, investors enjoyed large subsidies (exceeding 100 percent) through their negative effective tax rates duringthe holiday period. 1.7 As experience has accumulated, most countries have changed the design of their fiscal incentives. Tax holidays, once popular, are losing favor compared to more efficient incentives that are directly linkedto the amount of investment undertaken2As UNCTAD (2000) notes, "a low tax rate accompanied by loss carry-forwards for tax purposes and accelerated depreciation i s considered to be a major element in an effective tax system and one that i s highly attractive to foreign investors". Inthis regard, Hong Kong (China) and Ireland provide persuasive examples. Both these economies are small and poor in natural resources, and both have adopted a low flat company tax rate (17.5 percent in Hong Kong and 12.5 percent in Ireland) for all business sectors, with no discretionary tax incentives. Both economies have excelled in attracting foreign direct investment: measured by the ratio of foreign direct investment inflow to GDP, Hong Kong (China) and Ireland stand out as the second (20 percent) and fourth most successful country (14 percent) inthe world, basedon the statistics for 1997-2001, see Mintz (2004). - The disadvantages of tax holidays have been extensively documented. Recent references include UNCTAD (2000), and Mintz and Smart (2004). 4  2. INVESTMENTIMPACTOFTAXAND TAX INCENTIVESINMANUFACTURING 2.1 This chapter evaluates the impact of current tax incentives on capital investment inMalaysia's manufacturingindustry. It has three parts: Section A provides an overview of the current tax and tax incentive structure applicable to manufacturing in comparison with those applicable to agriculture and tourism. Section B calculates "effective" tax rates, using marginal effective tax rate (METR) analysis, to indicate the investment impact of tax and tax incentives on manufacturing, and Section C compares Malaysia's tax incentives for foreign investors with those of major competitor countries-China, Indonesia, Singapore, andThailand. A. THECURRENTTAXINCENTIVE PROGRAM: ANOVERVIEW 2.2 General business taxation in Malaysia consists of three major taxes (Table 2.1): the corporate income tax (CIT), transaction taxes on capital goods, and the property tax.3 2.3 The general CIT rate is 28 percent, and that for small and medium businesses (SMEs) i s 20 percent. The depreciation allowance for capital investment i s generous: in addition to the annual allowance, there i s an initial allowance of 20 percent for machinery and equipment and 10percent for buildings and structure. Business losses may be carried forward indefinitely. Malaysia China Indonesia Singapore Thailand 1. Capital Taxes Company incometax rate 20.0/28.0 15.0/24.0/30.0 10.0/15.0/30.0 22.0" 20.0/25.0/30.0 Revenue-based tax None None None None Yes Initialinvestmentallowance 10.0/20.0 None None 20.0 None Annual depreciationfate Buildinas 3.0 SL 5.0 SL 5.0 SL . 3.0 SL 5.0 SL Machinery - Manufacturing 14.0SL 15.0 SL 18.75SL 15-33.3 SL 20.0 SL Inventoryaccounting FIFO Optional FIFO/average Optionalho LIFO Optional I s s Carry-forward Indefinitely 5 years 5 years indefinitely 5 years ~~n dividends II None II 20.0 II 15.0 II None II 10.0II Annex 1SectionA provides a fuller overview of Malaysia's businesstaxation. 5  IProDertvtax I 0.45-0.55 onI 18.0on rental I 0.2 on costI 10.0on rentalI12.5 on rentalI Propertvtransfertax 3.0 3.0 5.0 5.0-10.0 1.o 2. indirectTaxes - Sales tax 10.0 17.0 VAT 18.0 VAT 5.0 VAT 7.0 VAT - import dutyon M&E Up to 30.0 9.6 Under 5.0 Generally0.0 Under 10.0 3. Tax incentives Taxholidays 5-10 years 5 years unclear 5-10 years 3-8 years Accelerafeddepreciation Yes No Yes Yes Yes Importduty exemption Yes Yes Yes NA Yes Notes:NA = not applicable, M&E= machinery and equipment, and S L = straight-line method. aTheCIT rate inSingaporewill be reducedto 20 percentfor the assessment year of 2005. According to the International Bureau of Fiscal Documentation, as an investment incentive, all the capital goods imported to UAE are exempted from the import duty; according to information published at the official website of Malaysia's Customs Department, import duty i s zero on all goods imported to UAE. Despitethe inconsistency, both sources indicate an effective zero import duty on machinery and equipment. 2.4 Malaysia has abolished import duties on a wide range of raw materials, components, and ma~hinery.~However, a sales tax i s applied to capital goods made domestically, at a general rate of 10per~ent.~Since this sales tax i s not a value-added tax, it is non-refundable to business taxpayers. As such, any sales tax paid by investors on their purchase of capital goods effectively increases the cost of capital investment. 2.5 Local authorities levy property tax, based on the cost of property, with rates ranging from 0.45 percent in remote areas to 0.55 percent in "prime" areas. The stamp duty on property transfer i s 3 percent on transaction prices above RM400,OOO. 2.6 Malaysia has a long history of offering a wide range of tax incentives for investments in manufacturing, agriculture, and tourism. The major incentives are provided under the Promotion of Investments Act (PIA) 1986 and its amendments, and others are providedunder the Income Tax Act (ITA) 1967 and its amendrnenk6 Tax Incentives Granted under PIA 2.7 These incentives are granted by the Malaysian Industrial Development Authority (MIDA), and consist of two alternatives: so-called Pioneer Status or the regular Investment Allowance. Eligibility for these alternative incentives i s based on certain priorities, including the firm's level of value-added, technology used, and industrial Therefore, it is reasonableto assume a zero import duty rate on capital goods for our METRcalculation later in this paper. Statistics provided by the Ministry of Finance indicate that about 30 percentof machinery and equipment usedby manufacturing industry i s made domestically. These are detailed inAnnex 3 Section C. 6  linkages. Eligible projects are termed to be "promoted activities" or "promoted products.'77 2.8 Under Pioneer Status, a qualified investment project may enjoy a limited period of tax holiday or tax reduction, applied to its statutory business income only. * Because the period i s limited, capital allowances must be deducted from adjusted income, and unabsorbedbusiness losses and capital allowances cannot be carried forward to the post- pioneer period. As an alternative to the Pioneer Status, however, the Investment Allowance granted to qualified capital expenditureg, if unabsorbed during the specified period, may be carried forward and deducted against adjusted income in subsequent years. 2.9 The originaE content of these alternatives was as follows. A company with Pioneer Status may enjoy a 70 percent exemption from the payment of income tax for a five-year period. Alternatively, a company may obtain an Investment Allowance of 60 percent'for its qualifying capital expenditure, up to 70 percent of its statutory income within five years from the date on which the first qualifying capital expenditure i s incurred. 2.10 The enhanced incentive under these two alternative programs is available for eligible projects located in "promoted areas."" With this enhanced incentive, a company may enjoy either Pioneer Status with a 100percent tax exemption on its statutory income or a 100 percent IA up to 100 percent of its statutory income. Both of these options last for a five-year period. However, there i s a time limit for applying this incentive, which i s December 31,2005. 'According to the International Bureau of Fiscal Documentation (IBFD), Malaysia's Minister of International Trade and Industry from time to time determines which activities or products are to be promoted. The factors for determining these promoted activities or products are the suitability of "commercial scale," the "prospects for further development", and "the national and strategic requirements of Malaysia." See IBFDMalaysia, p. 135. * According to MIMS, "statutory income i s derived after deducting revenue expenditure and capital allowances from the gross income," which can be seen as a proxy for taxable income at the margin in METR analysis. According to IBFD, "capital expenditure" qualifying for the investment allowance covers only that (for production purposes) incurred in manufacturing (factory, plant and machinery), agricultural sector (buildings, access roads including bridges, plant and machinery), hotel business (hotel buildings, plant, machinery and related facilities), and tourist projects (constructing and running the projects). See IBFD, Malaysia, pages 138-40. lo"Promoted areas" include the States of Sabah and Sarawak and the designated "Eastern Corridor" of Peninsular Malaysia, which covers the States of Kelantan, Terengganu and Pahang, and the district of Mersinginthe Stateof Johore. 7  2.11 The above alternative incentives have many variations, in terms of the percentage of tax exemption, the percentage of Investment Allowance, and the number of years for eligibility. The variations depend on industry- and subsector-specific features, and some involve even more generous terms than those just outlined. For example, more generous incentives are tailored to some branches of manufacturing, such as high technology companies, strategic projects that involve products or activities of national importance, SMEs, large firms that participate in an industrial linkage program, and firms that produce machinery and equipment. 2.12 Table 2.2 summarizes the resulting structure of investment incentives for manufacturing, agricultural, andtourism projects. l1 The tourism projects are eco-tourism and agro-tourism projects including hotel businesses (Le. construction of low-cost hotels up to a three-star category and expansiodmodernization of existing hotels), construction of holiday camps, recreationalprojects, and construction of convention centers with a capacity to accommodateat least 3,000 participants. 8  s z sz 5 . ^ 2 e0 0 a v) sz  Incentives Granted under the Income Tax Act (ITA) 2.13 The investment incentives under the ITA are granted by the Inland Revenue Board (IRB) and are targeted at reinvestment in selective sectors or projects. In manufacturing, a reinvestment allowance of 60 to 70 percent of statutory income is available for a firm that "has been in operation for at least 12 months and incurs qualifying capital expenditure to expand, modernize, or automate its existing business or diversify its existing businessinto any related products within the same industry" (MIMS 2004). This incentive i s given for a period of 15 consecutive years beginning in the year the first reinvestment is made, after which firms that reinvest in manufacturing of promoted products are eligible for an Accelerated Capital Allowance (ACA) that can be used within three years (Le. with an initial allowance of 40 percent and an annual allowance of 20 percent). In agriculture, a 15-year reinvestment allowance followed by an ACA, the same as that given to manufacturing industry, is available to any entity that has been engaged for at least 12 months in producing "essential" food, defined as rice, maize, vegetables, tubers, livestock, aquatic products, and any other activities approved by the Minister of Finance. Problems inthe ExistingInvestment Incentive Program 2.14 Malaysia's current investment incentive program i s complicated. It has been formulated on a highly selective basis interms of sectors, regions, and length of period to qualify for certain tax incentives. As Table 2.2 shows, an investment project within manufacturing may be entitled to one of nine different tax incentives depending on the nature of product, the location, the source of funding (i.e. initial investment vs. reinvestment from after-tax income), the nationality of ownership, the size of firm, the stage of business life, and the preference of the investor. 2.15 The incentive program is also inefficient. It has been tailored to investors' preferences, but it appears to have sacrificed the needed balance between encouraging capital investment and generating tax revenue. Further, since the system makes loss- carry-forward and accelerated depreciation allowances available to regular business taxpayers, tax holidays become less attractive to serious capital investors and leave room for tax abuse. The eligibility criteria for various incentives are tedious for firms to satisfy. This militates against market efficiency andcomplicates tax administration. 2.16 For example, an initial allowance of 60 percent (or 100 percent) up to 70 percent (or 100 percent) of income for five years, by artificially enforcing or skewing the investment decision in the officially desired direction, appears to be quite confusing. During the start-up period of an investment project, a typical firm does not necessarily expect to invest continuously each year unless it sees highly promising profitability andor has the financial capacity to make such investments. For firms that cannot afford to make intensive capital investments within the period favored by the government, the limitation on the period for granting initial allowances either induces distorted capital investment or provides very little tax benefit. A similar argument can be made against granting the initial allowance to reinvestment for a period of 15 consecutive years. Different manufacturing industries do not share the same investment cycle. Accordingly, there is no justification for the government to manipulate the business sector's capital 10  investment by scheduling the availability of tax incentives. Restricting the initial allowance to a fixed number of years is not in accord with the goal of market efficiency. B. THEIMPACTOFTAXES INCENTIVESONTHE COSTOFCAPITAL AND 2.17 To quantify the impact of the current tax and incentive provisions on capital investment in Malaysia, we use marginal effective tax rate (METR) analysis. We first explain the concept of marginal effective tax rate analysis and then present our estimates of the METR on capital investments affected by the different types of investment incentives inMalaysia. MarginalEffective Tax RateAnalysis: the Concept 2.18 The way to judge how provisions in the tax system affect investment i s to determine how much tax a firm pays on a marginal or new capital investment. The marginal effective tax rate i s a summary measure that quantifies the overall cumulative tax burden that i s placed on this new capital investment. Thus the METRincorporates the effects not only of statutory tax rates and related tax treatments (e.g., tax depreciation, tax credits, tax deductibility, tax holidays), but also of various economic factors that interact with these tax treatments (e.g., financial costs, the inflation rate, the structure of investment). As the result of this interaction between statutory tax provisions and actual economic and industrial conditions, effective tax rates can vary by industry, even under the same tax regime.12 2.19 The METR is expressedas the net amount of taxes paid as a percentageof income earned from an investment project on an annualized basis. For profit-maximizing firms, the rate of return on capital, net of economic depreciation, must be equal to the financing cost of capital, adjusted for taxes. The size of this adjustment for taxes on a new investment i s the marginal effective tax rate on capital. For example, if the gross rate of return to a new capital investment i s 20 percent and the net-of-tax rate of return i s 10 percent, then the marginal effective tax rate on capital i s 50 percent. 2.20 It should be noted that the analysis of marginaleffective tax rates providedinthis study deals only with profitable firms. By "profitable" we mean those firms that have taxable income and will pay income tax if not granted a tax holiday. This assumption is important because a tax holiday has no value to an unprofitable firm, which does not have to pay income taxes anyway. Effective Tax Rates on CapitalInvestment by Domestic Firms 2.21 Our estimates of the effective tax rates on capital incorporate company income tax, sales tax on machinery and eq~ipment'~,propertytax, property transfer tax, andtaxes l2See Annex 1Section C for informationregarding the impact of non-tax factors on the marginal effective tax rate. l3According to the Ministry of Finance, Tax Analysis Department, imported machinery and equipment (M&E) are exemptedfrom import duty andthe 10percentsales tax, which account for 70percentofM&E used by manufacturing industry. Therefore, we assume an effective sales tax rate of 3 percent on M&E in our h4ETRanalysis. 11  on financial investments. The estimates are made for manufacturing, manufacturing- related services, agriculture, and tourism. We consider six options: no incentive; basic Pioneer Status and initial allowance; enhanced Pioneer Status and initial allowance; Pioneer Status and initial allowance for strategic projects; Pioneer Status and initial allowance for SMEs; and acceleratedcapital allowance. For manufacturing industry, we also present effective tax rates by type of assets and by type of incentives. The four types of capital assets in this analysis are buildings, machinery and equipment, land, and inventory. Table 2.3 and Table 2.4 presentthe estimate^.'^ 2.22 Note that since the METRmeasures the tax impact on a single incremental or new unit of capital investment (e.g. an investment project), the initial allowance is relevant only when it i s counted for a given capital investment project. In other words, METR analysis does not count a "cluster" of initial allowances available for continued capital investment over a multi-year time horizon. Further, for simplicity, we assume that the initial allowance in our analysis i s fully utilized in all cases. Therefore, in both Table 2.3 and Table 2.4, the initial allowance i s presentedwithout indicatingeither the time horizon (e.g. for five or ten years) or restrictions on the deductible portion of taxable income. l4For detailsof the methodology, see Annex 1Section B. 12  2' ' ? cu '4 ' ' R/ ' I ; .-0QP c n .-P t 0 c xE I -E dc Q ' 2 v) K E .-01 C L c Y a I 0 : 3 v) ac 13  + 41 6 W 2 7 2 2 7 rc 7 7 +- 2 cq 7 7 2 Ln 2 7 7 II 7 2 Ln 2cu 7 2; .-UI E L ca 20 se a 14  2.23 The two tables show that: For a company with Pioneer Status, non-depreciable assets (i.e., land and inventory) are taxed less than depreciable assets (i.e. buildings and machinery and equipment). This is mainly becauseby being exempted from some or all income tax (because of their Pioneer Status), investment projects receive less benefit from income tax allowances such as deductibility for interest expense (unadjusted for inflation) and the annual depreciation allowance. While all types of assets incur interest expenses, only depreciable assets incur replacement cost as measured by accounting depreciation rates. Therefore; under Pioneer Status the loss of tax benefits i s greater for depreciable assets than that for non-depreciable assets. Except under Pioneer Status, machinery and equipment are the lowest taxed among the four types of assets. This is a result of the generous initial allowance given to depreciable assets. Although this initial allowance i s equally available for buildings, the annual allowance for buildings i s much lower than that for machinery. Moreover, buildings are subject to two taxes that do not apply to machinery and equipment: the property transfer tax and the property tax. Infact, machinery and equipment might be taxed even less if there were no sales tax collected from domestically made machinery and equipment. Of all the incentives available to manufacturing, the 100 percent initial allowance appears to be the most generous. Since buildings and machinery and equipment, on average, account for more than 70 percent of capital investment in manufacturing, the 100 percent initial allowance implies that on average more than 70 percent of total capital investment can be written off against taxable income. Therefore, this incentive i s particularly attractive to those manufacturing industries whose capital investment i s more intensive indepreciable assets-particularly machinery and equipment. The efSective tax rates on some types of assets are negative under certain incentive conditions, indicating an excessive tax subsidy to these types of investment. Of course, a negative effective tax rate does not imply that the government i s giving a cash handout as a tax subsidy. Instead, excessively subsidized investment in certain types of assets may make possible reductions in project-wide or firm-wide tax liability. Pioneer Status benefits the agricultural sector more, while initial allowances benefit the tourism and manufacturing sectors more. This i s mainly because, as mentioned above, Pioneer Status benefits non-depreciable assets more and the initial allowance benefits depreciable assets more; depreciable assets account for more than 70 percent of the capital investment undertaken by manufacturing industry, while non- depreciable assets account for almost 70 percent of the capital invested in agriculture.l5 l5Itshould be cautioned that the capital structure by industrythat underlies this calculation is adopted from Canadian data, and may not accurately reflect Malaysianconditions. See Table Al.1inAnnex 1Section 3. 15  The sales tax on domestically made capital goods, unlike a value-added tax, imposes a direct tax cost on capital investment. Our simulation shows that if the sales tax on capital goods were eliminated, the METR on capital investment for all three sectors in our analysis would be 1 to 6 percentage points lower than shown in Table 2.3 and Table 2.4. The larger the share of machinery and equipment in a capital investment, the more an investor would gain from the elimination of this sales tax. Not surprisingly, manufacturing would be the biggest winner from such a tax change. Among non-depreciable assets, land i s taxed more highly than inventory under all kinds of tax regimes, because of the property tax and property transfer tax, which apply only to land. 2.24 Given the above observations, we can conclude that the Malaysian incentive program is too complicated, inefficient, and has causedexcessive tax expenditures for the government. C. TAXIMPACTINVESTMENT:ONFOREIGN A CROSS-COUNTRY COMPARISON 2.25 How do Malaysia's investment incentives for foreign investors compare with those of competitor countries in the region? This i s an unavoidable question in attempts to restructure Malaysia's investment incentive program. In Malaysian eyes, these competitor countries include China, Indonesia, Singapore, and Thailand. We first outline the major features of the tax structure and incentive programs in those four countries.16 Then, using marginal effective tax rate analysis, we compare the tax cost on foreign capital investment inmanufacturing industry inMalaysia and the four countries. Incentive Programs inCompetitor Countries 2.26 There i s a great deal of similarity between Malaysia's statutory tax and incentive provisions and those of its competitor countries for foreign investment in Southeast Asia, as shown in Table 2.1. The exception is Singapore, which has been moving towards a simpler andmore efficient tax structure with fewer selective tax incentives. 2.27 As to statutory tax provisions: 0 All countries but Singapore use multiple income tax rates, and none uses a higher rate than 30 percent. China's lower corporate income tax (CIT) rates are for foreign investors only and vary by special zones. Inthe other three countries including Malaysia, the lower rates are for SMEs. Singapore has reduced its single CIT rate from 25.5 percent in2001 to 22 percent in2004 and will reduce it further to 20 percent in2005. Malaysia and Singapore provide an initial allowance for capital investment in general. Other countries do not. l6For more details see Annex 1Section D. 16  All countries have similar annual depreciation allowances of between 3 and 5 percent for buildings, but the annual allowance for machinery and equipment varies significantly from as low as 14 percent in China to as high as 33.3 in Singapore. Both Malaysia and Singapore allow operating losses to be carried forward indefinitely, which makes tax holidays less attractive. Other countries allow only five years inwhich businesslosses can be carried over. Both Malaysia and Indonesia levy property tax based on the purchase cost of property; other countries levy property tax basedon its rental value. All countries levy a property transfer tax, which ranges from 1 percent in Thailand to 10 percent inSingapore. All countries except Singapore exempt machinery and equipment from import duty. InSingapore, the import duty is generally zero and hence no exemption is needed. All countries butMalaysia have adoptedthe value-added tax, which by nature is supposed to have no impact on the cost of capital inve~tment.'~ Malaysia, asIn mentioned above, the sales tax imposed on domestically made capital goods has directly increasedthe cost of capital investment. 2.28 As to incentives, all countries except Indonesia grant tax holidays to certain types of investment projects, although the focus of these tax holidays differs. While Singapore has been scaling down its tax holiday provisions, China does not appear to be changing its massive conditional incentive programs, which benefit only foreign investors. 2.29 China provides foreign investors with various tax incentive packages. The most common i s a five-year tax incentive package that i s available to all foreign enterprises engaged in production and operating for ten years or more. This package comprises a two-year tax holiday and a subsequent three-year 50 percent tax reduction. The tax holiday starts with an enterprise's "first profit-making year," which i s defined to be the first profitable year after the enterprise has commenced production or operations and earlier losses have been carried forward and set off against profits." China also allows reduced tax rates of 24 percent and 15 percent for those foreign manufacturing firms located in designated areas, including special economic zones (15 percent), open coastal economic zones (24 percent), and the middle and western regions (15 percent). China offers a further reduced income tax rate of 10 percent to those foreign manufacturing firms in the specified zones ifthe firms are determined to be either technologically advanced or export-oriented. These tax holidays and reduced tax rates mean that a manufacturing firm located in a special economic zone and using advanced technology may enjoy a two-year tax holiday, a three-year reduced income tax rate of 7.5 percent, and a 10percent income tax rate for the rest of its business life. l7In reality, some countries may be slow in refunding input tax credits under the VAT, which can cause significant cash-flow problems for business andeffectively raise the cost of investment. This feature of tax holidays provided in China is unseen in any other tax regimes we have observed. Unfortunately, by focusing on the case of "profitable" firms, our METR analysis will not catch this level of generosity of tax holidays. 17  2.30 Indonesiaintroduced several incentives in 2000. Among them, the most relevant to our study i s the accelerateddepreciation allowance available to enterprises located in the development zones. This allowance doubles the regular depreciation rate available for buildings and machinery and equipment. As a result, the depreciation rate for manufacturing buildings i s 10 percent, using the straight-line method, while that for machinery and equipment i s 18.75 percent using the straight-line method, or 37.5 percent using the declining-balance method. This accelerateddepreciation allowance, if not fully utilizedduring any given year, may be carried forwardfor up to 10 years." 2.31 Singapore provides various tax incentives for different types of investment projects in manufacturing industry. The main incentive i s that of income tax relief through a five-year tax holiday and a subsequentfive-year tax reduction (normally with a 15 percent CIT rate) that i s granted to "pioneer companies." Tax reduction for a limited number of years i s also available to companies that manufacture for export. The alternative to these incentives for "pioneer companies" and exporters i s a generous investment allowance of up to 100 percent in addition to the regular annual allowance, which provides a write-off of up to 200 percent of capital expenditure against taxable income. The most commonly granted investment allowance is 30 percent. Notably, an official evaluation indicates that "it i s not clear whether the tax incentives add anything to Singapore's already attractive status as a relatively stable economy and therefore a good destination for foreign investment." As mentioned above, Singapore has been moving towards adopting a lower corporate tax rate, which would in itself suggest a lowering of the value of incentives to potential investors. In the meantime, Singapore has discontinued certain incentive programs and does not intend to revive them in the near future.2o 2.32 Thailand provides tax holidays and import duty reduction or exemption for "promoted investment" projects identifiedby their product andbusiness location. For this purpose, Thailand i s divided into three zones.21The length of the tax exemption i s three years for investments in Zone 1, up to five years for investments in Zone 2, and eight years for investments inZone 3. The amount of the exemption i s capped at 100percent of investment capital (excluding the costs of land and working capital).22This restriction makes the tax holiday more like a 100 percent investment allowance for depreciable assets. However, within any given length of a tax holiday, the total tax amount exempted may not be big enough to match the effective tax benefit that i s associated with a 100 percent investment allowance for depreciation assets. Thailand's Board of Investment also selected the following five targeted industries: agri-business, automobiles, fashion, As noted inAnnex 1SectionD,it is not clear whether Indonesiaprovides any tax holidays or income tax reductions. u,This paragraphis basedon informationprovidedat <www.wto.org/enelish/tratop eltm eIt~229e.htnv Zone 1 includes Bangkok and five provinces (e.g. Samut Prakan); Zone 2, 12 provinces (e.g. Ang Thong); and Zone 3 (all 58 remaining provinces). InAugust 2000, two provinces were added to Zone 2 and the 58 provinces in Zone 3 were divided into two areas based on development level of each province; incentives for the 40 more developed provinces are now more limited in scope than those for the 18 less- developedprovincesinZone 3. 22According to an amendmentto the Investment Promotion Act. 18  electronics, and high-valued services23for further incentives: these industries are eligible for exemptions of import duties on certain machinery and a corporate tax holiday for eight years, irrespective of the location of their investment projects.24Effective from 19 September 2002, Thailand relaxed its zoning requirements for various promoted activities: for example, projects that were previously required to be located in either Zones 2 or 3 now qualify if they are located in Zone 1. Certain export requirements tied to incentives granted to firms located in export processing zones were also eliminated on 1January2003. Malaysia's Tax Competitivenessinthe Region 2.33 We now apply marginal effective tax rate analysis to evaluate Malaysia's tax competitiveness for foreign investment in comparison with its major competitor countries. In our analysis, we take the United States as the origin of a representative foreign investor. 2.34 As is well known, income earned in a host country by a foreign investor may or may not be taxed inthe investor's home country. There are two models, of which one can be represented by the U S system and the other by the French tax system. For investors from the US, income earned in Malaysia will be taxed in the U S but with a credit for taxes paid in Malaysia.25For investors from France, by contrast, income earned in Malaysia i s not considered taxable inFrance. 2.35 Table 2.5 provides an effective tax rate comparison for a representative U S company investing in manufacturing in Malaysia, China, Indonesia, Singapore, and Thailand. The specific incentives includedinthis comparison are the following: 0 Malaysia: For investment projects subject to the general CIT rate of 28 percent: 0 No incentives 0 General Pioneer Status, which provides a 70 percent tax reduction for 5 years. 0 Enhanced Pioneer Status, which provides a 100percent tax exemption for ten years. 0 The 60 percent initial allowance for 5 years. 0 The enhancedinitial allowance of 100percent for five years. China 0 No incentives for a foreign firm located in an area other than special zones that are subject to the regular CIT rate of 30 percent. 0 A two-year tax holiday, followed by a three-year 50 percent tax ~~ 23 BO1Bulletin No.68145, 10April 2002 "BO1Targets Five Industries for Proactive Marketing Campaign". <http://www. boi.go.th/english/focus/boipr_target~industriesO2.html> 24 These incentives also seem to apply to enterprises engaged in other priority activities. Priority activities are: agriculture and agriculture products; technological and human resource development; public utilities ''andinfrastructure;implementation environmental protection; and other targeted industries. Owing to the of the post-1988 interest allocation rule, the taxation of foreign income in the US canbe slightly morecomplicated than that mentioned inthe text. 19  reduction for investment projects located in special zones that are subject to a 24 percent CIT rate. o A two-year tax holiday followed by a 3-year, 50 percent tax reduction for investment projects located in special zones that are subject to a 15 percent CIT rate. 0Indonesia o Noincentives. o The double annual allowance available to investment projects located inthe development zones. 0Singapore o No incentives. o A five-year tax holiday followed by a five-year reduced CIT rate of 15 percent. o A three-year fast write-off for machinery andequipment. Thailand o The three-yeartax holiday for projects located inZone 1. o The eight-year tax holiday for selective products regardless of the location of the investment projects. 20  21  2.36 Table 2.5 presents effective tax rates simulated for the cases listed above for each country. Note that without further investigation on how foreign manufacturing firms fall into a given category of incentives, the METR comparisons presented below may not be representative for certain types of manufacturing investment. But what can be observed are the relative advantages or disadvantages of similar types of incentives provided by different countries. 2.37 The simulations show, first, that if none of the five countries granted any incentives, foreign manufacturing firms would find Singapore to be the lowest-taxed jurisdiction, with a METR of 16 percent, and China the highest, with a METR of 29 percent. Malaysia, closely followed by Indonesia and Thailand, appears to be in the middle with a METR of 24 percent. As mentioned, Singapore's tax regime has been moving towards a lower CIT rate and broader tax base (i.e. fewer conditional incentives). Therefore, Singapore's low effective tax rate under the no-incentive case can be seen as broadly applicable to foreign manufacturing firms; it i s certainly more attractive than that offered by Malaysia and the other benchmark countries. On the other hand, according to lBFD,more than 80 percent of the foreign firms inChinaare locatedin special zones and hence enjoy a lower effective tax rate than that associatedwith the "no incentive" case. 2.38 Second, for tax holiday cases, the general and enhanced Pioneer Status in Malaysia, with an effective tax rate of 11 and 6 percent respectively, appears to be competitive with the tax holidays provided in other countries. For example, the effective tax rate for tax holiday cases ranges from 8 to 12 percent in China, 6 percent in Singapore, and 5 to 8 percent in Thailand. The differences mainly depend on the length of tax holiday and the degree of tax exemption. 2.39 Third, among the countries with accelerated depreciation allowances, both Indonesia and Singapore appear to provide a much simpler incentive structure that can achieve similar results without overburdening the tax administration and taxpayers. In contrast, the effect of Malaysia's generous initial allowance i s partially offset by the sales tax on domestically made machinery and equipment. It should also be noted that the Malaysian initial allowance i s structured in a very complex manner and made available only for a specified number of years. 2.40 Fourth, the gap intax burden measuredby METR between the non-incentive case and the most generous incentive case appears to be widest in China (29 percent vs. 10 percent), followed by Malaysia (24 percent vs. 6 percent) and Indonesia (25 percent vs. 12 percent). These large gaps indicate the significant tax distortion caused by the statutory tax and incentive structure in these countries. B y this measure Thailand, closely followed by Singapore, appears to provide the tax structure with the least distortion. 2.41 Finally, for illustrative purposes, a simulation for all the Malaysian cases i s made assumingthere were no sales tax on capital goods. This shows that without the sales tax, the effective tax rate in all the cases would drop at least 2 percentage points below its current level. 22  D. CONCLUSIONS 2.42 To sum up, by comparison with those of China, Indonesia, Singapore, and Thailand, Malaysia's incentive regime i s complex and difficult to understand. It provides an invitation to unfair outcomes at best and corruption at worst, and for some types of business it i s excessively generous. The incentives offered under the Promotion of Investment Act, in particular the whole or partial tax holidays for Pioneer Status investments, are particularly problematic. Where tax holidays are combined with the extremely generous and deferrable write-offs, Malaysia's system not only increases the after-tax profits of profitable companies but in many cases can transform an investment that i s not profitable before taxes into one that i s profitable after taxes. For most companies that get these incentives, the corporate income tax (CIT) i s effectively at or below zero. 2.43 Overall, these results suggest that Malaysia needs to simplify and increase the efficiency of its tax and incentive tax structure. Towards that end, recent changes made by the Singapore tax authority in eliminating many selective tax incentives and lowering the general corporate income tax rate couldcertainly provide an inspiration for Malaysia. 23  3. FISCALINCENTIVES FORRESEARCHAND DEVELOPMENT 3.1 In this chapter, we study the fiscal incentives that Malaysia uses to stimulate research and development (R&D) in Malaysia. Section A outlines the economic importance of R&D and the economic rationale for using fiscal incentives to stimulate R&D. Section B describes Malaysia's R&D tax incentives system, focusing not only on the incentives themselves but also on their institutionalunderpinnings and their operation inpractice, andthenoutlines approachesfor evaluating andcosting tax incentivepolicies. Section C assesses the cost of current R&D incentives in terms of forgone tax revenue, and then uses marginal effective tax rate (METR) analysis to assess how the R&D behavior of firms would change in response to a change in tax incentives; the case of the current Double Deduction (DD) for R&D i s used to illustrate the application of the methodology. Section D analyzes the context for Malaysia's R&D tax incentives; it describes the country's many other policies for encouraging R&D and benchmarks Malaysia's R&D incentives against those of competitor countries and against the average for OECD countries. SectionEoffers recommendations. A. CONCEPTUAL FRAMEWORK I 3.2 R&D is associated with the creation of new ideas, technologies, materials, and products that help to advance the quality of life. The economic literature accords it great importance as a means of sustaining economic growth, since growth can only be sustained by using the existing stock of capital more efficiently and hence increasing the productivity of the economy. The work of Romer (1990), Grossman and Helpman (1991), and Aghion and Howitt (1992) shows how R&D plays a crucial role inincreasing efficiency. This literature, now known as Endogenous Growth Theory, shifted the focus of economic growth theorists away from increasing the stock of capital towards the pursuit of efficient andproductive use of capital and labor as a means to achieve greater wealth. Why ShouldR&Dbesubsidized? 3.3 R&D is a public good, whose social return is greater than can be captured by individual investors undertaking it and whose value to society will never be reflected in marketprices.26Thus while afirm may enjoy a 10percent rate of returnon its own R&D, society may enjoy a much higher rate of return because of positive spillovers. 26The idea of public goods and positive externalities is used in a wide array o f areas. Education is subsidizedfor a range of reasons, one being the positive externality of having a well educatedpopulation is greater than just the private return to the individual. The converse is also true for negative externalities.. Factories are increasingly made to pay not just for the fuel they use, but also the pollution they produce as this has a strong negative externality for the people living in nearby areas as well as the environment in general. Externalities are often related to market failures (there tend to not be purchasers of pollution for example) and hence a role for the state inbridging the failures becomesfeasible via taxes or subsidies. See Stiglitz (2000) for further details. 24  3.4 Using this idea, Figure 3.1 provides a conceptual framework for thinking about the role of subsidies to R&D. Figure 3.1:ConceptualFramework Showing Social Return vs. Private Return Price I I I Q - Q" Quantity 3.5 Say the government felt that the volume of R&D being carried out (Q) was insufficientand wanted to increasethis activity to level Q*. The demand for R&D, say by a car manufacturer, i s shown as the downward sloping Private Return curve. The government can then award a subsidy S*, say by providing tax incentives to the car manufacturer to commission R&D, such that the car manufacturer will carry out more R&D. This increasedR&D will then coincide with the social optimal level Q*. 3.6 A number of points are worth noting. The government can subsidize too much (S>S*) or too little (S<S*) such that an excessive or insufficient amount of R&D i s done. The government could also subsidize the producers of R&D. This is equivalent to lowering the cost of R&D to the level of the Lower Cost of R&D curve whilst keeping the Private Return curve fixed, as in Figure 3.2. Inprinciple the government will achieve its objectives of Q 3 Q* whichever way it chooses to give the subsidy. However, the distribution of the immediate benefits to the producers and consumers of the R&D will differ depending on the type of subsidy. 25  Figure 3.2: ConceptualFramework showing Lower Cost of R&D Price I Lower Cost of R&D Private Return I Q - I Q" Quantity The Nature of Spillovers 3.7 Economists have traditionally debated the nature of R&D spillovers. There are essentially two views in this debate. The first view holds that ideas travel short distances most easily, and that the spillovers are strongly local in nature and occur mainly through personal contacts. As personal contacts tend, on average, to occur more when people are geographically close, the implication is that spillovers of R&D are localized. The other view maintains that with increasedglobalization and internationalization, knowledge and invention are able to move further afield. Potential mechanisms for this spread of knowledge are patents and innovation that are embodied innew products. 3.8 In a seminal paper, Jaffe, Trajtenberg, and Henderson (1993) found strong evidence for the hypothesis that spillovers were local in nature. They studied the geographic citations of patents and found that the closer two inventors are geographically, the more likely they are to learn from each other's work. Similar work by Jaffe and Trajtenberg (1999) showed that firms benefit most from R&D that i s done by other firms that are geographically close to them. 3.9 If knowledge spillovers are strongly local, government subsidies for R&D should focus on R&D that i s done locally. This i s a different principle from subsidizing R&D done by a localfinn, because a local firm may be carrying out its R&D in a foreign location, while conversely, a foreign firm may carry out R&D locally, with a large potential for positive spillovers. This understanding of spillover effects i s a primary reason why almost all OECD countries provide R&D incentives only for R&D that i s carried out within their borders.27 27The other mainreason is to curtail tax avoidance, as discussed later on. 26  B. OVERVIEWOFMALAYSIAN TAXINCENTIVES R&D 3.10 Malaysia's R&D incentives described here are the ones that are systematically available through the tax code and apply to all firms. Table 3.1 provides a summary of the R&D tax incentives. Type of Company Tax Allowance Period Requirements Contract R&D Tax exemptionon 5 years ProvidesR&D servicesto Company (Pioneer 100%of Statutory non-relatedcompanyin Status) Income Malaysia Contract R&D 100%of qualifying Expenditureincurred Provides R&Dservicesto Company (IA) capitalexpenditure within 10years non-relatedcompany in incurredwithin 10years Malaysia aaainst 70% of Siatutory Income R&D Company . - II100%of qualifying I Expenditureincurred IProvidesR&D services to capitalexpenditure within 10years relatedcompany or any incurredwithin 10years other companyin against 70%of Malaysia Siatutory Income In House R&D 50% of qualifying Expenditureincurred Only does researchfor capitalexpenditure within 10years relatedcompanies incurredwithin 10years I aaainst70% of Siatutory Income Source: MIDA 2004. 3.11 Companies that carry out R&D are of three types: contract R&D companies, R&D companies, and in-house R&D companies. Contract R&D companies can choose which tax status they prefer, that i s between Pioneer Status or receiving the Investment Allowance. A company's choice generally depends on the size of its capital expenditure and the revenue it expects to make. A company that expects to have high capital expenditure will likely choose the IA, whereas one that expects to have high revenues will probably choose Pioneer Status. An R&D company has two options--either to claim the IA or to forgo this allowance so that its parent or sister companies will be able to claim a Double Deduction on any R&D expenditure that it undertakes. (Note that if the R&D company chooses to take the IA, its parendsister organizations cannot claim a Double Deduction on R&D expenditure for this work.) 3.12 Companies that commission R&D are allowed to claim a Double Deduction for non-capital spending (DD) against their statutory income, provided that a related R&D company i s not already claiming an IA. Additionally, if R&D undertaken overseas i s less than 30 percent of the total cost, the company may be eligible for a DD. These cases are considered on acase-by-casebasis. StructuralAssessment 3.13 Below we describe the findings of our analysis of Malaysia's institutional arrangements for selecting, evaluating, and implementing R&D tax policy. Whilst this 27  discussion focuses on R&D tax policy, and i s basedon interviews2*and analysis of R&D tax policy, most of the recommendations are likely to apply more broadly to the whole of corporate tax policy. 3.14 The structural assessment i s based on four factors: design, operations, tax abuse, and R&D definition. Since policy recommendations need to be tailored to the Malaysian tax code and practice, we suggest undertaking consultations with government, tax practitioners, and the private sector in order to further illuminate these four factors and the issues associated with them. A full consultation i s vital to ensure that the tax code provides the necessary incentives without abuse or burdensome input and monitoring by government bodies. Annex 4 provides guidelines for undertaking consultations. Design 3.15 Malaysian R&D tax incentives are simple, which i s very important for encouraging uptake and for minimizing the administrative burden on companies and government. The Double Deduction operates as a simple volume system for current R&D. This differs significantly from the practice in the US and France, for example, where rolling bases are used to define the incentive. The key difference between the two systems (volume and rolling) i s that the rolling system provides incentives only on the R&D that is over and above the volume of R&D from some base year(s). B y contrast, the volume credit provides incentives on all the R&D that i s done. The advantage of the volume-based system i s its simplicity, because in order for these schemes to have an impact they needto be readily understood within firms as well as by tax accountants.For full details on the pros and cons of each system see Bloom et al. (2001). One of the strengths of the tax incentive system usedinMalaysiai s the fact that it i s volume-ba~ed.~~ Further, by channeling these tax credits through the corporate tax system (rather than through grants or a separate tax code) Malaysia's system uses an existing payment mechanism and so does not create any additional transactional burden. 3.16 There are, however, some design issues that should be addressedto strengthen the policy. The first of these i s providing incentives for all R&D undertaken in Malaysia. BecauseMalaysiaprovides the tax incentives to the party financing the R&D, rather than to the party undertaking the R&D, some of the R&D taking place within the country i s left ineligible for the R&D Double Deduction. This i s the case when the buyer of the R&D is foreign. We are aware that the IA/Pioneer scheme provides incentives to Malaysian companies doing R&D. But given the valuable local spillovers from R&D, as outlined earlier in this chapter, the government should consider also giving the Double Deduction to Malaysian firms that are undertaking R&D funded by foreign companies. We think it i s important to design the DDcorrectly, because the DDprovides much larger fiscal support for R&D than the W i o n e e r schemes, and also because elsewhere in this report we recommend winding down the IA/Pioneer schemes. 28See Annex 2 SectionEfor list of interviews. 29The OECD has recently classified the UK R&D tax credit as `best practice'. This judgment largely reflects the simplicity of the UK system, which uses a volume tax credit. 28  3.17 The second issue is that of diferential rates of provision on similar R&D. As noted above, different rates are available on potentially identical R&D depending on the sort of company doing the R&D. But since there i s no reasonto believe that the different company structures are generating different levels of R&D spillovers, they should in principle all receive the same level of subsidy. Further, in practice the range of different subsidy rates on R&D activity is distorting company behavior in a wasteful way, in that some companies we spoke to appearedto have changed their corporate structure solely to claim more R&D subsidies. We appreciate that the need to prevent tax abuse i s the primaryjustification for the current structure, and this i s entirely appropriate. But it i s still important to find out whether the current system can be rationalized without enabling tax abuse. Operation 3.18 We heard from a number of sources that while the design of the R&D credit system i s simple and transparent, its implementation by the IRB is not. One person we spoke to said that "The IRB seems determined to deduct claims from all my clients whether or not they are legitimate" and another stated "They feel the need to counteract the softness of the MIDA schemes by being tough on Double Deduction". The overall impression we received - which is confirmed for the whole of the Malaysian corporate tax system by the PICS survey - is that the credit i s well designed but inconsistently implemented. One respondent even quoted a foreign investor as being aware of this problemand seeingit as a disadvantageof movingR&D to Malaysia. 3.19 Some respondentsidentifiedIRB's lack of resources and the different application of definitions by MIDA and IRB as the causes of inconsistent implementation. They said that although IRB has some very high caliber staff, the staff were overburdened and unable to cope with the volume of claims, andthat this was leading to somewhat arbitrary decisions. These respondentsarguedthat any saving made by limitingIRB's resources i s a false economy, and strongly urged the government to provide more resources for IRB. One approach could be to reduce the generosity of the IA and Pioneer schemes, by unifying or removing them, and usingsome of the savings to fund more staff at IRB. This step could both increase the impact on R&D spending in Malaysia and reduce the effective budgetary cost to the Ministry of Finance. The resources that IRB needs to monitor the DD, IA, and Pioneer schemes should be assessed as soon as possible. A second component of this question is whether IRB staff are adequately trained or scientifically aware of which activities are R&D activities and which are not. One respondent we spoke to claimed that IRB officials "only think something is R&D if there i s heavy machinery involved". 3.20 Additionally we received comments from various companies that the time taken to approve DD claims was often very long. This tends to create a significant problem for smaller companies. The solution is to speed up processing time, but this must only be done if IRB's resources are increased significantly. We did not review IRB's internal processes, but there may be some scope for efficiency gains, so long as they do not underminethe quality of IRB's review. 29  Taxabuse 3.21 A worrying theme in our encounters with many firms was the potential for tax abuse in the schemes. This seemed to be particularly prevalent for the IA and Pioneer tax schemes. Though these schemes are little used in practice, their current design offers a number of areas for potential abuse, and in our discussions with tax practitioners it was clear that some clients were aware of these areas. In order of importance, the areas of most concern were: 0 Checks on R&D activities. There seemed to be significant variation in the frequency of checks upon firms to confirm that their activities were legitimately R&D. Among the sample of the seven firms we spoke to, two had not been checked in more than a decade! One firm had been confirmed as an R&D company by MIDA in 1993, but since then had been involved in a major restructuring, which entailed stopping the activity that had led to the granting of the R&D firm status and moving into an entirely different field. No IRB official had returned to confirm that the new activity was indeedR&D (most of it was) and eligible for the credit. Far more controls are needed, bothfor newfirms obtaining R&D credits and alsofor the current stock of firms claiming the credit. A three-year period between checks would seem reasonable. 0 Transfer pricing. There was concern that abuse of transfer pricing could lead to tax abuse. For example, Group AB owns Company A which makes cars and Company B which does R&D on cars. Company A buys $100 worth of R&D from Company B. If company A pays $100, Group AB obtains $100 worth of R&D double deductions. But if it charges a far higher fee - say $400 - then Group AB obtains $400 of R&D double-deductions, reducing its overall tax bill by $84. Hence, by simply increasing the internal transfer price of R&D, Group AB can reduce its tax liability. We were told that transfer-pricing abuse was fertile ground for tax avoidance as it was "very hard to check". The government is clearly aware of this issue and is currently providing guidelines on transfer pricing that are designed to clarify these issues. This i s an important step forward. Nevertheless, with the generosity of the current R&D tax credit scheme, it i s important to ensure that the transfer pricing rules are as strong as possible. We were told, for example, that the Singapore Government had introduced a legal framework on transfer pricing, and if so, this could provide a benchmark for Malaysia. It would be usefulfor the government to consider improvementsin its transfer pricing regime, and we have incorporated relevant questions in the guidefor consultations (Annex4). 0 Abuse of contract R&D schemes. Since an R&D contract company receives a much stronger fiscal incentive for R&D than does either an R&D company or an in-house R&D company, there are strong incentives to change corporate structure to obtain the higher rate of credits. As discussed above, this i s both conceptually undesirable (why should the same R&D receive different levels of subsidy?) and operationally undesirable (tax policy should try to avoid providing incentives for companies to change corporate structure solely for tax reasons). In addition, there i s potential for 30  abuse when a linked company succeeds in labeling itself an R&D contract company when infact it does R&D only or mainly for its parent company. We came across one example of this - a firm that designed and manufactured a range of products. This firm had separated its design arm from its production arm and then requested that clients procure these services separately. Under this set-up the design arm formally did no R&D for the manufacturing arm (because it contracted directly with the clients) and so was legally a "contract R&D company. But in practice the design arm was effectively doing R&D solely for the use of the manufacturing arm (or its clients). While this strategy does not amount to tax abuse as defined in the legislation it clearly defeatsthe purposeof the contract versus in-house definition. If the different rates across the IMPioneer schemes cannot be unified structurally, attention should be paid tofirms operating as contract R&D companies. These may not, in effect, be true contract companies but linked to their parent companies and involved in tax avoidanceor borderline behavior. 0 Checking of overseas R&D. The double-deduction scheme allows a firm to claim for R&D undertaken overseas as long as this accounts for less than 30 percent of the firm's total R&D bill. In effect this provision allows firms to claim tax credits for R&D that IRB has no ability to check, opening up the potential for substantial abuse. In addition, as outlined above, R&D spillovers are typically local in nature, so Malaysia's economy will not benefit from R&D undertaken abroad. 0 R&D capital resale. The IA and pioneer schemes provide R&D tax breaks against capital R&D expenditure. There appears to be no legal claw-back clause inthis, andif one exists no-one we spoke to was aware of it being imposed. The lack of such a clause means that a firm could buy a building in year 1 for $100, claim R&D tax reductions worth $28, and then sell the building in year 2 for $100 without having to return the tax deduction. Hence, by buying and reselling R&D capital, firms can obtain large R&D tax credits on R&D activity that does not take place. Becauseof the potential for such avoidance activity, no OECD country that we are aware of provides an R&D credit on capital expenditure. We would strongly advise that: (a) if a claw-back clause is not already present, one should be introduced into these tax schemes; and (b) IRB should regularly check firms' asset registers to confirm they still own these assets, and if not, request the return of the tax credits. 0 False company "rebirth ". The Pioneer and Investment Allowances provide generous credits for firms in the first five or ten years of their lives. This gives a strong incentive for a firm nearing the end of this period to reincarnate itself as a new company. For example, a group parent sets up a new wholly owned subsidiary and transfers all the activity into the new subsidiary from the current company. We were told that this type of activity was being carried out by larger multinationals (although not necessarily with R&D tax credit schemes in mind). If the government decides to maintain the IA and Pioneer incentives, how can it prevent this type of abuse from happening? One route i s through rigorous individual monitoring by IRB officials, although this will be time consuming and difficult. It i s worth noting that when we 31  discussed this scheme with officials from the UK Inland Revenue before coming to Malaysia, they said "we would never have a scheme like this because of the risk that companies just change their names and start up again". Thus having any scheme that runs for X years after birth may be structurally susceptible to tax avoidance and so innately difficult to police. 0 Firms obtaining credits on grants. The government would clearly be keen to prevent firms from obtaining 100 percent MOST1 grants to undertake R&D and then also claiming an additional double deduction against this R&D as well. We are not entirely certain what the exact position is, but suspect there may be some room for this type of abuse. LegalDefinition of R&D 3.22 The Malaysiandefinition of R&D is broader than the OECD Frascati definition of R&D on which most OECD statutes are based.30Under the OECD definition, R&D incentives apply to industrial drawing and design, industrial engineering, tooling up, and trial production undertaken for R&D and "feedback" purposes, but they do not apply to the same activities undertaken for production processes. Under the current Malaysian definition all trial, design, andprototype activity i s eligible for R&D tax incentives. 3.23 There i s no clear answer as to what should be defined as R&D and what should not, but there are two principles to guide this decision: 0 Subsidization of spillovers. As outlined earlier in this chapter, R&D within firms should be subsidized to ensure that an economically optimal level of R&D occurs. The evidence, however, is that the knowledge spillovers are greater for researchthan for development, because research i s more intangible and "blue-sky" and the knowledge it generates i s more generally applicable and typically flows more freely. Development work tends to be more applied and product-oriented, with a narrower scope of application and spillover (Griliches 1998). As a result, governments are typically keener to subsidize research rather than development, and'usually do so by directly funding research in universities. They tend to provide tax credits at the same level for both research and development, mainly because this i s a simpler option to administer. But in defining R&D they tend to include development with a spillover potential (research oriented, with broader application) and to exclude development that i s a specific private good (which only one firm can benefit from). Thus the Frascati definition of R&D includes "development R&D production" but not "process development ." Unless R&D spillovers in Malaysia take a different form from those in OECD (and we found no evidence that they do), this would argue in favor of moving to the narrower Frascati definition, providing a credit only on R&D for which spillovers accrue. 30See Table 2.2 on page 24 of the OECD document at <http://www 1.oecd.ord~ublications/e-booW929404lE.PDF> 32  0 Capability building. A second principle is to encourage the building of Malaysia's capability to undertake higher-end development and ultimately research within companies. Iti s correct to assume that a generous R&D tax credit will help to provide suitable conditions for R&D and so will help to build this capability. But the fundamental requirement for R&D capability in the long run i s a supply of highly trained scientists and engineers. Subsidizing relatively low-tech development work will help provide jobs for relatively less skilled scientists and engineers, but many of these people will never be sufficiently trained to move into higher-level occupations. A more cost-effective approach would be to fund university places in science and engineering for talented and able students. While capability building is an important aim, and a broadly defined R&D tax credit will probably help to achieve this, we question whether this is the most appropriate and cost-efective way of achieving this aim. 3.24 Moving from the conceptual to the practical, our interviews with companies in Malaysia pointedto a lack of clarity over what should and should not be defined as R&D. Malaysia i s not alone in this dilemma. After the UK introduced an R&D tax credit, there was so much uncertainty over the implementation of the definition that a second definitional consultation was undertaken. Such a process and document3' would provide useful guidance for Malaysian authorities and could be incorporated in any consultation done inMalaysia. Without detailed knowledge of the current tax code and issues around the definition it i s hard to say what should and should not be covered by such a consultation. But a fruitful approach would start by: (a) discussing with tax experts and IRB what areas typically cause the most problems and uncertainties; and (b) considering the questions outlined in the linked UK consultation. The results could then be used to educate firms and officials andbuildon the current guidance. 3.25 The UK and European revenue authorities typically do not involve R&D scientists in evaluating what i s and i s not R&D, but they do allow tax officials time to consider each case, and also provide extensive back-up documentation and guidance. So even if officials are unsure of what i s eligible they can consult a number of precedent examples and consult with other tax officials. Building up this formal and informal expertise takes time but it i s an important part of ensuring effective implementation. Oversight, Monitoring, and Evaluation 3.26 Evaluation i s essential for accurately assessing the impact of current and potential policies. Yet policy evaluation i s typically much harder than policy costing and i s an area that many OECD countries still struggle with. Only in the most efficiently run civil service ministries i s evaluation taken seriously - for example inthe US, UK, Canada and Australia - while many other countries even within the OECD pay evaluation no more than lip service. 3.27 Three general approachesare: 31See, for example, chttp:Nwww.inlandrevenue.gov.uk/randd/dti guidelines.htm> 33  0 Programevaluation for pilot schemes and policies inplace. 0 METR for proposedpolicies, andpilot schemes andpolicies inplace. e Survey and qualitative evaluation using large stratified samples of informed parties -for example, in the case of R&D tax credits, sampling 50 tax accountants who are aware of the claims (and explicitly not R&D managers) - for proposed policies, pilot schemes, and policies inplace. 3.28 Usingthe R&D fiscal incentives as an example, Malaysiacould take a number of routes to evaluate policies, as describedinwhat follows. Full program evaluation 3.29 Here the best approach in principle i s to set up a control group of companies and compare the impact of the policy on this control group versus the policy (or treated group). This involves calculating a double difference - the difference in the treated group before and after the introductionof the policy, and the difference in this change between the treated group and the control So, for example, if R&D tax incentives had been introduced in 2000, and only firms in engineering had been eligible, we would have compared the change in R&D spending by engineering firms between 1999 and 2001 with the change in R&D spending by all other firms. Any positive difference would indicate that engineering firms had increased their R&D spending more than all other firms, suggesting the R&D tax incentives hadhad an impact. 3.30 In practice, such program evaluations tend to be extremely hard to run for corporate tax policies because of the absence of a control group. (To generate a proper control group one would need to randomly exclude certain firms from the policy - a move that i s not likely to be politically or commercially acceptable. No OECD country has yet managed a formal program evaluation of R&D tax credits.) We could compare R&D spending in firms claiming favorable tax treatment against those that are tax exhausted and so cannot claim the credit, but these two groups are not necessarily comparable in other respects. Marginal effective tax rate (METR)evaluation 3.31 Without sufficient data to undertake a full program evaluation, the next best alternative i s a METR evaluation. METR analysis looks at the impact of the rate of tax on a marginal investment project. The analysis focuses on marginal projects because those are the projects that firms are indifferent over, so changes in tax policy will change behavior. To evaluatethe impact of the METR we need to: 1) Convert the tax policy into an impact on the user cost of capital. So using the King and Fullerton (1984) methodology, we go from a change incorporate tax to an impact on the user cost of capital for the affected group. This conceptual step estimates the impact of taxes on firms' marginal incentives to invest. 32 See, for example, Ravallion (2001) and Blundell and Costa-Dias (2000). 34  Convert changes from the user cost of capital into predicted changes inthe actual amount of R&D undertaken using an estimated R&D user cost ela~ticity.~~We do this by taking estimates of R&D user cost elasticities from OECD countries.34 Bloom, Griffiths, and Van Reenen (2001) look at eight OECD countries from 1979 to 1994 and estimate that the short-run elasticity of R&D to changes in the user cost i s -0.12 and that the long-run elasticity i s -0.86 (so that a 10percent fall inthe user-cost of R&D would lead to a 1.2 percent increaseinR&D inthe short run and a 8.6 percent increaseinthe long run).These estimates are confirmed by a survey of the literature by Hall and Van Reenen (2000), who report average short- and long-runcoefficients of -0.1 and -1respectively. A sample impact evaluation usingMETR analysis is presentedinSection C below. 3.32 METR calculations are useful in predicting the likely overall impact of R&D fiscal incentives on behavior-assuming there are no other important constraining factors. The main concern we have with using assumptions based on OECD countries i s that Malaysia may face constraints that OECD countries may not. For example, in Malaysia an important factor constraining firms' R&D expenditure i s the supply of skilled scientists and engineers. If firms cannot recruit sufficiently skilled employees to undertake R&D, they will be unable to increase R&D activity whatever the level of the tax credit. 3.33 Looking at tax reforms in the US for R&D and also for physical capital investment, Goolsbee (1998a, 1998b, 2003) finds that inthe short run (one year) about 50 percent of the impact goes towards pushing up wages, but that in the long run (ten+ years) wages slowly return to market levels and the full impact takes place. The economics of this process are as follows. The sudden increase in the rate of subsidy to R&D increases the demand for scientists, whose supply is relatively fixed in the short run, so their wages go up. With higher wages, fewer scientists leave the profession and morejoin. This increases the supply and pushes wages back down to their long-run level. Hence, skills are an important constraining factor in the short run, but by pushingup the wages of scientists the R&D tax credit will indirectly help to address the problem in the long runthrough increasing the supply of scientists. 3.34 constraint^.^^ As Goolsbee reports, the OECD countries face their own skills supply In the Malaysian case, skills might be a particularly binding constraint in the short run because the Malaysian (Double Deduction) incentives for R&D are much 33Elasticity i s the standardconcept used by economists to measure the impact of price changes on quantity. Elasticities are phrased in terms of a - (percent change in quantity)/ (percent change in price), so for example, an elasticity of -2 implies that a 10percentfall inprice will lead to a 20 percentrise inquantity. 34 For the METR analysis, we would have liked to use estimates of elasticities based on data from developing countries. We lookedfor estimates inthe literature for both the R&D user cost elasticity and the physical capital user cost elasticity but could not find any reliable estimates of either. We also contacted development economists including Simon Commander (London Business School), RobinBurgess (London Schoolof Economics), and Oriana Bandiera (London School of Economics). 35While OECD countries have more scientists and engineers per head than Malaysia they also spend more on R&D per head. As aresult the equilibrium balancebetween supply and demandmay not be significantly different from the Malaysian context. 35  more generous than those of OECD countries, and can be presumed to lead to a much stronger increased demand for scientists. This means that in the short run, Malaysian wages for scientists are likely to rise significantly. In turn, this implies that the effect of the DD i s likely to be towards the lower end of the elasticity measure (0.12) for a longer period, before making a transition to the higher level (0.86). This would occur with the side benefit of bringing more people into the engineering and scientific professions. 3.35 The underlying assumptions for the METR analysis are reasonably robust. The constraints that we observe are important but over the longer term, they become less so. Additionally, there is no reason to suspect that the proportional change in R&D activity caused by the tax credit will be very different in OECD and non-OECD countries. It i s the timing that is likely to be more affected. 3.36 A complication in analyzing the Malaysian situation is that the DD has been in place for several years. If the scheme has functioned well since its introduction, this would mean that some increase inR&D activity has already occurred. Inour calculations we use the working assumption that the DD has been functioning well and hence has a short-run and long-run elasticity of 0.12 and 0.86 respectively. Costing and Management of Malaysia's Current Policies 3.37 Probably the single most surprising aspect of the institutional background to Malaysia's R&D tax incentives i s the lack of costings. Without costings, policy management i s difficult. While we have subsequently received claims figures for the R&D tax credit - for which we are very grateful - the general lack of costings is a fundamental issue incorporate tax policy, because it makesit difficult to: 0 Evaluate the impact of a policy decision on the government's fiscal position, either at the time of the policy introductionor on an ongoing basis; 0 Undertake any cost-benefit analysis of policies; 0 Monitor policies over time to check on uptake and compliance. 3.38 Looking at Budget 2004, it i s clear that a number of policies that operate through the corporate tax system - such as Training Incentives, Incentives to Strengthen Industrial Linkages, and Incentives for Regional Distribution Centers - also do not appear to have any costings. Since corporation tax accounts for RM25 billion per year, some of these policies have potential costs of several RMlOO million, but without any costings this i s impossible to evaluate. 3.39 We recommend: 0 In the short run, draw up a stratified sample of around 200 firms36 and electronically compile their accounts to enable some initial cost estimates to be 36Stratification here would be heavily weighted towards large firms (who pay most corporate tax) but also include a few small firms. For example sample 100firms with more than 10,000 employees and 100firms with less than 10,000 employees.The Department of Statistics, which has stratification expertise, would be a good source to consult further on this. 36  made. This should be possible in time for Budget 2005 and should allow the Ministry of Finance to evaluate how much corporation tax i s being lost through the various schemes. 0 In the medium run, aim to electronically collate the corporation tax receipts of all the largest firms and of a large sample of smaller firms. The latter group should be selected on the basis of both their current and potential corporation tax payments, proxied by sales and company accounting profits.37The group might therefore include all firms paying more than RM1 million in corporation tax per year, plus all firms selling more than RMlO million per year, plus all firms reporting more than RM1million profits per year in their company accounts. B y compiling this sample the government will be able to analyze in far more detail, and with greater accuracy, where corporation tax i s gatheredand where it is lost. 0 In the long run, set up an electronic tax filing system that will allow for easy collation of company accounts. This i s an ambitious aim (which the UK, for example, has yet to complete) but it would provide a powerful policymaking tool. MIDAand costings 3.40 We were surprised by MIDA's lack of concern about the lack of policy costings. When asked whether they have any evaluation of the policies, MlDA staff respondedthat a high uptake would indicate an effective policy. But a high uptake in fact indicates a high fiscal cost, so that they were implicitly reporting that they evaluate a policy as being effective based on its tax cost. As an institutional framework this provides strong incentives for MIDA to pursue policies with a high fiscal cost, rather than policies that are cost effective. 3.41 Three steps would help to address this issue: 0 Require MIDA to provide full costings for all their schemes; 0 Includethese schemes inthe total MIDA departmental budget; 0 Link the departmental budget (or at least the component for the schemes) to evidence on its performance. 3.42 European countries have industry departments similar to MIDA, in recognition of the importance of promoting industry, but they control these departments much more closely. This involves monitoring how much the departments are spending, but also linking the departments' overall budgets to their demonstrated performance. This provides correct incentives for the departments to control their own costs and evaluate their own policies, devolving this task from the Ministry of Finance. Interactionbetween Departments 3.43 Our interviews with Malaysian government agencies were few and as such gave us only brief exposure to these organizations. Hence the views here are very preliminary and may be inaccurate. Further, any weakness seen in the interaction between departments is not unique to Malaysia. In the UK for example, the efforts in "Joined Up 37Company accountingprofits rather than tax accountingprofits, which may be manipulated. 37  Government" to get better interaction between departments have yet to yield significant results. 3.44 The two government organizations mainly concerned with tax incentives are MIDA and IRB.A number of issues arise from the interaction of these two departments. First,the definition of "R&D" is applied with varying levels of strictness. Firms findthis confusing. Additionally, it may cause financial difficulties for a firm, say if MIDA's advice has ledit to believe it will qualify for atax incentive and then, once the investment has been made, it discovers from IRBthat it does not qualify. The definitional differences may arise from the different aims of the two organizations. MIDA has a strong mandate to grant Pioneer and IA status to companies, and this may cause it to loosen the definition of what qualifies as R&D; indeed, we observed several organizations that had received MIDA approval for Pioneer or IA status but were not involved in R&D activities as defined on the MIDA web page. We also met briefly with officials of the Ministry of Science, Technology, and Industry, which awards grants. A question that we asked all three organizations was whether firms receiving grants would also qualify for tax incentives. All three gave different answers. This needs to be clarified. 3.45 Even if the level of interaction between IRB, MIDA, and MOST1 i s high, we suspect its quality must be low, becausethe three organizations lack the fundamental data (such as on costs, benefits, or long-run success of firms on various schemes) that would enable them to have meaningful discussions. For any friction to be realistically smoothed away, not only must there be a high level of interaction between departments but it must be supported by facts. 3.46 A general suggestion that would help firms to be aware of the various available incentives i s to have a single reference organization that can funnel the firm to the correct place. For small firms in particular, this sort of directorate would be enormously useful. Importantly though, the inconsistencies between the various government departments would first need to be worked out before any such scheme could succeed. C. SAMPLEIMPACTEVALUATIONANR&DTAXINCENTIVEOF 3.47 In this section we apply marginal effective tax rate (METR) analysis to evaluate the impact of tax changes, taking the current Double Deduction (DD) for R&D to illustrate the use of the methodology. 3.48 The impact evaluation for the DD requires estimates of the cost, in terms of forgone tax revenue, vs. the return, in terms of increased spending on research and development. We make two sets of estimates of the cost, first using data from the Mala sian Census of Production, and second using claims data supplied by MIDA and IRB.J3 38 SeeAnnex 2 Section A. 38  CostingBasedon Census Data 3.49 The Double Deduction scheme operates as an extra 100 percent deduction on current expenditure over and above the standard deduction available to all other current expenditure. From the census data we have information about total R&D expenditure and from this we make certain assumptions in order to make the problem tractable. We cost only the DD, because the census data do not permit us to cost the IA or the Pioneer Schemes.39 Costingthe IRB Double Deduction: key assumptions 3.50 We assume: 0 That 75 percent of all R&D expenditure i s claiming a DD. This seems a reasonablefigure to us, inthat some firms may be unable to claim the DDbecause they cannot deal with the extra administrative difficulty, while others may simply not have heard of the DD incentive. Both of these types are more likely to be small firms. This assumption can be varied to achieve upper (100 percent) and lower bounds (say 50 percent) of the participation rate. 0 That 80 percent of total R&D expenditure i s current expenditure. In OECD countries, this value i s typically 90 percent, but because Malaysia has a lower wage rate we use the lower 80 percent as a benchmark figure. 0 Corporation tax rate of 28 percent. We were made aware that the corporation tax may be reducedto 26 percent. 3.51 For completeness, we provide two sets of calculations, one assuming no tax exhaustion and one assuming full tax exhaustion, with the true figure probably lying somewhereinbetween4' The full tax exhaustion case makestwo further assumptions: 0 For lack of time series data on R&D in the census we have assumed there i s no carry-forward or carry-back of losses. For lack of census information on ultimate ownership, we have also assumed that each establishment i s a stand-alone enterprise, and thus that profits and losses cannot be traded off among establishments. 3.52 We focus our attention more on the "no tax exhaustion case", believing it more important to focus on the potential higher level figure. 41 39There are three main reasons why. The calculation requires data on gross tangible fixed assets, and we only have net assets in the census data. We also have no investment information for a given year, only the imperfect capital stock values. Third, we cannot work out what the carry-forward behavior for firms is, though this determines the revenue lost to the Government, and we cannot make simplifying assumptions to work around this problem. 40For details see Annex 2 Section C. 41Typically large companies that do R&D are profitable, even if individual establishments within any one year do not show aprofit inthe census data. Due to carry back and carry forward and group relief, they can usually claim full deduction and therefore would not be considered tax exhaust. This is why inthe analysis we have carried out in this study, we will mainly refer to the non-tax exhaust cases. The tax exhaust case serves as a lower bound to the amount of forgone revenue. 39  Costing the IRB Double Deduction: results 3.53 Table 3.2 shows the low and high estimates of revenue forgone by the government attributable to the DD, by industry class and as a total. The total value of R&Dclaims is provided by the census data andthe forgone revenue is calculated. Table 3.2: R&DExpenditureandForgoneRevenuebyIndustry (thousandRM) MIC2 digit 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 Ex 36 Manufacturln 3.54 A sample calculation of how we arrive at these values is shown below. Consider MIC2 code 32, `Electronics'. (Max Forgone Revenue)= (R&D claimed) x (CorporationTax) x (Take Up Rate) x (Current Exp. Share) (44586.4) = (265395) ~(0.28) x (0.75) x (0.8) Hence the maximum forgone revenue resulting from the Double Deduction in the year 2000 comes to roughly RM112 million. 40  Costs Based on Information fromMIDMRB 3.55 Information that we received from the IRB and MIDA on 17 June 2004 gives a breakdown of the costs and claims of each of the schemes and the numbers of companies usingthe schemes. IA/ Pioneer Status 3.56 The costings kindly provided by MIDA (Table 3.3) suggest that despite our concerns over the potential for extensive tax abuse under the MIDA schemes, these schemes have had a very limited uptake and a low cost. Only four firms, including some very large ones, claimed IAPioneer Status in 2002. Consultations with the private sector might usefully focus on the question of low uptake. Table 3.3: MIDA's InvestmentAllowance and Pioneer Schemes 1998 1999 2000 2001 2002 Cost IA and Pioneerschemes 10.6 16.7 29.4 0.3 1.2 RM million Number of companies 2 5 5 2 4 Double Deduction 3.57 These data are available only for 2001 and 2002 (Table 3.4). Nonetheless it i s informative to consider the significant levels of claims being made and the significant rise in the amount of claims being made injust one year. The claim per firm rose by 36 percent, from RM1.29 million (=286 milliod221) in 2001 to RM1.78 million (=449.7 milliod253). That is, in absoluteterms the revenue forgone by the government roseby 57 percent in twelve months. The trend is similar if we consider the DD claims of in-house R&D andR&D companies separately. Table 3.4: Breakdown of Claims on DD Year Total Total noof In-house Noof R&D No of Claim on Companies DDClaim Companies Company Companies DD (in RM (in RM mil) (In- house) DD claim (R&D mil) (in RM mil) Company) 2001 286.8 221 211.1 156 75.7 65 2002 449.7 253 347.3 181 102.4 72 Comparison of the Costsfrom Different Data Sources 3.58 Key findings are that: 0 The estimated forgone revenue as the result of DD i s RM80.3 for the year 2001, and between RM88 and 112 million for the year 2000. This i s a remarkably close fit, suggesting that the method we use may be a robust yet simple check. It also 41  suggests that this method may provide a good estimate of lost revenue on any future tax incentive schemes that the Malaysian government may choose to implement. 0 The 36 percent rise in per-firm claims of the DD from 2001-2002 i s a substantial increase. It most likely resulted from growth in awareness of the scheme, rather than from any suddenincreaseinR&D activity. 0 We have mentioned our concern that the Pioneer Status/IA scheme i s potentially open to abuse and spurious claims. The data do not back this up, but we question whether the data are indeed accurate. METRCalculations 3.59 The idea in what follows i s to derive a single quantitative measure to summarize the effects of a complex set of tax regulations on the costs of performing R&D. This measure i s the pre-tax real rate of return on the marginal investment project, which i s required to earn a minimumrate of return after tax. To focus on tax effects we will ignore changes in other variables such as inflation and interest rates, though in reality these clearly also play a role. 3.60 To calculate the impact of the Double Deduction on the METR we need to calculate the pre-tax user cost of capital both with and without the tax incentives. Taking the ratio of these two costs indicates the percentage change in the user cost-of-capital neededfor our calculations. 3.61 Followingthe methodology laidout inAnnex 2 Section B we calculate both costs, where r i s the firm's cost of capital (discount rate), d i s the depreciation rate, A i s the present discounted value of depreciation allowances, Ad i s the physical capital PDV (present discountedvalue) of depreciation rates, and Ac i s the research and development capital PDV of depreciation rates. Current pre-tax required return Post R&D double-deduction pre-tax required return =(r+d)x 1-Ad-Ac (1-t ) =1OOx( 1-Ad 1-Ad-Ac -1) Percentfall in the pre-tax required return =100x ( 1-0.254 1-0.254-0.8 *0.28 -1) (3) =43% 42  3.62 The calculations for (Ad) in (3) follow the standardPDV approach for calculating standard depreciation allowance^.^^ To estimate the impact on R&D spending in the absence of any other constraints we use the elasticity figures of Bloom, Griffith, and Van Reenen (2001), which are 1.2 percent for the short run and 8.6 percent for the long run. Here we take the short run as a one to two year period and the long run as a five to ten year period. The estimatedimpacts are as follows: Estimated short-run impact on R&D expenditure =1.2%~43% =5.2% Estimatedlong-run impact on R&D expenditure =8.6%x43% =37% Withdrawingthe Double Deduction 3.63 In the description above, the analysis is carried out assuming that the tax credit has yet to be given. Inthe Malaysian case, the tax credit has been in operation for some time. This means that we needto carry out the analysis inreverse, which i s to say that we estimate the reduction in R&D activity that would occur if the tax credit i s removed or reduced. Hence we assume that the current R&D expenditure (R&D EXPwihDD) from data providedby the Department of Statistics i s a result of the DoubleDeduction(DD) already being incorporatedinto the firm's decision.43 3.64 An additional stage inthe calculation is henceincorporated:44 Hence a DD given to a firm that had no previous tax credit would result in an increased R&D activity of which, using(2) & (3), can be rewritten as [(R&D EXpWithDD) / 1.371X 0.37 (4) 42For the depreciation schedules we assume R&D spending as 80 percent current and 20 percent capital. For the capital component adjusting standard figures this was assumed to be 10 percent information and communications technology (ICT), 50 percent plant and equipment, and 40 percent buildings. The depreciation schedules used for these three assets classes were three-year S L (straight line) write-off for ICT, eight-year SL write-off for plant and machinery, and 2.5 percent DB (declining balance) write-off for buildings. While the METR for the R&D tax credit is primarily driven by the generosity of the Double Deduction for R&D, it i s relatively insensitive to small changes in the assumptions over standard capital assets. 43Data for (R&D Exp are not available andhencemustbe calculated inthis way. The long-run impact of 37 percent is used in the calculation. To estimate the short-run impact, use 5.2 percent instead. 43  3.65 Additionally, we assume an uptake rate of 75 percent. Hence the net increase in the R&D due to the DD is given by: [(R&D EXPwithDD) / 1.371X 0.37 0.75 X The resulting estimates are incolumns 4 and 5 of Table 3.5. Table 3.5: Variation inR&DExpenditure Due to Differing Rates of Tax Incentive (thousandRM) Rateof Tax R&D Cost of Tax IncreasedR&D Increased R&D Incentive (%) Expenditure Incentive (long run) (short run) 200 (DD)- current 665374 111782.8 134774.6 24666.9 175 499024 83836.9 101081.O 18500.2 150 332687 55891.3 67387.3 12333.5 125 166343 27945.6 33693.7 6166.8 100 0 0 0 0 3.66 To estimate the effect of lowering the tax credit, we assume a linear change inthe volume of R&D activity. As can be seen from Table 3.5, there i s a roughly one-to-one ratio of the cost of the tax incentive to the long-run increase inthe volume of R&D. 3.67 Table 3.5 shows the estimated impact of the current DoubleDeduction scheme by analyzing the level of additional R&D inthe short run and inthe longrunthat the scheme i s supporting. This i s equivalent to estimating what the fall in R&D spending would be if the scheme were abandoned immediately. The estimate shows how much R&D the scheme currently supports by calculating the short and long differences between R&D now and R&D without the Double Deduction. In the short run this difference would be small, because firms would take time to reduce their R&D expenditure, but in the long runthe difference would be much greater. 3.68 Table 3.6 through Table 3.11 give an idea of the distribution of the effects of the DoubleDeductionon R&D across industries, across establishments of different sizes, and across Malaysianregions. The tables refer to non-tax exhaust cases, because these are the more common: typically, large companies that do R&D are profitable, even if individual establishments within any one year do not show profit in the census data. Due to carry- back and carry-forward and group relief, they can usually claim full deduction and therefore would not be consideredtax exhaust. However, tax exhaustion is an important issue in the design of tax incentives for smaller start-up companies, which often are not making profits and hence cannot benefit from tax incentive^.^^ For further discussion of this issue, see Annex 2 Section C. 45 Inthe UK, smaller start-up companies receive such benefits via other tax channels, such as the employer's contributions to employee pensions or Social Security, 44  Table 3.6: By Industry(Short Run) (thousandRM) 33 7747 5674 287.2 210.3 34 89822 85319 3329.9 3163.0 35 1364 693 50.6 25.7 36 5377 3489 199.3 129.3 45  Table 3.7: By Sales (Short Run) (thousandRM) Max Min Value Value of of Total R&D Total R&D Nottax Increased Increased Sales Expenditure Exhaustat Margin R&D R&D Sales c 250 23 6 0.9 0.2 250 c Sales c 1500 519 311 19.2 11.5 1500c Sales c 10000 7914 3287 293.4 121.9 10000c Sales ~100000 101007 57019 3744.6 2113.8 Table 3.8: By Region(Short Run) (thousandRM) This is because the data received from MIDA for the states were the actual amounts of DD claimed inthe year 2002. For all other distribution analyses, we estimate the amount of DD carried out, from the total expenditure on R&D for the year 2000. To retain consistency with the other distributional analyses, we assume the same DD distribution for all the states as given by MIDA is the same as it was historically, i.e. in the year 2000. We then normalize the values such that the total amounts of R&D expenditure are the same as in2000. The data received from MIDA for the states do not include `profit' and hence we are not able to carry out the analysisfor the `non-Tax exhaust' case. 46  Table 3.9: By Industry(Long Run) (thousandRM) Table 3.10: By Sales (Long Run) (thousandliM) Value of Sales (in RM Total R&D Total R&D Not Tax Exhaust _I Increased Increased f thousands) Expenditure at Margin ~ D v ~ ~ ~ 4.7 ~ ~ Sales e 250 23 250 e Sales e 1500 519 105.1 1500e Sales c 10000 7914 1603.0 665.8 I 10000c Sales <100000 101007 57019 20459.4 11549.5 100000e Sales 555911 AII SalesRanges 605374 47  Table3.11: By Region(LongRun) (thousand RM) Kelantan 0.0 I 0.0 Notes: The calculation for the distribution for Malawian states differs from that of the other distributions. This is because the data receivedfrom MIDA for tie states were the actual amounts of DD claimed inthe year 2002. For all other distribution analyses, we estimate the amount of DD carried out from the Total Expenditure on R&D for the year 2000. To retain consistency with the other distributional analyses, we assumethe same DD distribution for all the states as given by MIDA is the same as it was historically, i.e. in the year 2000. We then normalize the values such that the total amounts of R&D expenditure are the same as in2000. The data received from MIDA for the states do not include `profit' and hence we are not able to carry out the analysis for the `non-Tax exhaust' case. D. THEOPERATIONALCONTEXT:NON-TAX REGIONALSYSTEMS AND 3.69 To see the role played by Malaysia's tax incentives for R&D in context, it i s important to review the other types of incentives being providedfor the same purpose, as well as those offered by other countries that seek to encourage R&D. Non-Tax Incentives 3.70 The bulk of government expenditure on R&D i s undertaken by MOST1through its system of grants. Though we didnot have time to assess the grant system in any detail, we didhear some anecdotal evidence that concerned us: 0 Applications for large grants are evaluated by other Malaysian scientists. The problem i s that in a country the size of Malaysia, all high-level scientists will be personally acquainted, so that obtaining independent assessments within the country will be impossible. For this reason smaller countries like the UK and Sweden, for example, pay external (typically US) scientists to evaluate their grants. 48  0 Malaysian scientists are capped in the number of grants they can receive. This i s particularly true for the IRPA. While equity considerations and the fact that researchers have to teach as well are suggested as rationale for capping, it i s worth noting that this situation causes a number of problems. First, it limits the amount of money that can flow to the best scientists so they cannot expand their research laboratories; second, it risks forcing them abroad (to countries like the U S where there are no caps); and third, it helps to ensure easier funding for less able scientists, reducing their incentives to join the private sector. Government grants may even have a negative effect if they "crowd-out" private sector research by employing a large share of all potential researchersinunproductive public sector research. 0 A strong focus on applied research. While important for a small share of research funding, such a focus often tends to breed poor research. While it is important to promote industry-university links, this should not be at the expense of university quality. The experience of the UK and US funding councils shows that forcing universities to link with companies often results in lower quality research, which in the long run i s less economically useful than higher quality basic research. Hence these councils now typically only require industry-university links for a small share of their funding grants. Therefore, it is worth noting that there is a potential trade-off between Malaysia's current mix, giving the bulk of funding to industry-university linked research, and the development of a long-run base of high quality Malaysian research, andthe training of top quality scientists and engineers. 3.71 As Figure 3.3 indicates, Malaysia has at least 22 grant, loan, or equity-based schemes for encouraging research and development, operating at the firm, institution, or individual level. Though designed to support virtually every stage of the corporate lifecycle and the assumed process from concept to commercial product, these schemes have not substantially augmented research and development spending in the private sector. Yet the Intensification of Research in Priority Areas, venture capital funds like MAVCAP, a number of funds gearedtoward supporting SMEs, skill development grants, industrial research grants, and low-interest loans certainly impose a greater fiscal burden than the tax incentives discussedabove. 3.72 This raises the possibility that government funding has crowded out private demand for Malaysia's limited pool of researchers. An impact assessment of the underlying research system and its governance is beyond the scope of the current study, but would be strongly advisable. The quality of researchoutputs, including international peer review systems, the degree of competition for funding of grants, and mechanisms to improve coordination across various schemes may be important issues. A rough estimate of the available funding is RM 300 million a year-a fairly robust amount of funding considering that there are approximately 643,000 researchersinMalaysia. 49  Figure 3.3: Malaysia's Tax Incentives for R ~ L DTip~ :the Iceberg? ~ of Tax Incentives Double Deduction Venture Capital and Grants Business Development / Commercialization Experimental Applied Research Research Technology Commercialization Acquisition of Researchand Prioritized Intensification Fund (TAF) I1 Develgop;:; Fund Research Of Research RM 238 Million In Priority Areas (IRPA) -- Industrial MAVCAP, IPTA Research Industrial IPTA Business MIMOS funds Grants Grant Scheme Development RM 230 Mill Assistance Fund Entrepreneurs RM600 Million Fund RM e00 Mill MOST1Business Development Demonstrator Multimedia Application Supercorridor Grant (DAGS) I Grant (MGS) I RM 90 Mill Special high-tech EconomicZones / Platforms (Y Skill DevelopmentFunds/ Institutions Multimedia Bio-Valley Malaysia (Not yet operational) Universityof Resources I 1 Corridor Science & Dev Fund Technolo State-level Skills Retraining Parks Technology parks Technology- Development Fund Based Inst. Fund RM 100 Million 46The list of institutions included inFigure 3-3 is not exhaustive.For example, the R&D&C grant schemes under the purviewof MOST1include: Intensification of ResearchinPriority Areas (IRPA); Industry Research and Development Grant Scheme (IGS); Multimedia Super Corridor ResearchandDevelopment Grant Scheme (MGS); Demonstrator Application Grant Scheme(DAGS); Commercialization of Research and DevelopmentFund(CRDF); Technology Acquisition Fund(TAF); Technology Acquisition Fund- Women (TAF-W); Scientific Advancement Grant Allocation (SAGA); NationalBiotechnology Directorate BIOTECHR&D Grant Scheme; andNational Oceanography Directorate R&D Grant Scheme. 50  BenchmarkingMalaysia's Fiscal Incentives 3.73 Fiscal incentives explain only a small partinthe success of other countries inEast Asia in encouraging R&D. Other critical factors are the quality and flexibility of public institutions andthe autonomy of public universities. 3.74 Table 3.12 summarizes the tax codes with respect to R&D of several major countries, primarily in OECD. The descriptions here are by no means complete.47But the international comparison does make clear that Malaysia's tax incentives are much more generous than those offered by most OECD countries and are among the most generous inEastAsia. Table 3.12: R&Dtax incentives inselectedcountries Country Year R&D R&DCapital Tax Typeof Special Foreign R&Dby Introduced Deprec. Deprec. Rate Credit Tax Treatment R&D by Foreign Rate Rate Credit for SMEs Domestic Firm Firms Canada 1960s 100% lOO%or20% 20%, Volume Yes Expenses, 20% only DE, 20% ITC. 35% for No ITC Not buildings SME France 1983 100%or 3 year SL (not 50% Rolling Yes No Unknown 5 year buildings) cap Germany nla 100% 30% DE, 4% None None Yes Unknown Unknown SL buildings Italy nla 100%or Accelerated None None Yes Unknown Unknown 5 yr. cap Japan 1966 100%or Accelerated 20% (up Roiling Yes 6% credit 20% on 5 yr. cap 5% TC to max for coop royalties buildings of 10% with foreign of CT) labs UK 2000(SME), 100% 100% 50% Volume Yes Yes 25% on 2002(other) (SME), royalties 25% us (other) 1981 100% 3 year, 15 20% Rolling Yes No Same as (contract year for foreign research buildings firm 35%) India n/a 100% 100%(except None None 30-50%on 30-50% land) royalties on royalties South nla 100% 18-20Yo lo%, Rolling Yes 10-16% on No Korea depreciation 25% royalties 5.6% on buildings Taiwan 1991 100% Accelerated 5-25% Rolling Yes 35%on 35%on depreciation royalties royalties Australia 1986 125% 3 year SL (not None None Yes up to 10% No buildings) of project costs 47For amore thorough guide see Hall and vanReenen(1998)and Sawyer (2004). 51  3.75 Given the long-standing consensus inEast Asia and the Pacific on the importance of knowledge-intensive activities to medium-term competitiveness and industrial upgrading, it i s not surprising that the economic strategy of nearly every country in the region includes measures to raise private sector spending on research and development. Countries use a wide variety of incentives for this purpose. Singapore has leveraged a combination of openness to FDI, trade, and human capital investment. Korea's strategy was based initially on protecting large domestic manufacturers, followed by a gradual opening, with government investment in technology and investment in human capital. Taiwan, by contrast, has encouraged overseas Chinese to invest inTaiwan through fiscal incentives and technology parks, often gearedtoward start-up and small businesses.Both Singapore andTaiwan have stimulated R&D through defense spending. Table 3.13: Scienceand Technology: Summary Indicators for SelectedEast Asian Economies Country Researchers 1999TIMMS-R R&D %GDP PrivateSector Yo (`000) ScienceI Math Total of R&D Notes: UNESCO statistical database, Figures for Taiwan: Indicators of Science and Technology, ROC, 2003. Malaysia Private Sector Share figure fromMASTIC (2002). 3.76 From the East Asian experience, there appears to be no strong correlation between the volume of tax incentives and either R&D as a share of GDP or private sector R&D expenditure (Table 3.13). The countries that offer the highest tax incentives - Thailand and Malaysia (Table 2.1 above) - do the least research and development. The countries that appear to do the most research and development - Taiwan and Korea - are the least generous in tax terms, but they offer non-tax incentives. Singapore falls inbetween these two poles: it offers very generousincentives and also performs well on comparative R&D indicators. However, its superior performance i s more likely related to the strength of its researchbase and basic science and math education. 3.77 There i s a much stronger correlation overall between comparative performance in research and development and the quality of the human capital base, as indicated by TIMMS-R scores (Table 3.14), than there is with tax incentives. 52  Table 3.14: TIMMS-R Scores Third InternationalMath and S -- Repeat(TIMMS-R) Mathematics - 1 Science . Nation ,, Average I _ _ II 1" 1 ~- Nation - Average _i 2 1. Singapore -- 1 -604 1. ChineseTaipei 569 2. Korea,Republicof ' 587 2. Singapore 568 " " " x _. _I- ~" 3. ChineseTaipei t 585 - 3. Hungary 552 " - . ~ - 7 - 4. Hong Kong SAR 582 4. Japan 1 550 , *- _ * I I ^ ^ i 5. Japan- ,I . 579 I_ ~ 1 - ~ " I_ 6. Belgium-Flemish -.558 6. Netherlands 545 ~ ~ "_ " _ 7. Netherlands 540 7. Australia 540 8. Slovak Republic 8. Czech Republic- I 539 9. Hungary ' 532 9. England I 538 535 * " 1 5 - 4 _ _ ~ Federation 526 12. Belgium-Flemish 535 13. ' - ! _ I I I _ I x i^- 14. Finland 520 14. Canada 533 "_ 15. Czech Republic _ -520 I 15. - I 16. - . i _ . 519 j 16. Russian Federation 529 I . -17. - ."' 511 - 17.- Bulgaria - " 518 18. Latvia-LSS 505 - - 18. UnitedStates 515 ~x 19. NewZealand 510 - " 20. Latvia-LSS 1 503 21, -NewZealand : 491 21. Italy ' 493 ~ _ " _ I I ^ 22. Lithuania _ 1 492 I 22.- Malaysia - " _ " _ iI ~ 23. Italy , 479 23. Lithuania , 488 X I r ~ " I i 476 24. Thailand 482 . * . L - 472 25. Romania 472 .. 26. Moldova 1 469 26. Israel 468 27. Thailand 467 E. CONCLUSIONSANDRECOMMENDATIONS 3.78 Malaysia's volume-based Double Deduction on current R&D expenditure i s generous and well designed. Our METR evaluation suggests that the DD has a cost of RMllO million, a short-run impact on R&D of RM25 million, and a long-run impact of RM135 million. This benefitkost ratio will be significantly reduced, however, if there are serious shortages of skilled scientists and if the operation, design, and definition of the scheme are not improved. 53  0 The operation of the scheme seems to be limitedby inadequateIRB resources. We recommendreviewingthe levels of IRBresourceswith aview to significant increases as soon aspossible. 0 To help optimize the operation, design, and definition of the scheme, we recommend consulting on a set of reforms, to be enactedinthe context of Budget 2006 at the latest. These reforms would involve focusing the schemes more on Malaysian-based R&D, consistently applying an agreedR&D definition across all government agencies, and developing proceduresfor tackling potentially serious tax avoidance. 3.79 Malaysia also has an extremely generous system of incentives for capital R&D via the IA and Pioneer schemes. The interaction of the three systems - IA, Pioneer, and DD - however, causes complexity and friction. The structural nature of the IA and Pioneer schemes also provides scope for widespreadtax abuse, which may be difficult for MIDA and the IRB to control without extensive, costly, and invasive monitoring. Firms' aversion to this heavy-handed monitoring may explain the very low uptake of the IA and Pioneer schemes. 0 We recommend phasing out these schemes, startinginBudget 2005, by granting no further IAPioneer incentives. 3.80 Though we only examined R&D schemes, the general design principles of these schemes raised concerns about complexity and the potential for tax abuse which should be investigated. Iflwhere problems are identified, the government could act to address these inor before Budget 2006. 3.81 It is important to cost all major corporate tax schemes, including those for R&D, not only as an essential check on the operation and uptake of schemes, but also as a vital component of their cost-benefit analysis. We recommend costing the R&D schemes and extending costing to all major corporate tax schemes as soon as possible, with some of the additional IRB resources beingtargeted towards this. 3.82 Over the longer term, there i s a need to continually reexamine, within the context of Malaysia's next five-year plan, the rationale for R&D tax credits. If the credits are in place to encourage positive spillovers from R&D, then this calls for a tightening up of the definitions of R&D and hence tax incentives. But if the goal i s to encourage local R&D to develop (i.e. to buildcapacity), then this is likely to require the opposite. 3.83 The volume tax credit being used in Malaysia is a good option because it is simple; it should remain a volume tax credit. However, there are concerns about its delivery and monitoring systems. There are significant gains to be made by carrying out some basic costing checks. This i s also likely to apply more generally across the whole Malaysian corporate tax system. 3.84 Finally, we feel that it i s necessary to look also at the more fundamental issues of human capital development. This will include nurturing top-class scientists and engineers as well as developing abusinessenvironment that i s conducive to R&D. 54  4. IMPACT OFFISCALINCENTIVES ONINVESTMENT INMANUFACTURING 4.1 Like many countries, Malaysia offers special incentives and R&D-promoting schemes to spur inward investment. Advocates argue that such incentives promote investments and jobs, while opponents contend that they are ineffective, have high revenue costs, and distort investment. Using data from the Productivity and Investment Climate Survey (PICS), this chapter investigates the impact of incentives on the investment rates of manufacturing establishments in Malaysia. Although data limitations prevent strong conclusions for most types of incentives and schemes, there are enough observations to conduct an extensive analysis of the reinvestment allowance and acceleratedcapital allowance. 4.2 Section A describes relevant findings of the survey. The econometric analysis in the following sections i s conductedto ascertainwhether incentives have the desired effect of increasing the investment rate of establishments.We use standardlinear regression and a semi-parametric statistical matching approach to determine the impact of the reinvestment allowance/accelerated capital allowance on the investment rate and to analyze the causal relationship between the two. In Section B we use a flexible accelerator model of firm investment to investigate the effect of the reinvestment allowance/accelerated capital allowance on the investment rate, controlling for important financial variables. In Section C, we use propensity score matching to compare the investment rates of establishments that are using the reinvestment allowance/accelerated capital allowance with the investment rates of "matched" establishments that are not usingthis allowance. 4.3 Unfortunately, the limited information on incentive use and corporate taxes in the PICS, and more importantly, the cross-sectional nature of the data (lack of panel data), prevent a cost-benefit analysis of tax incentives. While two years of financial and income recall data are available in the PICS, the question regarding incentive use does not specify the year in which the incentive was used. Furthermore, of 902 total manufacturing establishmentscovered by the PICS, data on corporate tax are missingfor 620 in 1999 and for 507 in 2000. Thus the PICS data do not capture any dynamic effects that tax incentives may have had in terms of increasing productivity, output, and income. Consequently, the evaluation of short-term government revenue loss versus medium- and long-termgains i s beyond the scope of the PICS data. A. INCENTIVES USEDBYMANUFACTURING ESTABLISHMENTS IN MALAYSIA 4.4 The PICS data show that Malaysian manufacturing establishments make moderate use of a wide array of different incentive programs and R&D-promoting schemes. Figure 4.1 shows that 47.8 percent of the 902 manufacturing establishments surveyed have gained from at least one type of incentive or scheme for their most recent investment. Many more establishments use incentives (43.7 percent) than schemes (15.7 percent). 55  Figure 4.1: Use of Incentives 70 60 50 1 I G 40 30 20 10 0 Incentives& Schemes Schemes 4.5 Second, the difference in the use of incentives by establishments of different sizes i s striking: 60.7 percent of large-scale establishments (those with equity more than RM2.5 million) have benefited from incentives for their most recent investment, whereas only 38.5 percent of small- and medium-scale establishments have done so. 4.6 The most frequently used incentives are the reinvestment allowance/accelerated capital allowance (33 percent) and Pioneer Statusfinvestment Allowance (24 percent) (Figure 4.2).48These more frequently used incentives benefit a greater percentage of large establishmentsincomparison to small-scale establishments. 48 Itis important to note, however, that there may be considerablemeasurement error in terms of identifying the correct underlying sample of eligible establishmentsfor each specific incentive. The lists of promoted activities and products for the individual incentives do not matchexactly with available 5-digit ISICcodes given for establishmentsinthe PICS. Many items on the list of eligible products are more detailed than the descriptionsat the five-digit ISIC code level. 56  Figure 4.2: Types of Incentives Percentageot Eliglbleinvesting Establishmentsby Sizethat Benefitedfrom the Followingincentivesfor Their MostRecent Investment 10 All Firms I Large-Scale(Equity>RM2.5mm) I end Medium-Scale(Equityc=RM2.5mm) Small- I contrast, R&D schemes have had little success in helping Malaysian manufacturing establishments benefit from investment. Figure 4.3 shows that very few establishments have benefited from particular R&D schemes: Double Deduction incentive for approved R&D project (5.6 percent), Investment Allowance on capital expenditure for R&D (5.7 percent), exemption of duty/tax on R&D machinery (6.7 percent), capital allowance on R&D plant and machinery (7.7 percent), industrial building allowance for R&D (2.1 percent), and commercialization of technology from public researchinstitute (0.7 percent). Figure 4.3: Firms' Awareness of R&D11ncentives Has Your FirmBenefittedfrom the Following IncentiveSchemes? 55 - 50 - 55 - 50 - e 45 - 40 E B 35 9I30 f 25 Yes No, INever heard NO,Iapplied and was NO.I never applied No, scheme not No, results of about this scheme turned down because process too relevant application are long and complicated pending W Doubledeductionincentivefor approved R&D project 0investmenttax allowanceoncapitalexpenditurefor R&D Exemptionof duty, tax on R&D machinery ECapital allowanceon R&Dplant and machinery ElIndustrialbuildingallowancefor R&D UComrnerciaiizationof technologyfrom public researchinstitute 57  4.8 A possible reason for such low usage of R&D schemes may be the lack of information disseminated to establishments. More than 55 percent of the establishments surveyed have never heard about any of these schemes. A secondary issue i s that about a fourth of all establishments believe the process of applying to any of these schemes is too long and complicated, outweighing any benefit they couldreceive.49 4.9 Although the reported percentage of establishments that use the R&D incentive schemes i s quite small, more than 75 percent of the users believe the benefits to be of at least moderate value, and more than half feel they are of major or critical value (Figure 4.4). This finding, along with the fact that more than half the establishments have never heard of the R&D schemes, suggests that more effort may be required to inform establishments of their options. Once managers realize the potential benefits available from such R&D schemes, a higher uptake i s likely to follow. Figure 4.4: Value of R&DIncentives Doublededuction Investmenttax Exemptionof duty, tax Capital allowanceon Industrialbuilding Commercializationof incentivefor approved allowanceon capital on R&D machinery R&Dplant and allowancefor R&D technologyfrom public R&D project expenditurefor R&D machinery researchinstitute 49Establishments may not have heard of a particular scheme for the good reason that they are ineligible to apply for it. If this is the case, we may have overestimated the number o f establishments that have never heard of a particular scheme, and underestimated the number o f establishments that have benefited from a particular scheme. A more accurate measure and conclusive analysis o f incentive and schemeuserequires more detailed questions and better data from establishments inthis area, possibly from future survey work. 58  4.10 A positive findingis that more establishmentshad benefited from R&D schemes in the last three years prior to the survey (80 establishments) than in the entire period before 2000 (69 establishments), peaking in2001 (Figure 4.5). Figure 4.5: R&DIncentive Use over Time Year that Firm ReceivedBenefit 1972 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Doublededuction incentive for approved R&D project 0Investmenttax allowanceoncapitalexpenditurefor R&D 0Exemptionof duty,tax on R&D machinery Capitalallowanceon R&D plant andmachinery Industrialbuddingallowancefor R&D !l oftechnologyfrompublicresearchinstitute Commercialization 4.11 The use and ratings of incentives among small- and medium-scale establishments are similar to those among large-scale establishments. More than 55 percent of small- and medium-scale establishments have never heard of these schemes, and very few have benefited from them. Figure 4.6 shows that fewer than five percent of small- and medium-scale establishments use schemes under the Industrial Technical Adjustment Fund, while only one percent use the SME Technology Acquisition Program, 1.3 percent use the S M E Modernization and Automation Loan Scheme, and 1.7 percent use the Soft Loan Scheme for Quality Enhancement for Bumiputera small and medium ind~stries.~' Nearly 25 percent of small- and medium-scale establishments decide against applying for these schemes becausethey see the process as too long and complicated. 50However, the high number of non-usersmay again be the result of measurementerror. 59  Figure 4.6: Awareness of Incentives for Small- and Medium-ScaleFirms HasYour Firm Benefitedfromthe Following IncentiveSchemesfor Small- and Medium-Scale Enterprises? 65 Yes No, INeverheard No, Iappliedandwas No, Ineverapplied No, scheme not No, resultsof aboutthis scheme turneddown becauseprocesstoo relevant applicationare longand complicated vending ITAF Consutancy Scheme OlTAF ProaLct DeveloprnenVDesig Scheme ITAF OJalit and ProdLCtivit ImprovementScheme .. .MarketDeveopmet Scneme OSME Techno!oav AcaLis&ton8rwramme 0ITAF.Moderniza1,onandAutomationLoanScheme SME 4.12 Nonetheless, most of the SMEs that have usedthe schemes targeted to small- and medium-scale enterprises rate the benefits as valuable, and more than half feel they are of major or critical value (Figure 4.7). Again, this finding suggests that more effort i s needed to inform SMEs about the options open to them. Similar to the sample of large- scale establishments, many more small- and medium-scale establishments had benefited from R&D schemes in the last three years prior to the survey (79 firms) than the entire periodbefore 2000 (30 firms), peaking in2002 (Figure 4.8). 60  Figure 4.7: Value of Incentives for Small- and Medium-scale Firms How Criticalwas this Small-and Medium-ScaleEnterprisesScheme to Your Firm? 90 1 No Value 0 MinorValue Moderate Value ElMaior Value !ElCritical Value I 80 70 60 EE v) -z550 0) n i40 30 20 10 0 ITAF Consultancy - ITAF Product - ITAF Qualityand - ITAF Market - SMETechnology SMEModernization SOHLoan Scheme Scheme OevelopmenVDesig Productivity DevelopmetScheme Acquisiton andAutomationLoan lor Quality Scheme Improvement Programme Scheme Enhancementlor Scheme BumiputeraSMI Figure 4.8: Use of Incentives by Small- and Medium-scale Firmsover Time Year that Firm ReceivedBenefitfrom Incentive Scheme for Small- and Medium-Scale Enterprises 40 35 30 v)25 GE c 20 d2 5 15 10 5 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 HITAF ConsultancyScheme 0ITAF ProductDevelopmenffDesig Scheme ITAF Qualit and Productivit ImprovementScheme .. p1ITAF Market DevelopmetScheme -- OSME TechnoyogyAcquisitonJro ramme 0SMEModernizationandAutomationLoanScheme ESoft Loan Schemefor QualityEnaancementfor BumiputeraSMI 61  4.13 Descriptive statistics show that large establishments that use incentives andor schemes have significantly higher average investment rates than large-scale establishments that do not use them. However, there i s no such pattern for the sample of small-scale establishments (Le., establishmentswith equity less than RM2.5 million). The investment rate for large establishments that use incentives andor schemes was 0.35 in 2000 and 0.22 in2001, much higher than the investment rates of 0.12 in 2000 and 0.10 in 2001 for those large establishmentsthat didnot use any (Figure 4.9). 4.14 Conversely, small- and medium-scale establishments that benefited from incentives andor schemes averaged a lower investment rate than small- and medium- scale establishments that did not use any in 2000 (0.34 versus 0.40), while in 2001, the opposite was true (0.26 versus 0.22). 0.40 I havebenefitedfromincentiveandlorschemes Firmsthat 0.35 -. 0Firmsthat havenotusedincentivesorschemes 0.30 - 0.25 - 0.20 - 0.15 - 0.10 - 0.05 - I 0.00 7 2000 I 2001 2000 1 2001 Ail Firmswith i>O Ail Large-Scale Firmswith bO All Small- and Medium-Scale FirmswRh i>O B. LINEAR REGRESSION ESTIMATE OFTHE IMPACT OFINCENTIVESON INVESTMENT RATES 4.15 This section presents standard linear regression estimates of the impact of the reinvestment allowance/accelerated capital allowance, as well as regulatory constraints, on investment rates. Among the fiscal incentives, we use the reinvestment allowance/accelerated capital allowance for this analysis because all business establishments qualify for this incentive and hence the PICS dataset provides enough observations to permit reliable conclusions. 4.16 A flexible accelerator specification of the investment function is estimated to determine the factors that explain investment and then i s augmented to examine the relationship between incentive use and the level of investment. The investment rate in periodt i s of the following form: 62  where -i s the investment rate for new machinery and equipment, -i s the ratio of " AV Kt-1 Kt-1 the change in value addedto the capital stock, i s the ratio of profit to the capital stock, i s the ratio of long-term debt to the capital stock, and u, i s an error term. 1-1 The-Kt-1 and AV (a) (E) terms have an expectedpositive impact on the investment t-1 (alt-l rate(p,,p, >0) as bothreflect changes inthe desired capital stock. A positive profit rate effect i s also possible if liquidity constraints are binding. The borrowingterm is assumedto have a negative impact on the investment rate (p3<0),reflectingincreasing costs of borrowing. The sign on the squaredborrowing term ( ~ ) will~ be positive or 1 negative dependingon whether the marginal effect of borrowingon investment is decreasingor increasing. 4.17 The flexible accelerator model is augmented to allow for the size of the firm (measuredby the lagged number of permanent workers), firm age, industry, infrastructure investment (proxied by location), and several types of incentives and regulations. The size of the firm i s included to allow for firm- or size-specific capital costs and indivisibilities in investment. The age of the firm may affect both the decision to invest and the amount of investment undertaken, since longer-established firms may have greater access to lower-cost capital and more experience in acquiring capital. Industry dummies at the two-digit ISIC level are included to allow for industry-specific capital costs. Public sector investment facilitates a better business environment and may attract more private investment. This factor is proxiedby the region in which the firm i s located. Annex 3 Section A presents the descriptive statistics for the variables in the flexible acceleratormodel. 4.18 The regression-basedestimates of the flexible accelerator model are presentedin Table 4.1 for the sample of large establishments (total equity greater than RM2.5 million) and in Table 4.2 for the sample of small- and medium-scale establishments (with equity less than RM2.5 million). Inthese tables, Column 1contains the results of a probit for the decision to invest; Column 2 presents the marginal probabilities for the probit estimates; Column 3 reports results of the equation for the level of investment of investing establishments; Column 4 includes a dummy variable for reinvestment allowance/accelerated 63  capital allowance users; Column 5 includes dummy variables for several regulatory constraints on doing business; and Column 6 reports results for the full model. 4.19 Between the two groups of establishments, there are a number of differences and similarities in the estimated coefficients of the flexible accelerator model variables. For large establishments, the change in value added i s not a significant determinant of the decision to invest, and for large-scale investing establishments it i s only a marginally significant determinant of the investment rate. For small- and medium-scale establishments, the change in value added is a marginally significant determinant of the decision to invest, but among those SMEs that invest, it i s a very significant determinant of the investment rate. Profitability is a significant determinant of both the decision to invest and the amount of investment for large establishments, but it is insignificant for small ones. Long-term debt i s a significant obstacle to investment for small establishments, but not for large ones. The size of establishment, as measured by the number of permanent workers, is a highly significant determinant of the decision to invest for both samples. The age of the establishment i s not a significant determinant. 4.20 Among the large establishments in the sample, use of the reinvestment allowance/accelerated capital allowance, as well as ownership regulation, are significant determinants of the investment rate. The results in Column 6 of Table 4.1 show that among large establishments, those using the reinvestment allowance/accelerated capital allowance have investment rates 9.4 percent higher than those not usingthese incentives. Conversely, among large establishments, those that indicate ownership regulation i s the biggest obstacle to doing business in Malaysia have on average a 20.3 percent lower investmentrate than those not ratingownership regulationinthis way. 64  Table 4.1:InvestmentEquations: LargeEstablishments Table4.1. InvestmentEquations[Sample:Large.de establishments totalequitygreater thanRM2.5 million]. - - Probit Marginal - - OLS OLS _. OLS - OLS =1ifinvestment; probabilitiesfor Investment hVeStment Investment t/ Investment DependentVariable t/ t/ t / =O if noinvestment estimatesin[l] Capital Capital Capital Capital InvestmentM InvestmenbO InvestmenW) InvestmenW) Reinvestmentallowanceand/or 0.087 ' 0.094 * acceleratedcapitalallowanceuser (0.047) (0.049) Ownership regulationis the biggest -0.205*" -0.203 *** obstacle (0.061) (0.057) Labor regulationsis the biggest -0.035 -0.063 obstacle (0.068) (0.076) Importregulationis the biggest -0.115 -0.13 obstacle (0.077) (0.080) Bureaucraticburdenisthe biggest 0.1250 0.1280 obstacle (0.101) (0.097) (value Added 0.222 0.055 0.046* 0.048 ** 0.039 ' 0.041 ' Capital [., - ValueAdded )/ (0.171) (0.039) (0.024) (0.024) (0.022) (0.023) (ProfiUCapital) cI 0.200 **' 0.050'** 0,023* 0.023 ** 0.018 ' 0.018 (0.072) (0.017) (0.012) (0.012) (0.011) (0.010) (Long-termDebUCapital) [., -0.331 -0.082 -0.034 -0.055 -0.046 -0.065 (0.359) (0.090) (0.055) (0.057) (0.058) (0.060) (Long-termDebVCapital) 0,080 0.020 0.005 0.011 0.005 0.011 (0.096) (0.024) (0.012) (0.013) (0.013) (0.013) 0.468 *** 0.116 "' 0.032 0.015 0.028 0.012 (0.160) (0.036) (0.019) (0.021) (0.020) (0.022) FirmAge 0.005 0.001 0.001 0.001 0.002 0.003 (0.018) (0.005) (0.003) (0.003) (0.003) (0.003) Year 2000 -0.244 -0.061 -0.015 -0.016 -0.014 -0.014 (0.196) (0.050) (0.045) (0.045) (0.047) (0.047) Constant -1.363 0.020 0.065 0.024 0.059 (0.967) (0.145) (0.149) (0.154) (0.158) IndustryDummies Yes Yes Yes Yes Yes Yes RegionDummies Yes Yes Yes Yes Yes Yes Observations 168 168 133 133 133 133 Number of establishments 109 109 89 89 89 89 Psuedo/Adjusted R 0.172 0.172 -0.010 0.006 0.009 0.029 Nom :Standarderrors adjustedfor clusteringonestablishrrentaredenotedinparentheses. Vdue ad&d i s calculatedas totalsales minusdirect materialcost. Capitalis given by net boak value of new machineryandequipment. Size isequal to the total numberof permaDent employees. Long-term debt isequal to domstic c m u c y denominatedlong-tenn liabilities. There are 8 industrydummiesdefinedat t k two-digit ISIClevel and5 regiondummies. * deirotessignificance at 10%level: '' denotes significanceat 5% level: "'denotes significanceat 1% level. 4.21 Among the sample of small- and medium-scale establishments, there are no significant differences between reinvestment allowance/accelerated capital allowance users and non-users (Column 6 of Table 4.2). Conversely, those that indicate import regulation i s the biggest obstacle to doing business in Malaysia have on average a 13.1 percent lower investment rate than their counterparts who do not rate importregulationin this way. 65  Table4.2: InvestmentEquations:Small- andMedium-scaleEstablishments Table4.2. InvestmentEquations[Sample:Small-andmedium-scaleestablishments-total equity less than RM2.5 million]. Probit Marginal - - - - OLS OLS OLS OLS =1ifinvestment; probabilitiesfor Investment t/Investment 1/ Investment l/ Investment l/ DependentVariable =O if noinvestment estimatesin[l] Capital t.l Capital l.l Capital l.l Capital l.l InvestmenMl Investment9 InvestmenMl InvestmentM [I1 PI [SI [41 PI [a Reinvestmentallowanceandor 0.008 0.01s acceleratedcapitalallowanceuser (0.040) (0.041) Ownershipregulationis the biggest 0.0380 0.0380 obstacle (0.072) (0.074) Labor regulationsis the biggest -0.07 -0.068 obstacle (0.058) (0.059) Importregulationis the biggest -0.129 *' -0.131 obstacle (0.073) (0.076) Bureaucraticburdenis the biggest -0.0640 -0.0700 obstacle (0.068) (0.072) (Value Added 0.135' 0.038 ' 0.055 **' 0.055 "* 0.052*** 0.051'* Capital ,., ,-Value Added ,-,) / (0.072) (0.020) (0:ozo) (0.020) (0.020) (0.020) (ProftKapital) ,-, 0.035 0.010 O.Oo0 O.Oo0 -0.002 -0.002 (0.043) (0.012) (0.008) (0.008) (0.008) (0.008) (Long-termDebVCapital) ,-, -0.819 *'* -0.230 *'* -0.006 -0.005 -0.003 -0.003 (0.255) (0.071) (0.059) (0.059) (0.062) (0.062) (Long-termDebVCapital) ',., 0.193 *** 0.054 *" (0.001) (0.002) (0.002) (0.002) (0.069) (0.019) (0.015) (0.015) (0.015) (0.015) In(Size) ,., 0.244 "* 0.069 '** 0.019 0.017 0.022 * 0.020 (0.078) (0.023) (0.013) (0.014) (0.013) (0.014) FirmAge ,., -0.012 -0.003 -0.003 -0.003 -0.003 -0.003 (0.012) (0.003) (0.002) (0.002) (0.002) (0.002) Year ZOO0 -0.002 O.Oo0 0.070** 0.070 ** 0.074 *' 0.074 ** (0.123) (0.035) (0.031) (0.031) (0.032) (0.032) Constant 0.756 0.147 ** 0.149 ** 0.156 ** 0.160 ** (0.497) (0.071) (0.071) (0.072) (0.071) Industry Dummies Yes Yes Yes Yes Yes Yes Reion Dummies Yes Yes Yes Yes Yes Yes Observations 403 403 297 297 297 297 Numberof establishments 241 241 189 189 189 189 PsueddAdjusted R 0.156 0.156 0.094 0.091 0.091 0.088 Notes : Standarderrorsadjustedfor clusteringonestablishment aredenoted inparentheses. Valueadded is calculatedas totalsales ininus directmaterialcost. Capitalis givenby net bookvalue of newmachineryandequipment. Size is equalto the total numberof permanent employees. Long-termdebt is equalto domesticcurrency denominatedlong-termliabilities. Thereare8 imlushydummiesdefnedat the two-digit ISICleveland5 regiondummies. * denotes significance at 10%level; *' denotes significance at 5 6 level; *"denotes significanceat 1% level. C. CAUSAL INFERENCE SCOREMATCHING THROUGH PROPENSITY 4.22 This section briefly describes the propensity score matching method and then presents the estimated effect of the reinvestment allowance/accelerated capital allowance on the investment rates of firms. The EvaluationProblem 4.23 Does the reinvestment allowance/accelerated capital allowance cause firms to increase their investment rate? Or are firms that already have high investment rates more likely to use this incentive than firms that have lower investment rates? To answer these questions, we require information on the investment rates for users in the period before 66  and after usingthe reinvestment allowance/acceleratedcapital allowance, so that the rates can be compared. Unfortunately, such data are unavailable in the PICS, and traditional linear regression techniques are unable to determine the causal relationship between the use of this incentive and the investment rates of firms. 4.24 Propensity score matching methods can be used to estimate the hypothetical average investment rate for users had they not taken advantage of the incentive. This rate .is estimated using a control group of non-users. Note that the average investment rate of non-users i s not a suitable variable for this calculation, because users and non-users may differ in characteristics that affect their investment rates. That is, users are not a random sample of the population; they may be using the reinvestment allowance/accelerated capital allowance because they have characteristics that also influence their investment level. For this reason, we match users with a control group of non-users similar in characteristics that influence both the probability of incentive use and the investment rate.51 TheMatchingMethod 4.25 For the sample of large-scale establishments, users are matched to non-users that have similar probabilities of using the reinvestment allowance/accelerated capital allowance (i.e. propensity scores). The probabilities are estimated using a probit model that includes factors that are expected to affect the use of the reinvestment allowance/accelerated capital allowance as well as the investment rate. The specification for the propensity scores regresses the firm's use of the reinvestment allowance/accelerated capital allowance on the following variables: (a) the ratio of the change in value added to the capital stock; (b) the ratio of profit to the capital stock; (c) the ratio of long-term debt to the capital stock, size, firm age, industry, region; (d) dummy variables indicating whether the firm benefited from Pioneer Statushvestment Tax Allowance, incentives for high technology companies, incentives for strategic projects, incentives to strengthen industrial linkages, incentives for manufacture of machinery and equipment, tax exemption on the value of increased exports, and any R&D scheme; (e) dummy variables indicating regulatory constraints (ownership regulations, labor regulations, import regulations, and bureaucratic burden); (f) a dummy variable indicating that the firm's most important way of acquiring technology i s through the purchase of new machinery and equipment; and (g) capacity utilization in the previous year and its square. This specification passes the balancing test as described in DehejiaandWahba (1999,2002) andSmith andTodd (2004).52 4.26 The Gaussian kernel estimator i s used to estimate the hypothetical non-using investment rate for users-that is, the rate at which users of the incentives would be investing if they were not using the incentives. Several methods can be used for matching; they differ in the specific weights assigned to each comparison group 51Basic statistical matching estimators solve the problemof selection bias if and only if the selectionbias is completely determinedby selection on observables. Thus, if there is selection on unobservables,matching methodsdo not identify the causaleffect of an endogenous treatment. See Annex 3 SectionB for a detailed explanationof the propensity score matching methodology usedinthis chapter. 52See Annex 3 SectionB for adescription of the balancing test usedinthe analysis. 67  observation. All matching estimators construct an estimate of the expected unobserved counterfactual for each treated observation (i.e., user) by taking a weighted average of the outcomes of the untreated observations (Le., non-users). What differs among the various matching estimators is the specific form of the weights. The Gaussiankernel estimator i s used because, as Figure 4.10 shows, comparison observations are many but asymmetrically distributed. Kernel matching i s helpful inthis case because it will use the additional data where they exist, but will not accept bad matches where data are lacking.53 Figure 4.10: Distribution of the PropensityScores QTreated:On support Treated:Off support c]Untreated 0 . o ~ ' .A75.225.275.325,375.425.475,525,575.625.675.725,775.825.a75.925.975 MatchingEstimates of the Impact of Reinvestment Allowance/Accelerated Capital Allowance The estimated effect of the reinvestment all'owance/acceleratedcapital allowance on users i s a 14 percent higher investment rate on average, and it i s statistically significant. As seen inTable 4.3. for the propensity score matchingestimate, the bootstrap standard error based on 2000 replications i s 0.050. Each bootstrap includes re-estimation of the propensity scores used in the matching on the bootstrap sample. This estimate i s higher than the ordinary least squares estimate of 9.4 percent reproduced in the third row of Table 4.3. 53 Inimplementation, a fixed, global bandwidth parameter using Silverman's (1986) rule-of-thumb method i s selected. 68  Table 4.3: Propensity Score Estimates of the Effect of Pioneer Status and/or Investment Tax Allowance Use Difference betweenthe investmentrate of users and non-users Unmatched 0.084 Average Treatment Effect on the Treated 0.140*** (0.050) OLS Estimate 0.094. (0.049) Notes:Standarderrors are inparentheses. For the propensity score matchingestimate, bootstrap standard errors are basedon 2000 replications. For the OLS estimate, standarderrors are adjusted for clusteringon establishment. *** denotes significance at 10%level; ** denotes significanceat 5% level; ***denotes significance at 1% level. D. CONCLUSIONS 4.27 While the evidence presented in this chapter i s limited, for most types of incentives it provides some support for the effectiveness of the reinvestment allowance/accelerated capital allowance in increasing the investment rate of large-scale establishments. Standard ordinary least squares (OLS) estimates show that large-scale establishments that have benefited from the reinvestment allowance/accelerated capital allowance have investment rates 9.4 percent higher than large-scale establishments that have not benefited from the reinvestment allowance/accelerated capital allowance. The propensity score matching estimate of the reinvestment allowance/accelerated capital allowance effect on the investment rate i s 14 percent, which i s consistent with the OLS estimate. These results suggest that the reinvestment allowance/accelerated capital allowance has a causal and statistically significant impact on the investment rate of large establishments. Findingsfor other types of incentives and schemes are inconclusive, due to the insufficientnumbers of users. 4.28 Regulatory constraints-namely ownership regulations for large establishments and import regulations for SMEs-are found to have strong negative effects on the investment rate. Among large establishments, those that indicate ownership regulation i s the biggest obstacle to doing business in Malaysia have on average a 20.3 percent lower investment rate than those that do not record this problem. Among small- and medium- scale establishments, those that indicate import regulation i s the biggest obstacle to doing business inMalaysia have on average a 13.1 percent lower investment rate than those that do not record this problem. 69  5. THE COSTSOFFISCALINCENTIVES FOR INVESTMENT 5.1 As documented earlier in this report, tax incentives for capital do encourage investment. However, the more complex the system of incentives the greater the costs of administering and monitoring them, and of complying with their requirements. Chapter 4 indicated that only 60.7 percent of large firms and only 38.5 percent of SMEs inMalaysia take advantage of the incentive schemes. The primary reasons for this lack of use, particularly among small and medium enterprises, appear to be that firms lack information about the incentive programs and that the process of applying for the programs i s long and complicated. 5.2 Another important cost of tax incentives, and the focus of this chapter, i s the tax revenue that could have otherwise been raised if the incentives were not in place. This forgone tax revenue could have been usedto reduce other taxes, such as personal income taxes, consumption taxes, or, most importantly, taxes on capital that does not receive incentives. 5.3 Section A applies the concept of the marginal cost of public funds to investigate the sensitivity of income in the private sector to changes in taxation. It then derives estimates of the cost to the Malaysian economy of various corporate tax incentives. The results raise questions about the economic justification of some of the current incentives. Section B extends the analytical framework to investigate the level of subsidy that is appropriate for R&D. Section C concludes. A. ACCOUNTING FORTHE COSTOFFORGONE REVENUES TAX 5.4 One way of accounting for the cost of forgone tax revenues is to employ a concept called the marginal cost of public funds (MCF). The MCFmeasuresthe total cost to the private sector associated with raising one more ringgit in tax revenue from a particular tax by increasing the effective tax rate by a small amount. Conceptually the MCF of atax leviedon tax baseX is: MCF,= Change in IncomeKhange in Revenue 5.5 The numerator measures the reduction in national income associated with a small increase in the tax on activity X, while the denominator measures the change in tax revenue. The reduction in capital income will generally exceed the revenue generated from the tax increase, which means that the MCF i s greater than one, because of the disincentive effects associatedwith the tax increase. 5.6 As such, increasing the tax on activity X does more than merely transfer one ringgit from activity X to the government, it discourages the activity being taxed, which 70  gives rise to efficiency costs over and above the additional ringgit in tax revenue raised, This can perhaps be seen more clearly by writing the MCFalternatively as: MCF,=l+MEC,, (1) where MEC, i s the marginal efficiency cost of increasing the tax on activity X by a very small amount. The MEB is the incremental efficiency cost, measured in terms of the reduction in national income, arising from increasing the tax on activity X. These costs arise because scarce capital i s reallocated as a result of tax distortions from more productive to less productive uses. 5.7 The efficiency cost of capital taxes i s thus a measure of the value of the output forgone when taxes divert capital from its best uses. This formulation emphasizes that the marginal cost to the private sector associated with raising an additional ringgit in tax revenue i s composed of two parts - the one ringgit in tax revenue raised plus the marginal efficiency cost associated with the higher tax rate. It i s because of the presence of MEC associated with a tax increase that the MCF exceeds the additional tax revenue generated. 5.8 The specific formula for the MCF depends upon the type of tax being levied. For taxes imposed on corporate capital - such as the corporate income tax, property taxes, etc. -the MCFformula is the following: where: MCFKi s the marginal cost of public funds for taxes leviedon the returnto capital; t is the marginal incometax rate; z i s the marginal effective tax rate (METR) on capital; K i s the change in the capital stock when an additional ringgit of tax i s levied on the return to capital. It measures the tax sensitivity of investment activity, and is a negative number. Insymbols, K=dW(Kdz), where K i s the stock of capital; ris the after-tax rateof returnrequiredon aunitof capital. 5.9 The numerator of expression (2) i s the harm to the private sector, as measured by the reduction in national income, that results from a small increase in the effective tax rate on capital; it is normalized to unity. The denominator i s the change in tax revenue resultingfrom a small increaseinthe effectivetax rate on capital. 5.10 Expressing the MCFK in this way, with the reduction in national income (the numerator) normalized to unity, differs slightly from the intuitive explanation of the MCF provided above. With the loss in private sector income normalized to unity, the increase in tax revenue associated with a small tax increase in the METR on capital will be less than one, because of the MEB associated with the tax increase. The idea i s the same, 71  however, and, as above, the MCFK will be greater than one. In the denominator of the formula, the l+mcomponent is the direct revenue increase associated with the increase inthe effective tax rate on capital, takinginto account the shrinkage of the capital stock, and therefore of the effective capital tax base, that results from the increase in the effective tax rate (via the term K, which recall i s negative). The term t(r+T)K represents the indirect revenue loss arisingfrom the reduction intotal output causedby the decrease inthe capital stock. The term (r+z) is the gross-of-tax rate of return on a unit of capital. Multiplying this by the reduction in capital due to the increaseinthe effective tax rate, K, gives the reduction in output attributable to the tax increase. Multiplying this in turn by the marginal income tax rate, t, gives the indirect reduction intax revenue. Sensitivity of Investment to Taxation 5.11 From the above formula we can see that the MCFK i s increasing in two key parameters - the tax sensitivity parameter (K) and the METR on capital (2).That is, the more sensitive is capital investment to changes in the tax rate - the greater the incentive effects - the higher will be the costs of generating one more ringgit in tax revenue by raising the METR on capital. Similarly, the higher i s the METR to begin with, the higher will be the marginal costs associated with raising the rate slightly. Note that if capital investment i s completely insensitive to the tax rate (Le., K=O), the MCFK=~,because increasing the tax rate on capital has no costs over and above the tax revenue raised. This i s because when the capital tax base i s completely insensitive to the tax system, there are no incentive effects and therefore no efficiency costs. In this case, a tax on capital i s a lumpsumtax. 5.12 Of course, as stressed throughout this report, that i s not the case in practice: capital investment is sensitive to taxes, and to subsidies, and taxes imposedon capital are not lump sum. There is, however, some debate on how sensitive it is. Although early empirical studies of the impact of business taxes on investment (based on aggregate data) tended to find rather small effects, more recent evidence (based on firm-level data and new empirical techniques) suggests that corporate taxes have a significant impact on investment. For example, a recent study by Chrinko, Fazzari, and Meyer (1999) suggests that in the United States a one percent increase in the tax-adjusted user cost of capital - a concept closely related to the marginaleffective tax rate discussed above -leads to a 0.25 percent reduction in the capital stock. Similarly, in a multi-country study, Cummins, Hasset, and Hubbard (1998) found that a 1percent increase in the cost of capital can lead to a decrease in investment in machinery and equipment by as much as 1percent. High rates of tax on capital do, it seems, tend to depress investment. 5.13 Evidence also supports the proposition that business taxes affect the location decisions of multinational firms. For example, Devereux and Griffith (1998) found that effective tax rates were an important determinant in the location of foreign direct investment by U S multinationals in Europe. A similar study of U S states found that differences in tax rates across states helped explain differences in manufacturing firm start-ups across states. In particular, after controlling for state- and industry-specific effects, a high state marginal effective tax rate reduced the number of new firm "births" inmorethan half of the industries examined(Papke 1991). 72  5.14 The evidence just cited suggests that taxing capital at the corporate level distorts investment decisions and lowers the capital stock. It i s based on OECD countries. There have been few studies of developing countries although, as reported in Table 1.1, evidence reported in Shah (1995) for developing countries suggests that investment can be quite sensitive to certain types of fiscal incentives. Shah reports additional investment per dollar of forgone incentives in developing countries that ranges from around .25 dollars to 2.4 dollars inthe long run, dependingupon the nature of the tax incentive. Most effective in this regard are accelerated depreciation, investment allowances, and tax credits. 5.15 The empirical analysis in Chapter 4 above also provides evidence that investment in Malaysia is sensitive to tax incentives. For example, linear regression results from a flexible accelerator model suggest that the investment rate of large corporations who utilize the reinvestment allowance/acceleratedcapital cost allowance i s 9.4 percent higher than corporations who do not take advantage of these incentives. While this need not mean that all of the difference in investment rates between the two groups of companies i s due to the use of the incentive, the result i s telling. 5.16 One of the problems with the flexible accelerator model of investment i s that it does not explicitly incorporate the cost of capital (related to the marginal effective tax rate on capital) in the empirical specification. It was not possible to do this in our case, for lack of firm-level data on the cost of capital (a persistent problem in studies of this type). 5.17 However, we can do a "back of the envelope" calculation based on the empirical results to see if the tax incentive effects suggestedby the model are consistent with those in other studies. To do this, consider that, as mentioned previously, the percentage increase in investment that might be attributed to the use of the Investment Allowance/accelerated depreciation incentive i s 9.5 percent. From the METR calculations in Chapter 2, the METR for large corporations under the general (i.e. no incentives) tax systemi s 19.7 percent for manufacturing companies, andthe METR under the Investment Allowance/accelerated depreciation incentive i s 16.4 percent (Table 2.3). As such, the percentage reduction in the METR due to this incentive is 16.75 percent. This suggestsan elasticity of investment with respect to the METRof around 0.56, which i s in the range of elasticity estimates from other studies discussedabove. 5.18 Table 5.1 reports the MCF of taxes imposed on the return to capital for the manufacturing sector inMalaysia under the general (no incentives) tax regime and under various incentive schemes. Each regime i s associated with a METR as reported in Table 2.3. For each regime, the MCF of capital taxes i s reported under a range of assumptions regarding the tax sensitivity parameter, IC, which ranges from .25 to 1.5. As suggested above, the estimates inthe .5 to 1.0columns might be consideredto be the base cases. 5.19 The way to interpret the MCF calculations inTable 5.1 i s as follows. Focusing on the third column with a tax sensitivity parameter of 1.0, the MCF of 1.429 under the general tax regime means that raising one more ringgit in tax revenue by increasing the tax on capital in the general tax regime by a small amount results in a reduction in 73  income from the private sector of 1.429 ringgit. Of this, 1ringgiti s due to the 1ringgit in higher taxes collected and .429 ringgit i s due to the marginal efficiency costs associated with the higher tax rate. Notice that the MCF increases with the tax sensitivity of investment. So, with the higher tax sensitivity parameter of 1.5, the MCF in the general tax case i s 1.819, of which 0.819 ringgit i s attributable to the increaseinefficiency costs. 5.20 Notice also that as the METR decreases due to the presence of various incentives, the M C F also decreases. Consider, for example, the acceleratedcapital cost allowance for the general case. This lowers the METR on capital from 19.7 percent to 16.4 percent. The reduction in the METR due to this incentive lowers the MCF of capital taxes from 1.429 to 1.294. Table 5.1: MarginalCost of PublicFundsfor Taxeson the Return to Capital inMalaysian Manufacturing InvestmentSensitivityParameter METR on Capital for Various Incentives -.25 -.5 -1.0 -1.5 (%) General (19.7) 1.081 1.176 1.429 1.819 SME (16.3) 1.069 1.147 1.347 1.629 Basic Pioneer Status 1.061 1.130 1.298 1.525 Opportunity Costs of Tax Incentives 5.21 This difference in the MCF associated with investments taxed under the general tax regime and that associated with investments taxed under an incentive regime allows us to calculate the opportunity cost of tax incentives. The MCF calculation cuts both ways: it tells us the economic cost to the private sector of raising a tax by one ringgit, and it also tells us the economic benefit to the private sector of lowering a tax by one ringgit. The difference between the MCF associated with the general, no incentive, regime and the MCF associated with the incentive regime suggests that the METR on capital under the general regime could be reduced slightly so as to lose one ringgit in tax revenue, 74  generating a private benefit of 1.429 ringgit, and that the METR on capital under the incentive regime could be increased slightly so as to gain one ringgit in tax revenue, generating a private sector cost of 1.294 ringgit. This small increase in the incentive regime's METR and decrease inthe general regime's METR i s revenue neutral - the one ringgit lost due to the reduction in the METR under the general regime i s made up by the one ringgit gained by reducing the incentive - but it generates a net gain to the private sector of ,135 ringgit (=1.429-1.294). 5.22 Thus the cost of offering a one ringgit tax incentive-in the form of accelerated depreciation deductions that lower the METR from 19.7 percent to 16.4 percent-is that the overall METR under the general system is higher than it otherwise would be, and, under the assumptions underlying the MCF calculations, this cost exceeds the revenue generatedfrom this higher tax rate. 5.23 The above example i s illustrative, intended to highlight the opportunity cost of tax incentives. In this case, reducing the tax incentive in favor of a more general decline in taxes would enhance efficiency. As with the METR calculations that underlie the MCF calculations, several simplifying assumptions were made. One that i s important to consider i s that the tax sensitivity parameter i s the same for both types of investments - those that take place under the general regime and those that take place under the incentive regime. It may be that investments eligible for incentives are more sensitive to taxes than investments that are not eligible for incentive^.^^ 5.24 Of course, we are not able to ascertain whether this is or is not the case. But the point remains that every ringgit in investment incentives offered under an incentive scheme must be offset by a ringgit in taxes on other activities, and that the full cost of that extra ringgit in taxes exceeds one ringgit. The MCF concept stresses that, in order to justify a tax incentive that lowers the METR below the general rate for a certain type of investment, the MCF for that investment must be higher than for investments made under the general regime. For this to be the case, the tax sensitivity of the targeted investments mustbe higher than investments taxed underthe general, no incentives, regime. 5.25 To illustrate this, Table 5.2 presents the percentageincrease in the tax sensitivity of investments that take place under each of the incentives, relative to the general base case of no incentives, that would be required to justify the current lower METRs on efficiency grounds (assuming a base case sensitivity for investments under the general regime of -1.0). It shows, for example, that the 13.9 percent METR for investments eligible for accelerateddepreciation isjustifiable from an efficiency perspective if the tax sensitivity of investments eligible for this incentive i s 32 percent higher than for investments subject to the general, no incentive, regime (i.e., the tax sensitivity parameter 54For example, perhaps the appropriate tax sensitivity parameterfor investments under the incentive regime is 1.5 rather than the 1.O under the generalregime. Inthis case the MCF for investments madeunder the incentive regime (accelerateddepreciation) is, from the table, 1.517 compared to the 1.429 for investmentstaxed under the generalregime, suggestingthat arevenue-neutralefficiency-enhancing change intax rates would be to lower the METRunder the incentiveregime evenfurther. 75  i s -1.32 rather than -LO)? As can be seen from Table 5.2, investments under some incentive regimes must be considerably more sensitive to taxes in order to justify the current low METRs on efficiency grounds. As another example, inorder to justify the 6.8 percent METR on investments that take place under the enhancedInvestment Allowance, these investments must be more than twice as sensitive to taxation (220 percent ) as are investments that are subject to the general (no incentives) regime (and thus face a METR of 19.7 percent). Intuitively, such large variations in the sensitivity of firms seem unlikely. Table 5.2: Percentage Increase inTax Sensitivity Relative to the General Base Case Required to Justify the Lower METR B. OPTIMALTAXTREATMENT RESEARCHAND DEVELOPMENT OF 5.26 One of the simplifying assumptions underlying Table 5.1 i s that there are no externalities or spillovers associated with the capital that i s being taxed. Clearly this i s not the case for some types of capital investment, including R&D. As discussed in Chapter 3, a convincing case for subsidizing R&D can be made on the grounds that the social rate of return to R&D exceeds the private rate of return due to the presence of spillovers. 5.27 The MCF formula for taxes on capital can be adjusted to reflect the presence of positive spillovers associated with investments in R&D. The augmented formula i s the following: MCFRD={l-(l-t)rex}/{l+zx+t(r,+ r+z)x}, where the notation i s as above, with now the METR on R&D, and with the addition that T reis the excess of the social rate of return over the private rate of return r. (Thus, reis a measure of the positive spillovers from R&D.) As can be seen, this expression collapses to the previous expression when spillovers (re)are equal to zero. 55 Inother words, with a tax sensitivity of -1.32 the MCF for investments with a METR of 13.9 percent is the same (1.429) as the MCF for generic investments with a METR of 19.7 percent and which have a tax sensitivity of -1.O. 76  5.28 How much subsidy i s appropriate i s a difficult and contentious question, whose answer depends critically on three variables that are, to varying degrees, somewhat difficult to measure: the size of the spillover, the sensitivity of R&D to tax incentives, and the opportunity costs of the funds usedto finance the tax subsidy to R&D. 5.29 The MCFframework we have developed here can help us at least think about this issue. To see how, we can use the above formula for the MCFRDto solve for the optimal METR on R&D, z, by equating MCFRDto the MCF from taxing an alternative tax base that i s available to the government. Setting MCFRD=MCFand solving for z yields, z=[1-(l-t)reic]/[rcMCF(2+t)]-[1+tic(r+re)]/[ic(1+t)] (1) Setting this equal to zero and solving for K gives a boundary condition relating the tax sensitivity parameter (K) to the size of the spillover (re)for a given MCF. K=( 1-MCF)/[t(r+r,)MCF+( 1-t)re] (2) 5.30 This allows us to investigate the combinations of these three parameters that are required in order to justify a subsidy for R&D (i.e., a negative METR on R&D capital). As indicated above, given the uncertainty over the values of these parameters, the best we can do i s consider a range of possibilities, and the boundary formulation i s very convenient inthis regard. 5.31 Several empirical studies attempt to estimate the size of the spillover, but there is no consensus on how large it is. Nadiri (1993), a pioneer in this area, concludes that "the indirect and social rates of return [on R&D] often vary from 20 percent to 100 percent with an average close to 50 percent" (page 35). As such, perhaps the best one can do i s to consider a broadrange of estimates. 5.32 There i s somewhat more agreement on the sensitivity of R&D to tax incentives, as discussed in Chapter 4. The general view i s that the long-run elasticity of R&D capital with respect to the user cost of R&D i s in the neighborhood of 1, so that a ringgit worth of subsidy to R&D generatesroughly a ringgit worth of R&D. 5.33 Inorder to subsidize R&D, taxes must be raised on other activities. As such, the opportunity cost of funds used to finance the R&D tax subsidy i s the MCF of taxation. What MCF measurement to use i s a matter ofjudgment. Here we use the MCFassociated with the general income tax on large corporations under two assumptions regarding the sensitivity of capital with respect to taxes. 5.34 Figure 5.1 shows the combinations of the social spillover and tax sensitivities that define the R&D subsidy boundary for Malaysia in the case where the MCF i s 1.429. Any combination of spillover and tax sensitivity above the boundary requires a tax subsidy (i.e. a negative METR) on R&D, while any combination below the boundary requires R&D to be taxed (Le. to be subject to a positive METR). As discussed above, most 77  studies estimate the tax sensitivity of R&D to be in the neighborhood of 1.0. Inthis case, and given the MCF of 1.429, the social spillover must be above 35 percent in order to justify a tax subsidy for R&D. Figure5.1: Optimal Tax Treatment of R&DwhenMarginalCost of Public Funds4.429 2.5 g 2.0 a x- .-05 .-.- c 1.5 r CI 8X 1.0 0.5 0.0 0.0 0.2 0.4 0.6 0.8 1.o 1.2 1.4 1.6 Spillover 5.35 Figure 5.2 illustrates the same thing for a MCF of 1.819. This means that the opportunity cost of the R&D subsidy interms of higher taxes on other activities i s larger. As one would expect, for a tax sensitivity of unity the social spillover required tojustify a tax subsidy for R&D is higher inthis case, at about 60 percent. Figure 5.2: OptimalTax Treatment of R&Dwhen Marginal Cost of Public Funds=1.819 Optimal Tax Treatment of R&D When MCFz1.819 4.0 - _ -- _ _ _ _. 0.04 I 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 SpiHover 78  C. CONCLUSIONS 5.36 The marginal cost of public funds (MCF) measures the total cost of raising one more ringgit in tax revenue by taxing a particular activity. The MCF for taxes on capital inMalaysia differs widely depending upon investors' eligibility for various incentives - inour basecase the MCFrangesfrom 1.429 for investments not eligible for incentives to 1.160 for projects eligible for the Investment Allowance. The reason the MCF i s greater than one i s that aside from the additional ringgit in tax revenue generated, raising a tax rate increases the efficiency costs of the tax system by distorting economic activity. 5.37 Economic efficiency requires the MCF on all activities, in this case investment in different types of capital, to be the same. The fact that they are not, due to the plethora of incentives offered in Malaysia, suggests that there i s scope to improve the efficiency of the tax system, by raising the effective tax rate on capital for investment projects with a low MCF and loweringthe effective tax rate on capital for investments with ahighMCF. 79  6. FINDINGSAND RECOMMENDATIONS A. FINDINGS Current Incentives 6.1 We find three important problems with Malaysia's current investment incentive system. First, the system i s complicated and inefficient. An investment project in manufacturing industry may be entitled to one of nine different tax incentives depending on the nature of product, the location, the source of funding (i.e. initial investment vs. reinvestment from after tax income), the nationality of ownership, the size of firm, the stage of business life, and the preferenceof the investor. Second, since the Malaysian tax system provides loss carry-forward and accelerated depreciation allowances that are available to regular business taxpayers, tax holidays become less attractive to serious capital investors and leave room for tax abuse. Third, tedious selective conditions on eligibility for various incentives militate against market efficiency and complicate tax administration. 6.2 The marginal effective tax rate (METR) calculations in Chapter 2 show that the METR for large corporations under the general (no incentives) tax system is 19.7 percent for manufacturing companies, and that the METR under the Investment Allowance/Accelerated Capital Allowance incentive i s 16.4 percent (Table 2.3). Hence the percentage reduction in the METR due to this incentive is 16.75 percent. This suggests an elasticity of investment with respect to the METR of around 0.56, which i s in the range of elasticity estimates from other studies in the literature. However, METR calculations also suggest that Malaysia's Pioneer Status may be a disincentive - corroborating the international evidence presented in Chapter 1. The Pioneer Status incentive raises the lifetime tax burden for firms that invest inlong-lived assets and are in a loss position in their formative years, but it provides a positive incentive for firms that invest in short-lived assets and are profitable from the very start. 6.3 B y comparison with those of China, Indonesia, Singapore, and Thailand, Malaysia's incentive regime i s complex and difficult to understand. It provides an invitation to unfair outcomes at best and corruption at worst, and for some types of business i s excessively generous. The incentives offered under the Promotion of Investment Act, in particular the whole or partial tax holidays for Pioneer Status investments, are particularly problematic. Where tax holidays are combined with the extremely generous and deferrable write-offs, Malaysia's system not only increases the after-tax profits of profitable companies but in many cases can transform an investment that i s not profitable before taxes into one that i s profitable after taxes. For most companies that get these incentives, the corporate income tax (CIT) i s effectively at or below zero. 6.4 We recommend a rationalization of the incentive system in Malaysia, and in particular the incentives offered under the Promotion of Investment Act, with a view to 80  phasing out ineffective incentives such as the Pioneer Status scheme. Measures such as a lower CIT rate, more broadly available accelerated write-offs (but not as generous as those that are now available to a limited set of companies), investment allowances, and/or investment tax credits, are more effective investment incentives. They can be built into a . system that i s simpler, more transparent, and automatic, and thus easier to enforce. International evidence suggests that they are generally more effective than tax holiday type incentives. This basic approach meets the many, and sometimes contradictory, objectives of a business tax system. Moreover, it has worked well for a number of successful countries, including Chile, Hong Kong, Hungary, and Mauritius, and more recently has been adopted by, for example, Estonia, Latvia, Ireland, Romania, and Poland. 6.5 Some consideration should be given to completely removing fiscal incentives from the Promotion of Investment Act, and consolidating a simplified set of incentives - including accelerated depreciation, investment allowances, and tax credits - into the IncomeTax Act. 6.6 Our results in Chapter 5 suggest that taxes are too high on some types of capital investments (those not receiving incentives) and too low on others, and that a reduction or elimination of some of the tax incentives that generateextraordinarily low effective tax rates on some types of investment, and a more general reduction inthe corporate tax rate, would raise efficiency and stimulate overall economic activity. Incentives for Researchand Development 6.7 Fromthe East Asian experience, there appears to be no strong correlation between the volume of tax incentives and either R&D as a share of GDP or the absolute amount of private sector R&D expenditure. The countries that offer the highest tax incentives - Thailand and Malaysia - do the least research and development. The countries that appear to do the most research and development - Taiwan (China) and Korea - are least generousintax terms, but offer non-tax incentives. 6.8 Malaysia has at least 22 grant, loan, or equity-based schemes operating at the firm, institutional, or individual level. Though designed to support almost every stage of the corporate lifecycle and the assumed process from concept to commercial product, these schemes have clearly not raised research and development investment by the private sector. Programs for the intensification of research in priority areas, venture capital funds like MAVCAP, a number of funds geared toward supporting SMEs, skill development grants, industrial research grants, and low-interest loans certainly constitute a greater fiscal burden to the government than the tax incentives. This raises the possibility that government funding has crowded out private demand for the limited number of available researchers. An impact assessment of the underlyingresearch system and its governance is beyond the scope of the current study, but would be strongly advisable, A rough estimate of the available funding exceeds RM 300 million per year, which i s a fairly robust amount of funding considering that there are about 6,000 - 8,000 researchersinMalaysia. 81  6.9 Malaysia's volume-based Double Deduction on current R&D expenditure i s generous and well designed. A METR evaluation suggests the Double Deduction has a cost of RMllO million, a short-run impact of RM25 million, and a long-run impact of RM135 million. Its benefitkost ratio will be significantly reduced, however, if there are serious shortages of skilled scientists and if the operation, design, and definition of the scheme i s not improved. Its operation appears to be constrainedby IRB resources, and we recommend reviewing these resources with a view to significantly increasing them. We recommend undertaking consultations on a number of proposed reforms: focusing the schemes more on Malaysian-based R&D, tightening the definition of R&D, and developing procedures for tackling potentially serious tax avoidance. It would be worthwhile to undertake a quick consultation in time for Budget 2006 on these issues. A framework for structuring consultations is providedinAnnex 4. 6.10 Malaysia also has an extremely generous system of incentives for capital R&D via the IA and Pioneer schemes. The interaction of the three systems (IA, Pioneer Status, and DD), however, causes complexity and friction. The structural nature of the schemes also provides scope for widespread tax abuse, which may be difficult for MIDA and IRB to control without extensive, costly, and invasive monitoring. Firms' aversion to this heavy-handed monitoring may explain the very low uptake of these schemes. 6.11 The current system of tax incentives for R&D i s effective in the sense that a ringgit of incentives generates on the order of a ringgit of expenditure, on average, but this does not mean that it is economically efficient. Some of the increased spending no doubt reflects the rising costs of R&D personnel, rather than more R&D activity. More generally, when considering the overall level of subsidy that should be offered to R&D, three factors are important: the sensitivity of R&D to tax subsidies, the magnitude of social spillovers resulting from R&D, and the opportunity cost of the tax subsidyinterms of higher taxes levied elsewhere. The last factor - the opportunity cost of the tax subsidy interms of forgone revenue-is often ignoredbut canbe very important. The opportunity cost of an additional ringgit in subsidy offered to R&D i s the marginal cost of funds associatedwith the higher taxes that must be leviedelsewhere. Very roughcalculations of the marginal cost of funds and "reasonable guesses" as to the sensitivity of R&D to tax incentives and the magnitude of the social rate of return suggest that an R&D subsidy in Malaysia is justified on efficiency grounds. However, our preliminary view i s that the existing subsidy may come at too higha cost, and i s therefore overly generous. 6.12 We recommend a more systematic evaluation of the Investment Allowance (IA) and Pioneer Status schemes. Unless strong evidence is found to support them, we recommend slowly winding them down by not admitting any new applicants to the schemes. 6.13 Not only as an essential check on the operation and uptake of schemes, but also as a vital component of their cost-benefit analysis, we recommend costing all the major corporate tax schemes. If a need arises to make budgetary savings, based on our preliminaryfield trip andresearch, we recommend considering: 82  0 First, dropping the R&D IA and Pioneer schemes and investigating the cost- benefit ratio of these schemes for non-R&D projects with a view to potentialcuts; 0 Second, tightening the definition of the R&D schemes; and 0 Third, revisiting the rate of the Double Deduction after a cost-benefit evaluation of this incentive versus all other government funding for R&D (such as MOST1 grants). Use of Incentives and Impact on Plant-level Investment Rates 6.14 Descriptive statistics from PICS data show a striking difference in the uptake of incentives by establishments of different sizes: 60 percent of large establishments had benefited from incentives for their most recent investment, whereas only 38 percent of small- and medium-scale establishments had done so. Also, there i s only a moderate uptake of incentives and R&D-promoting schemes: fewer than half (48 percent) of the establishments surveyedhadgained from any incentive or scheme. 6.15 Regression-based estimates using the survey data show that the Reinvestment Allowance/Accelerated Capital Allowance has a significant effect on investment only among large establishments, not among small and medium ones. Among large establishments, those that usedthis incentive hadinvestment rates 9.4 percent higher than those that didnot use it. 6.16 From propensity score matching, we estimate that the Reinvestment Allowance/Accelerated Capital Allowance raises the investment rate by 14 percent, on average (the estimate i s statistically significant). This result i s consistent with the regression-based estimates, and together the results suggest that the Reinvestment Allowance/Accelerated Capital Allowance raises the investment rate of large establishments. Findings for other types of incentives and schemes are inconclusive, due to the small numbersof users covered by the data. I 6.17 Regulatory constraints-namely ownership regulations for large establishments and import regulations for small- and medium-scale establishments-are found to have strong negative effects on investment rates. Among large establishments, those that indicate that ownership regulation i s their biggest obstacle to doing business in Malaysia have on average a 20.3 percent lower investment rate than those not rating ownership regulation in this way. Small and medium establishments that indicate import regulation i s their biggest obstacle to doing business in Malaysia have on average a 13.1 percent lower investment rate than their counterparts who do not rate import regulation in this way. 6.18 These results suggest that in Malaysia, non-tax disincentives may be working at cross-purposes with tax incentives. Reducing the regulatory burden and improving the overall investment climate might well have more influence than a generous system of tax-incentives on a firm's decision to invest. Pressuring the tax system to compensatefor difficulties inthe investment climate may represent a huge lost opportunity for Malaysia. 83  Cost of Incentives 6.19 The marginal cost of funds (MCF) for taxes on capital inMalaysia differs widely depending upon the eligibility of an investment for various incentives. In our base case, the MCF ranges from 1.429 for investments not eligible for incentives to 1.160 for projects eligible for the Investment Allowance. Raising one more ringgit in tax revenue by slightly increasing the tax on capital in the general regime results in a reduction in income from the private sector of 1.429 ringgit. Of this, 1ringgit i s due to the 1ringgit in higher taxes collected and .429 ringgit is due to the marginal efficiency costs associated with the higher tax rate.56Therefore, reducing the tax incentive infavor of a more general decline intaxes would enhance efficiency. 6.20 To achieve economic efficiency requires the marginal cost of funds on all activities, in this case investment in different types of capital, to be the same. The fact that the MCF varies, in reflection of the plethora of incentives offered in Malaysia, suggests that there i s scope for raising the efficiency of the tax system by raising the effective tax rate on capital for investment projects with a low MCF and lowering the effective tax rate on capital for investments with a highMCF. B. NEXTSTEPS 6.21 The recommendations provided in this report are relatively general. While the basic idea behind them - a rationalization of the incentive system, and the use of automatic incentives offered through the tax system in the form of accelerated allowances, investment allowances, and investment tax credits - i s sound, the details require more study. Many of the calculations underlying our analysis are "back of the envelope" and some of the empirical work, such as the study of the impact of tax incentives on investment, i s preliminary. A more detailed cost-benefit analysis of specific incentives schemes, along the lines of what we did for R&D, i s required, as is more detailed empirical work on the impact of specific incentive schemes. 6.22 More work should also be done on the revenue implications of reform proposals, which we were not able to address in the current report. A Phase I1study would involve an in-depth analysis of more specific proposals. For example, if the existing set of incentives were to be rationalized and replaced with a simplified incentive system in the Income Tax Act, a set of METR and MCF calculations for specific scenarios could be developed. This, of course, would require more detail on specific options. 6.23 Finally, a broader examination of the way that incentives fit in to the entire tax system would be an important part of any comprehensive reform package. Other taxes, such as excise taxes, import duties, personal income taxes (and their integration with the corporate tax), sales taxes, interact with business taxes to create an overall tax 56This result is obtained using a tax sensitivity parameterof 1.With an assumption of higher tax sensitivity parameter (1.5 for example), the MCF in the generaltax case i s 1.819, of which 0.819 Ringgit i s due to the increase inefficiency costs. 84  environment that can be important for foreign investment. A detailed analysis of these other aspects of the tax environment inMalaysia would be an important part of a PhaseI1 study. 85  REFERENCES Aghion, P., and P. Howitt (1992), "A Model of Growth through CreativeDestruction," Econometrica 60 (March): 323-51. AItshuler, Rosanne, andJack Mintz (1964), "US Interest Allocation Rules:Effects and Policy", Mimeograph, NBER working paper. Bigsten,Arne, PaulCollier, StefanDercon, BernardGauthier, JanWillemGunning,Anders Isaksson, Abena Oduro, RemcoOostendorp, Cathy Pattillo, MBns Soderbom, Michel Sylvain, Francis Teal, andAlbert Zeufack (1999), "Investment inAfrica's Manufacturing Sector: A Four-Country PanelDataAnalysis." Oxford Bulletin of Economicsand Statistics 61 (4): 489- 512. Black, Dan, and Jeffrey Smith(2004), "How robust is the evidenceon the effects of college quality?Evidence from Matching." Journal of Econometrics 121(1): 99-124. Bloom, N.,L.Chennells,R. Griffith, andJ. van Reenen(2000), "How has tax affected the changing costs of.R&D? Evidencefrom eight countries", IFS working paperW97/3, Institute for Fiscal Studies. Bloom, N.,R. Griffith, andJ. van Reenen(2002), "Do R&Dtax credits work: evidencefrom an internationalpanelof countries 1979-97", Journal of Public Economics 85, 1-31. Blundell,R., andM.CostaDias (2000), "Evaluation Methods for Non-Experimental Data," Fiscal Studies 21 (4): 427-468. Boadway, R., N.Bruce, andJ. Mintz (1984) "Taxation, Inflation, andthe Effective MarginalTax RateInCanada," CanadianJournal of Economics 27: 286-99. Brealey, R., and S. Myers (2004), Principles of Corporate Finance. Sixth edition, McGraw-Hill publishers. Chen, Duanjie, andJackMintz (1993), "Taxation of Business Capital inCanada: An Inter- Industry andInter-provincial Comparison". InBusiness Taxation in Ontario. University of Toronto Press. Chen, Duanjie, andRichard Bird(2002), "Effective Tax Rates:Linking Two DistinctMeasures," TaxNotes International, 3 June: 1129-1139. Dehejia, RajeevH.,and SadekWahba (2002), "Propensity Score-MatchingMethods for Nonexperimental CausalStudies," Review of Economics and Statistics 84 (1): 151-161. Dehejia, RajeevH.,and SadekWahba (1999), "Causal Effects inNonexperimental Studies: Reevaluatingthe Evaluationof Training Programs." Journal of theAmerican Statistical Association 94 (448): 1053-1062. Devereux,M.,R. Griffith, andH.Simpson(1999), "The geographic distribution of production activity inthe UK",IFS Working PapersW99/26, Institutefor Fiscal Studies. Goolsbee, A. (1998a), "Investment Tax Incentives, Prices, andthe Supply of Capital Goods", Quarterly Journal of Economics 113 (l), 121 148. February: - Goolsbee,A. (1998b), "Does R&D Policy PrimarilyBenefit Scientists andEngineers?' American Economic Review (Papers andProceedings), 88 (2), May: 298-302. 86  Goolsbee, A. (2003), "Investment Tax Subsidies and the Wages of Capital Goods Workers: To the Workers Go the Spoils?" National TaxJournal LVI(1, part 2), March: 153-166. Griliches, Z. (1998), R&D and Productivity: TheEconometricEvidence.University of Chicago Press. Grossman, G.,and E.Helpman (1991), Innovation and Growth in the Global Economy. Cambridge, MA: MIT Press. Hagen, Tobias (2002), "Do Temporary Workers Receive RiskPremiums?Assessing the Wage Effects of Fixed-Term Contracts inWest Germany by a MatchingEstimator Compared with Parametric Approaches." Labor 16 (4): 667-705. Hall, B., and J. van Reenen (2000), "How effective are fiscal incentives for R&D? A review of the evidence", ResearchPolicy: 449-469. Heckman, James, Hidehiko Ichimura, and Petra Todd (1997), "Matching as an Econometric EvaluationEstimator: Evidencefrom Evaluating a Job Training Program." Review of EconomicStudies 64 (4): 605-654. Heckman, James, Hidehiko Ichimura, and Petra Todd (1998), "Matching as an Econometric EvaluationEstimator." Review of Economic Studies 65 (2): 261-294. Heckman, James, Hidehiko Ichimura, Jeffrey Smith, and Petra Todd (1998), "Characterizing Selection Bias UsingExperimentalData,'' Econometrica66 (5): 1017-1098. InternationalBureauof Fiscal Documentation(IBFD),Country Chapters (inloose-leaf form, up to date) for China, Indonesia, Malaysia, Singapore, and Thailand. Jaffe, A. B.,M.Trajtenberg, and R. Henderson (1993), "Geographic Localization of Knowledge Spillovers as Evidencedby Patent Citations", Quarterly Journal of Economics 79: 577-598. Jaffe, A., and M.Trajtenberg (1999), "International Knowledge Flows: Evidence from Patent Citations", Economics of Innovation and New Technology,Vol. 8: 105-136. King, M.,and D.Fullerton (1984), Thetaxation of incomefrom capital. University of Chicago Press. King, MervynA., and DonFullerton (1984), TheTaxation of Incomefrom Capital: a ComparativeStudy of the United States,the United Kingdom, Sweden, and West Germany. University of Chicago Press. Lattimore, R. (1997), "Research and Development Fiscal IncentivesinAustralia: Impact and Policy Lessons". InOECD, Policy Evaluation in Innovation and Technology: TowardsBest Practice. Leechor, Chad, and Jack Mintz (1993), "On the Taxation of Multinational Corporate Investment When the Deferral Methodi s Usedby the CapitalExporting Country", Journal of Public Economics,Vol. 51, Issue 1:75-96. Leuven, Edwin, and Barbara Sianesi (2003), "PSMATCH2: Statamodule to performfull Mahalanobis and propensity score matching, common support graphing, and covariate imbalance testing".~http://ideas.repec.org/c/boc/bs432001 .html.Version 1.2.3> MalaysianAssociation of the Institute of Chartered Secretaries and Administrators (2004), 2004 Budget Commentary & Tax Information. MalaysianInstitute of Accountants, Malaysian Instituteof Certified Public Accountants, and Malaysian Institute of Taxation (MIMS) (2004), Malaysia Investment in theManufacturing Sector. 87  MIDA (2004) <http://www.mida.gov.my/> Mintz, Jack (1990), "Corporate Tax Holidays and Investment," WorldBank Economic Review 4 (1):81-102. Mintz, Jack (2004) "Conduit Entities: Implications of Indirect Tax-Efficient Financing Structures for Real Investment", International Taxand Public Finance, August, 11(4), 419-434. Mintz, Jack and Michael Smart (2004) "Income shifting,investment, and tax competition: theory and evidence from provincial taxation inCanada" Journal of Public Economics, with, 88 (6), pp. 1149-78. NaudC, Willem, Remco Oostendorp, and Philip Serumage-Zake (2000), "Determinants of Investment and Exports of South African Manufacturing Firms:Firm-Level Survey Results". Mimeo. OECD(1991), Taxing Profits ina Global Economy: Domestic and International Issues.Paris. Ravallion, M.(2001), "The Mystery of the Vanishing Benefits: An Introduction to Impact Evaluation", WorldBank Economic Review 15 (1): 115-140. Romer, P. (1990), "Endogenous Technological Change", Journal of Political Economy 98 (October): S71-S-102. Rosenbaum, Paul R., and Donald B. Rubin(1983), "The Central Role of the Propensity Score in ObservationalStudies for CausalEffects,'' Biometrika,70 (1):41-55. Rosenbaum, Paul R., and Donald B. Rubin(1985), "Constructing a Control Group Using Multivariate MatchedSamplingMethods that Incorporate the Propensity Score," The American Statistician 39 (1): 33-38. Sawyer, A. (2004), "Reflections on Providing Tax Incentives for Research and Development: New Zealand at the Cross Roads" (unpublished). Shankar, Raja, and Anwar Shah (2003), "Bridging the economic divide within countries: a scorecard on the performance of regional policies inreducingregional income disparities," World Development 31 (8), August: 1421-1441. Silverman, B. W. (1986), Density Estimationfor Statistics and Data Analysis. Monographs on Statistics and Applied Probability. London: Chapman and Hall. Smith, Jeffrey, and PetraTodd (2005), "Does Matching Overcome LaLonde's Critique of NonexperimentalEstimators?" Journal of Econometrics, forthcoming. Stiglitz, J. (2000), Economics of the Public Sector.W. W. Nortonand Company. UNCTAD (2000), TaxIncentives and Foreign Direct Investment: A Global Survey. Geneva and New York United Nations Publications. 88  ANNEXES ANNEXES: CONTENTS ANNEX 1 ......................................................................................................................... 90 Annex 1Section A: BusinessTaxation inMalaysia: An Overview ............................. 90 Annex 1Section B:Methodology for Estimating Marginal Effective Tax Rates on Capital Investment ........................................................................................................ 94 Annex 1Section C: Impact of Non-Tax Parameters on the Estimate of Effective Tax Rates............................................................................................................................ 100 Annex 1Section D:Tax andIncentive Systems in China. Indonesia. Singapore. and Thailand ...................................................................................................................... 102 ANNEX 2 ....................................................................................................................... 112 Annex 2 Section A: Data and Statistics ...................................................................... 112 Annex 2 Section B:METRMethodology .................................................................. 116 Annex 2 Section C: Tax Exhaustion. Profits. andCosting......................................... 118 Annex 2 Section D:R&D Incentives inEast Asia ..................................................... 121 Annex 2 Section E:Interview Schedule ..................................................................... 126 ANNEX 3 ....................................................................................................................... 127 Annex 3 Section A: Descriptive Statistics: Investment Function............................... 127 Annex 3 Section B: Propensity Score Matching......................................................... 128 Annex 3 Section C: Content of Malaysia's Incentives for the Manufacturing Sector 134 Annex 4: ConductingPolicy Consultations ............................................................... 160 89  ANNEX 1 ANNEX 1SECTIONA: BUSINESSTAXATIONINMALAYSIA:OVERVIEW5' AN This section provides an overview of two major taxes inMalaysia: direct taxes oncapital investment and transaction taxes on capital inputs. Against this background, the investment incentives granted to manufacturing, agricultural and tourist industry are also reviewed. 1.Taxes oncapitalinvestments As Table 1shows, capital taxes include the direct tax on income generated from capital investment, personal income taxes on investment income5*, and the property taxes. The company income taxes Business income i s subject to the income tax under a two-rate scheme. The higher rate of 28 percent applies to companies with a paid-up capital above RM2.5 million. The lower rate of 20 percent applies to companies with a paid-up capital of under RM2.5 million and on their first RM500,OOO of taxable income with the subsequent taxable income being subject to the 28 percent rate. Capital-related tax allowances. Depreciation allowance i s calculated by the straight-line (SL) method. For machinery and equipment, the initial allowance i s 20 percent and the annual allowance ranges from 10 percent to 20 percent, with 14 percent for most plant and machinery. For industrialbuildings,the initial allowance i s 10percent andthe annual allowance 3 percent. Accelerated capital allowance i s also available to certain assets, industries, or activities such as those used exclusively for the conservation of energy, which i s irrelevant to our study. Furthermore, this type of acceleratedallowance does not apply to a company that i s granted any incentives. Inventory accounting method According to Ernst & Young,59trading inventory i s valued at the lower of cost or net realizable value. Cost must be determined under the first-in first-out (FIFO) method; the last-in, first-out (LIFO) methodi s not accepted. Loss carry-over According to the International Bureau of Fiscal Documentation (IBFD,Malaysia, page64b), businesslossesmay be carriedforward indefinitely. 5' Unless otherwise specified, this appendix is based on the following documents: 1) Malaysia Investment in the Manufacturing Sector: Policies, Incentives, and Facilities (MIMS). 2) 2004 Budget Commentary & Tax Information, published by the Malaysian Institute of Certified Public Accountants (MICPA), Malaysian Institute of Accountants (MIA) and Malaysian Institute of Taxation (MIT). ''Asillustrated by the effective tax rate analysis, taxes on any personal investment income could affect the cost of capital investmentthrough the financing of capital. 59 Ernst& Young (2001),page416. 90  Tax on dividends, interest, and capitalgains According to IBFD, both dividends and interest incomes are taxed as ordinary income although dividends are taxed using an imputation system. There i s no withholding tax on these incomes received by residents. For non-residents, the withholding tax i s 15 percent on interest unless specified under a double-tax treaty. Capital gains from financial investment are not treated as income and hence are not taxable. However, income arisingfrom recurring capital transactions i s considered to be derived from trade and taxed as business income. Capital gains derived from the disposition of real property (and shares in a real property company) are subject to real property gains tax. Property taxes According to IBFD, "Property tax (assessment rates and quit rent) i s payable by property owners pursuant to the legislation of local or municipal authorities on real properties situated within their jurisdiction. Local authorities in the State of Johore impose rates on the basis of improved value. Other local authorities levy rates on the basis of the annual value or capital value of the property." However, no information on specific property tax rates is available. 2. TransactionTaxeson BusinessInputs Malaysiahas abolished import duties on a wide range of raw materials, components, and machinery. Therefore, it is reasonableto assume a zero import duty rate on capital goods for purposesof our METR calculations. The other major transaction tax i s the sales tax. This i s a consumption tax on imported or manufactured goods. The general rate i s 10percent. However, according to the Ministry of Finance, importedcapital goods are exempted from this sales tax. TheStructure of Tax Incentives Malaysia provides an extensive range of tax incentives. These incentives are granted under the Promotion oflnvestments Act 1986and its amendments (PIA) and the Income TaxAct and its amendments (ITA). A. Thetax incentives granted under PIA These incentives can be categorized as two alternatives: the Pioneer Status (PS) or Investment Allowance (IA). These incentives are granted by the Malaysian Industrial Development Authority (MIDA). Eligibility for these alternative incentives i s based on certain priorities, including the levels of value-added, technology used and industrial linkages. Such eligible projects are termed as "promoted activities" or "promoted products." 91  The originaE content of these alternatives are the following: for Pioneer status, a company enjoys a 70 percent exemption from the payment of income tax for a 5-year period; and, alternatively, a company may obtain an Investment Allowance of 60 percent for its qualifying capital expenditure up to 70 percent of its statutory income6' within five years from the date on which the first qualifying capital expenditure i s incurred. The enhanced incentive under these two alternative programs i s the following: a company may enjoy either the Pioneer Status with a 100 percent tax exemption on their statutory income or a 100 percent IA up to 100 percent of the statutory income, both of which last for a 5-year period. This enhanced incentive i s available for eligible projects located in the "promoted areas"61provided that the project applications are received by December 31,2005. However, there are many variations of the above alternative incentives interms of the percentage of tax exemption, the percentage of IA, and the number of years for eligibility. These variations are granted on a differentiated basis according to industry and subsector features. For example, the aforementioned original and enhanced investment incentives are available to eligible manufacturing projects in general, but more generous incentives are tailored to some subsectors within manufacturing industry such as high technology (W)companies, strategic projects (SP) that involve products or activities of national importance, small- and medium-scale enterprises (SMEs), large firms that participate in an industrial linkage program (ILP), and firms producing machinery and equipment (M&E). Figure 1 provides a summary of this complicated structure of investment incentives for manufacturing, agricultural andtourism projects62. There is also an incentive for export promotion inthe form of a double deduction for expenses incurred in the area of marketingand sales (IBFD, Malaysia, pp. 124-145. However, this type of incentive does not affect our h4ETR analysis at the margin and hence i s ignoredinthis study. B. The incentivesgranted under the Income Tax Act (ITA) The investment incentives under the ITA are targeted to reinvestment inselective sectors or projects. These incentives are granted by the Inland Revenue Board (IRB). For manufacturing industry, a reinvestment allowance of 60 percent up to 70 percent of 6o According to MIMS, "statutory income is derived alter deducting revenue expenditure and capital allowances from the gross income," which can be seen as a proxy for taxable income at the margin for the purposes of METR analysis. 61 Promoted areas include the States of Sabah and Sarawak and the designated "Eastern Corridor" of Peninsular Malaysia which covers the States of Kelantan, Terengganu and Pahang, and the district of Mersinginthe State of Johore. These are eco-tourism and agro-tourism projects including hotel businesses (Le. construction of low-cost hotels up to a three-star category and expansiodmodernization of existing hotels), construction of holiday camps, recreationalprojects, and construction of convention centers with a capacity that can accommodate at least 3,000 participants. 92  statutory income i s available for a manufacturing firm that "has been in operation for at least 12 months and incurs qualifying capital expenditure to expand, modernize or automate its existing business or diversify its existing business into any related products within the same industry." (MIMS 2004) This incentive i s given for a period of 15 consecutive years beginning from the year the first reinvestment i s made, after which firms that reinvest in the manufacture of promoted products are eligible for an Accelerated Capital Allowance (ACA) that can be utilizedwithin three years (i.e. with an initial allowance of 40 percent and an annual allowance of 20 percent). For agricultural industry, a 15-year reinvestment allowance followed by an ACA, the same as that given to the manufacturing industry, i s available to any entity engaged for at least 12months in the production of "essential" food such as rice, maize, vegetables, tubers, livestock, aquatic products, and any other activities approved by the Minister of Finance. 93  ANNEX1SECTIONB:METHODOLOGY FOR ESTIMATINGMARGINAL EFFECTIVE RATESONCAPITALINVESTMENT TAX 1. The Concept The marginal effective tax rate on capital investment calculated in our study i s an effective tax rate on capital investment made by any legal entity that generates active income and i s subject to the company income tax. It combines all the taxes that affect capital investment at the firm level instead of at the personal level. Personal income taxes, however, may be incorporated into our calculation when they affect capital investment at the level of the firm. For example, withholding taxes on dividends are common and do affect the cost of capital invested at the firm level through their impact on the rate of return to equity required by shareholders. The marginal effective tax rate (METR) measures the impact of a tax system on an incremental unit of capital investment. It incorporates the effects not only of statutory tax rates and other tax provisions (e.g. tax depreciation, tax credits, tax deductibility, tax holidays, etc.) but also various economic factors interacting with these tax provisions (e.g. financial costs, the inflation rate, and the structure of investment). In other words, METR i s a summary indicator of the overall tax burden imposed by a tax system on a new investment within a given economic environment. The primary assumption underlying the calculation for our estimate of METR i s that of profit-maximization. Profit-maximizing firms base their investment decisions on the present value of foreseeable net revenues from investment. Taxes reduce the portion of the profits accruing to the investor, while tax allowances or credits mitigate such a reduction in accrued profits. For profit-maximizing firms, the gross rate of return on capital (net of economic depreciation) must be equal to the financing cost of capital, adjusted for taxes. The size of this adjustment for taxes on a new investment i s the marginal effective tax rate, or METR, on capital. For example, if the gross-of-tax rate of return to capital i s 20 per cent and the net-of-tax rate of return i s 10 per cent, then the METR on capital is 50 per cent. Owing to the interaction between these statutory tax provisions and actual economic/industrial conditions (e.g., financing conditions and capital structure), METRs may vary by industry even under the same tax regime. The second important assumption used in our METR analysis i s the concept of a "small open economy." In a small open economy, firms have the option of acquiring investment funds from both domestic and international markets; and the domestic market interest rate for a host country i s determined by international trading of currencies. Accordingly, even though international statistics may show variations in interest rates among countries under comparison, when foreign capital investment i s involved, it i s reasonable to assume that the real interest rate i s the same across all the home countries under comparison. Finally, to maximize profits, firms will always want to minimize the cost of financing by issuingdebt until the tax inclusive cost of debt financing equals the tax inclusive cost of equity financing. Therefore, besides interest deductibility for income tax purposes at the firm level, the gap between the personal income tax on interest and 94  that on equity income (including capital gains and dividends) can also affect the financing structure of firms as measuredby their debt to equity ratio. The METR analysis in this study deals with only "profitable" firms. B y "profitable" we mean those firms that have taxable income and are not in a loss-cmy- over position. This assumption i s important because a tax holiday i s irrelevant to an unprofitablefirm that does not have to pay income taxes. Furthermore, calculating METR for "tax-loss" firms would require data on the average number of years requiredfor these firms to write off their losses andbecome taxable, which i s beyond our policy concerns in the present report. 2. Methodology The standard method used to estimate effective tax rates has been extensively d~cumented~~.The formula based on this method has been modified by incorporating some miscellaneous taxes64such as capital tax, property tax, and tax on transfer of property as in Malaysia, and for foreign capital inve~tment~~. The following are the general formulas usedinthis study. Effective tax rate (t) The effective tax rate on a given type of capital i s defined as the proportional tax rate of return required by an investor (p).rGi s the marginal revenue product (or user difference between the gross-of-tax rate of return requiredby a firm (rG)andthe net-of- cost of capital, inequilibrium) net of economic depreciation. The after-tax rate of return i s the weighted average of the returnto debt andequity securities heldby the financial investor. Thus, the effective tax rate (t)i s defined as Thereal cost of financing ($, For domestic investors, the real cost of financing (r`) i s defined by withJ = the ratio of debt to assets ratio, i = cost of debt, U = the statutory corporate income tax rate, p = cost of equity, and n=inflation rate. That is, the cost of financing for an investor is the weighted-average cost of financing net of the inflation rate. For foreign investors, a realistic assumption i s that the investors finance their capital inthe host country usingtwo sources of funds. The first i s debt raised inthe host country andthe secondis equity investedinthe host country, the latter beingcomprised G3 Boadway,Bruce, andMintz (1984) and KingandFullerton (1984). G4 Chenand Mintz (1993). G5 MintzandTsiopoulos(1992). 95  of both debt andequity raisedinthe home country. Therefore, there are three components inthe real cost of financing (yf) for aforeign investor: debt andequityraisedat home and debt raised inthe host country. With differenttax regimes inhomecountries, the cost of debt raised inthe host country may be treateddifferently by the home tax regime. When the cost of debt raisedinthe host country can be fully deducted for income tax purposes at home, the real cost of financing for a foreign investor i s definedby the following equation: Inthis equation, all the home country variables are denotedby the symbol `,with J" = the ratio of debt to assets ratio inthe home country, i' = cost of debt inthe home country, U' = the statutory corporate income tax rate in the home country, p' = cost of equity in the home country, and d = the inflation rate in the home country. In addition, y = the ratio of debt raised in the host country to total investment funds, x = the weighted average withholding tax rate in the host country, i = cost of debt in the host country, U = the statutory corporate income tax rate in the host country, and n= the inflation rate in the host country. As equation (2') states, the real cost of financing to a foreign investor is the weighted-average of the costs of its investment funds taken from the home country and the debt raised in the host country, net of the home country inflation rate. The former i s the weighted average cost of financing at home net of withholding tax payable inthe host country, and the latter i s the cost of debt in the host country adjusted by income tax deductibility and any difference inthe inflationrate between the home and host countries. Equation (2') i s applicable to the French case in which all the taxes paid in a foreign country by French multinationals are "spared.," so as the tax benefit available in the foreign country including the interest deductibility (e.g. expressed by term i(1-U) in the equation). This equation may also be applicable to the US case assuming that US investors in Malaysia are in a "dejkit credit" position regarding their income tax liability at home (see below for explanation). The US case is different from the French one in that taxes paid in foreign countries are not "spared" but creditable against the home tax liabilities. A further complication i s that, since the implementation of the post 1986 interest allocation rule, the U S multinationals are not allowed to allocate their foreign interest cost across different foreign investment projects. As a result, for each foreign country where a US multinational i s doing business and paying tax, its foreign tax creditability for that given country may be in either a "deficit credit" position or an "excess credit" position. In the "deficit credit" position, the foreign tax paid in a given country by a US investor is below his tax liability at home corresponding to that given foreign income and hence can be fully credited against its home tax liability. B y way of contrast, in the "excess credit" position, the foreign tax paid in a given country by a US investor surpasses his home tax liability corresponding to that given foreign income and hence cannot be fully credited against his home tax liability. Inthis latter situation, all the formulas for foreign investors presentedinthis appendix need to be modified by adding a 96  "penalty" term to reflect the less-than-full credibility of foreign income tax paid66.In our study, we assume that US investors in Malaysia and other countries covered in this study are in a "deficit credit" position and hence do not incur any tax cost beyond their tax liability at home corresponding to their income generated in Malaysia. This assumption than that in the US (or at most equals that in the US if there is no income tax levied at the is based on thefact that the overall Malaysian income tax rate (up to 35 percent) is lower state level). Of course, when the tax allowances (e.g. tax depreciation allowance) in a host country i s less generous than that at home, even a lower income tax rate in the host country compared to that at home may lead to an "excess credit" position for a multinational relating to its investment in the given host country. We assume this situation i s irrelevant to US investment in Malaysia. Also note that, for a given U S investor, assuming an "excess credit" position will generally result ina lower METRthan assuming a "deficit credit" position. Therefore, the assumption in our METR comparison across different host countries that a representative US investor i s in a "deficit credit" position does not affect the METR ranking in our cross-country comparison to any material extent. Net-of-tax rate of return on capital (p) For domestic investors, the net-of-tax rate of return on capital is defined by the formula This is the rate of return on capital requiredby financial investors, or suppliers of investment funds to firms. Note that financial investors often include firms themselves when there i s equity generatedinternally. For foreign investors, the formula i s This is the net-of-tax rate of return on capital required by financial investors including foreign investors themselves and their creditors in both home and host countries. Applying (3) and (3') to equation (l), respectively, results in the effective company tax rate on capital for domestic investors andthat for foreign investors. Leechor andMintz (1993) 97  Gross-of-tax rate of return (rG)on capital67 A. Depreciable assets (i.e. buildingsandmachinery andequipment) For domestic investors rG= (l+tm)(f+s)(l- k)[l - A +z(l-U)/(a+f+nj]/[(l-U)(l-tp -tg)] - S (5) Where tin = tax on transfer of property, or a transaction tax (e.g., import duty) on capital goods wherever this i s applicable, S= economic depreciation rate, k = investment tax credit rate, A = the present value of tax benefit from the Investment Allowance and depreciation allowance, z = capital tax rate, a= tax depreciation rate, tp = property tax rate based on the rental value, and tg = gross receipts tax rate, or presumptive tax that i s basedon the gross revenue. For foreign investors rG'= (l+tm)(f'+s)(l- k)[l - A +z(l-U)/(a+f'+nj]/[(l-U)(l-tp -tg)] - S (5') B.Inventory For domestic investors Where tin = sales tax on raw materials where it i s applicable, and <= 1for the FIFO accounting method, 0 for LIFO, and 0.5 for the average cost method. For foreign investors, the formula i s the same except that the financing cost should be the one relevant to the foreign investors. That is, f should be replacedby f'. C.Land For domestic investors Where tm =property transfer tax. For foreign investors, the formula i s the same except that the financing cost should be the one relevant to foreign investors, i.e., f should be replaced by f'. 67Formulas provided here are only for the regular case where companies are profitable and pay taxes. For the tax-holiday case or the case of tax reduction for a limitedperiod, the formulas for depreciable assets are more complicated in that the present value of tax allowances needs to be adjustedto reflect the true impact of the tax holiday or tax reduction on METR. For non-depreciable assets such as inventory and land, the formula for the tax-holiday case is the same as that for the taxpaying case except that the statutory tax rate(s) may,differdue to the lower or zero rate arising from tax incentives. 98  Aggregation The effective tax rate for a given industry i s the proportional difference between the weighted averageof the before-tax rate of returnby asset type and the after-tax rate of return; the latter i s the same across asset types within a given sector@. That is, the marginal effective tax rate for industry i. ti, i scalculated as following: where j denotes asset type (Le. investments in buildings, machinery, inventories, and land), andwij denotesthe weight of asset typej inindustry i. Inter-sectoral METR dispersion The inter-sectoral METR dispersion coefficient is the weighted average of the standarddeviations of METR by type of assets across industry. It i s estimated as where tu i s the effective tax rate on assetj for industry i, and tj i s the average effective tax rate for assetj across all industries. The net-of-tax rate of return can be the same across all the sectors when there is no sectoral differentiation inthe statutory income tax rate. 99  ANNEX 1SECTIONC: IMPACTOFNON-TAX PARAMETERSONTHE ESTIMATE EFFECTIVE RATES OF TAX 1. ExpectedInflationRate The expected inflation rate affects the effective tax rate on capital through its impact on the nominal interest rate. For a given real interest rate, the higher the inflation rate, the higher the nominal interest rate will be. When there is nu regulation fur adjusting the impact of inflation, the nominal interest rate interacts with taxes mainly through the following three channels. Firstly, interest cost i s deductible for income tax purposes at the nominal rate. As aresult, the higher the nominalinterest rate inrelationto a fixed real interest rate, the lower the real after-tax financing cost, and hence the lower the METR. This effect i s particularly favorable for leveraged land financing. Secondly, the accumulated present value of a given annual tax depreciation allowance decreases as the nominal interest rate rises. Since higher inflation lowers the present value of tax depreciation allowances, it increases METR on depreciable assets. And finally, if the first-in-first-out method i s usedfor inventory accounting, it may result in inflatedtaxable income and, hence, a higher METR on inventory when prices rise. Since inflation thus affects METR on different assets in different directions, its net impact on capital will depend upon the capital structure related to a given industry. (See the end section of this Appendix for further explanation of the capital structure by industry.) 2. Expected Real Interest Rate The impact of the real interest rate on the effective tax rate i s inpart similar to the impact of inflation. For example, as the real rate rises, so will the nominal rate, thus increasing the effective tax rate on depreciable assets. For a given debt-asset ratio, however, unless inflation i s high, there i s unlikely to be much of a distortion in effective tax rates arising from the deductibility of interest. We use the US real interest rate for our study assuming there i s full mobility of investment funds within the international financing market. As shown in Table A2, the real interest rate in the US i s 4 percent, corresponding to a nominal interest rate of 6 percent and an inflationrate of 2 percent. 3. Debt-Asset Ratio The ratio of debt to assets i s sometimes referred to as the financing structure. As already noted, the impact of this ratio on the effective tax rate i s related to the expected inflation rate and (real) interest rate. For a given inflation rate and real interest rate, the higher the debt-asset ratio, the more the potential benefit from tax deductibility for debt financing cost, or interest expenses. A higher debt-asset ratio may thus reduce effective tax rates through lowering the real after-tax cost of financing. For simplicity, we apply a debt to assetsratio of 40 percent across sectors and across borders inour study. 100  4. Economic Depreciation The economic depreciation rate interacts with the tax depreciation allowance to affect the effective tax rate. Under our assumption of full mobility of capital and technology, a given type of machinery i s depreciated at the same economic rate everywhere around the world. Investment in this given type of machinery will incur a higher (lower) effective tax rate in those countries that provide a lower (higher) tax depreciation allowance, andhence may bediscouraged (encouraged) inthese countries, 5. CapitalStructure A capital investment generally involves two categoriesof capital: depreciable and non-depreciable assets. These two categories can be further divided into four types: buildings and machinery (both depreciable) and inventory and land (non-depreciable). Capital investments in different industries are structured differently by nature of their business. Moreover, under the same statutory tax rate(s), different types of assets may incur different effective tax rates due to the various interactions between tax provisions and non-tax parameters discussed above. Inthe absence of other information, we use the Canadian capital structure by industry to aggregate these differentiated effective tax rates on various types of capital for a given industry. Table Al.1: Non-Tax Parameters (%) for FDIHost Countries, 2004 Part 1:by country Malaysia China Indonesia Singapore Thailand Inflationratea 1.2 4.4 6.8 2.0 3.0 Interest rateb 5.2 8.4 10.8 6.5 7.8 Debtto assets ratio Debt raisedabroad to home 40 40 40 40 40 capital Debt to assets ratio in home 29 29 29 29 29 Part 2: by industry Manufacturing Manufacturing- Agriculture Tourism relatedservices Economicdepreciationrate Buildings 3.7 3.4 4.1 3.6 Machinery 18.1 22.4 14.2 24.8 Capitalstructure byasset type Manufacturing Buildings 23.8 48.4 17.2 40.7 Machinery 47.2 16.6 14.1 33.3 Inventory 27.7 22.8 47.9 14.7 Land 1.3 12.2 20.8 11.3 Jote:Unless notedotherwise, numbers presentedinthis table are basedon Canadianstatistics and were obtainedfrom the InternationalTax Program, University of Toronto. aThesenumbers are the annualchanges inconsumer prices published inTheEconomist, July 10"-16", 2004. Considering capital mobility, we assumeareal interestrate of 4 per cent for foreign direct investment in all host countries.This assumptionis made basedon the interest rate on corporatebonds and annualchange inconsumer pricesintheUnitedStates, published inTheEconomist, July 10~-16",2004. 101  ANNEX 1SECTIOND:TAXAND INCENTIVE SYSTEMSINCHINA,INDONESIA SINGAPORE, AND THAILAND6' This section presentsan outline of businesstaxation inMalaysia's four main competitor countries for foreign investment: China, Indonesia, Singapore, andThailand. This outline covers three major tax components: the capital tax provisionwhich describes how the income from capital investment i s taxed, the transaction taxes that may affect the cost of capital goods, and the tax incentives that are intended to attract foreign capital investment to the country. To enable comparisons to be made, the tax provisions reviewed here are those relevant to foreign manufacturing firms. A summary of the main features of each country's corporate tax system i s providedinTable A2. CHINA Foreign firms or foreign investment enterprises (FIEs), including both "wholly foreign-owned enterprises" and joint ventures, are taxed under the Foreign Zncome Tax Law and the Foreign Income Tax Regulations. The Minister of Finance stated in 2002 that the existing tax incentives available to FIEs would be terminated in 2003, with a view to the unification of the income tax regime for FIEs and that for domestic enterprises. However, there i s no timetable for the drafting of the required legislation or regulations inrespectof the proposed unification of the two tax regimes. 1. CapitalTax Provisions The corporate income tax rate Taxable income derived by foreign investment enterprises i s taxed at the rate of 33 percent, with 30 percent as nationalincome tax and 3 percent local. In practice, since more than 80 percent of foreign investment i s located in the special economic zones and other special areas where the national income tax rate is reduced and the local tax i s waived, the effective rate applying to foreign manufacturing firms i s either 15 percent (in special economic zones and in the middle and western regions) or 24 percent (in other specified zones). For those foreign manufacturing firms that are either using advanced technology or are export-oriented, the tax rate i s normally 10percent. (See below for more details.) The tax depreciation rule Depreciation for fixed assets i s computed annually on a straight-line basis on the assumption that there will be a residual salvage value of 10 percent of the original cost. Inother words, only 90 percent of the cost of an asset may be written off. The minimumdepreciation periodi s 20 years for buildings and structures, 10 years for general machinery and equipment, and 5 years for electronic equipment. Accelerated depreciation i s available to machinery and equipment used with high intensity (e.g. multi-shifts). Assuming that machinery and equipment used for manufacturing purpose comprises a big portion of the value of electronic equipment and i s often used more intensively, a 7.5-year of useful life for machinery and equipment i s usedfor our METR analysis. 69 Unless otherwise specified, this outline is written based on the country information published by the International Bureauof FiscalDocumentation. 102  Inventory accounting method Any of the following conventional accounting methods are allowed for tax purposes: first-in-first-out (FIFO), last-in-first-out (LIFO), moving average, or weighted average. However, once a method of accounting for inventory has been chosen, only by application and upon approval from the tax authority, may a taxpayer changethe chosenmethod. Loss curry-overs Losses may be carried forward for 5 years and no carry-back i s allowed. However, no distinction i s made between capital losses and ordinary business losses since capital gains are taxed as ordinary income. Taxes on investment income Dividends and interest income are taxed at a flat rate of 20 percent although inter-firm dividends are exempted. Capital gains are taxed as ordinary income and hence subject to different income tax rates depending on the recipients' tax situation. However, according to the double taxation treaty between China and the United States, interest and dividends are subject to a 10 percent withholding tax rate (Emst & Young, 2001). Also, there were several official "notices" issued during the late 1990s stating that capital gains on shares of listed companies were to be exempt from the individual income tax. There has been no official elimination of this type of exemptions (IBFD/China, pp. 134-135). The vropertv tux The only property tax applicable to FIEs i s the urban real estate tax. The tax i s leviedon buildings on an annual basis at the rate of 1.2 percent of the "residual value" or 18 percent of the standard rental of the real estate. The tax i s levied on land at the rate of 1.8 percent of the assessed value, which naturally varies by location. There i s also a stamp duty on documents for the transfer of property. The tax rate i s 0.05 percent on the value of transferred property. 2. Transaction Taxes The general VAT rate i s 17 percent, which i s refundable if paid on business inputs. The average rate of the import duty for machinery (and equipment) is 9.6 percent (IBFD-China, page 189). However, import duty i s exempted for imported equipment for approved projects. This tax exemption i s based on the new Guideline Catalogue of Foreign Investment Industries. 103  3. Tax Incentives The main form of tax incentive for foreign investment is tax exemption and reduction, which includes income tax holidays and reductions, exemption from import duty and VAT for importation of machinery and equipment used for designated projects, and exemption from other taxes such as property tax, urban maintenance and development tax, and the 3 percent localincome tax. The most common income tax holiday and reductions that are relevant to manufacturing industry i s the five-year tax incentive package. This package i s available to all foreign investment enterprises engaged in production that operate for 10 years or more. The incentive comprises a 2-year tax holiday and a subsequent 3-year 50 percent tax reduction. The tax holiday starts with an enterprise's "first profit-making year," which i s defined to be the first profitable year after the enterprise has commenced productionor operations andearlier losses have been carried forward and set off against profits. In addition to this five-year tax incentive package, reduced tax rates of 24 percent and 15 percent are provided to those foreign manufacturing firms located in the special areas of China including special economic zones (15 percent), open coastal economic zones (24 percent), and the middle and western regions (15 percent). A further reduced income tax rate of 10 percent is available to those foreign manufacturing firms in the specified zones ifthey are either determined to be technologically advanced or export- oriented. As a result of the combination of the aforementioned tax holidays and reduced rates, a manufacturing firm located in a special economic zone and using advanced technology may enjoy a two-year tax holiday, three-year reduced income tax rate of 7.5 percent, and the 10percent income tax rate for the rest of its business life. Best of all, this firm does not haveto worry about its income tax liability untilit has fully recovered from its start-up losses, neither does it have to worry, after becoming profitable, about any taxes other than the income tax and the urban real estate tax. INDONESIA Indonesia taxes domestic and foreign ownedfirms equally although it imposes some industrial restrictions on foreign investment. 1. CapitalTax Provisions The corporate income tax rate. Effective from January 1, 2001, the incomes of resident companies are taxed under a progressive schedule with the following three income tax rates : 10 percent (under IDR 50 million), 15 percent (IDR 50-100 million), and 30 percent (above IDR 100million). It is noteworthy that the deduction for interest expenses is regulated under a thin- capitalization rule. As such, the deduction of interest may be limited if the ratio between debt andequity capital exceeds 3:1. 104  The tax depreciation rule. Taxpayers may opt to use either the straight-line or declining-balance method to depreciate any depreciable assets except buildings. These assets are classified into four groups; and each of these four groups i s assigned both a straight-line depreciation rate and a declining-balance rate. The declining-balance rate i s double the straight-line rate. Machinery and equipment used by manufacturing industry falls into Groups 2 and 3. The average of declining-balance rates for these two groups i s 18.75 percent, which i s the rate usedfor our METR analysis. For buildings, only the straight-line meth.odof depreciation may be used. The depreciation rate for permanent buildings is 5 percent. Inventory accounting method. Both FIFO and average cost methods are allowed for inventory accounting. Once a method i s chosen, it must be used consistently and continuously. The LIFO method is not allowed. Loss carry-overs. Losses incurred inany given year may be carried forward 5 years. No carry-back i s allowed. Taxes on investment income. Dividends are subject to a withholding tax, which i s 15 percent for residents and creditable against overall tax payable. Inter-company distribution i s exempted from further taxes. Interest incomes are subject to a withholding tax of 15 percent. For resident individuals, this tax i s final. For resident companies, this tax i s creditable against the corporate income tax due. For individuals whose dividends andor interest incomes do not exceed the annual personal allowance, they are exempt from withholding taxes. For non-residents, the withholding tax rate on both dividends and interest i s 20 percent. Capital gains are taxed as ordinary income while capital losses are deductible for purposes of the income tax. The property tax. A property tax i s leviedon "fair-market sales value." With effect from the 1994 fiscal year, the effective rate i s 0.2 percent for buildings with an assessed value of IDR 1billion or more (see IBFD, Indonesia, page 70, for more details). There i s also a 5 percent tax on acquisitions of real property, which took effect in 30 June 1998. 2. TransactionTaxes The general VAT rate is 10percent. All taxable enterprises can claim a monthly refund, provided that the input VAT i s higher than the VAT payable. Furthermore, VAT in respect of capital goods imported or delivered on or after 9 March 1998, being machinery and factory equipment required for the process of producing taxable goods, i s borne by the government. This incentive covers VAT otherwise payable on the importation andthe purchaseof these items. 105  Import duties are levied on either an ad valorem or specific basis. The average import tariff for machinery i s under 5 percent. However, by a Decree of 4 January 1999, both foreign and domestic investment enterprises are exempt from import duty on import of machines plus spareparts and components. 3. Tax Incentives The h o m e Tax Law of 2000 introduced several incentives, of which the one most relevant to our METR analysis is the accelerated depreciation allowance available to enterprises located in the development zones. This accelerateddepreciation allowance doubles the regular rate for buildings and each of four groups of machinery and equipment. As a result, the depreciation rate for manufacturing buildings i s 10 percent using the straight-line method, and that for machinery and equipment 18.75 percent, or 37.5 percent usingthe declining-balance method. This accelerateddepreciation allowance may be carriedforward for up to 10years if not utilized. Other incentives include an extended period for carry-forward of losses (up to 8 years) applicable to less developed provinces and certain resource industries, and reduction of withholding tax on dividends to 10 percent unless a lower rate applies according to a treaty. There i s a conflict in information on tax holidays on the same page of IBFD- Indonesia document (page 83). It states clearly, in Section 48.1, that "there are no full tax holidays under Indonesian tax law (2000) although a reduction of net taxable income i s available depending on the amount invested." And then it says, inSection48.3, that "by a Presidential Decree publishedon 26 January 1999, Indonesia will offer tax holidays of up to 8 years for new projects of 22 categories of industries" that include many manufacturing sectors. Since the 2001 edition of Ernst & Young's Worldwide Corporate Tax Guide does not indicate any provision for tax holidays in Indonesia, the IBFD's contradictory information on tax holidays i s ignoredinour METR analysis. SINGAPORE 1. CapitalTax Provisions The corporate income tax rate The corporate income tax rate i s 22 percent. As of 1 January 2003, the corporate tax structure was changed; tax i s charged on adjusted net profits less depreciation allowances but excluding all dividends paid, which are now tax exempt. The corporate tax rate has declined steadily from 26 percent at the time of the last review in 2000 to 22 percent from 1January 2003; there i s a partialtax exemption for 106  announced that the corporate income tax rate will be reduced to 20 percent with effect from Year of Assessment 200570. , The tax depreciation rule There are several allowances available to taxpayers. The first i s the initial allowance, which i s 25 percent for industrial buildings and 20 percent for plant and machinery. The second is the regular annual allowance which i s 3 percent for industrial buildings and varies for plant and machinery according to the type of assets. The third i s the accelerated allowance for plant and machinery, or a 3-year depreciation allowance, at a rate of 331m percent; taxpayers opting for this accelerated allowance, however, are not entitled to the initial allowance. Finally, a 100 percent allowance i s provided for computers and certain automation equipment. (Since there is no information available on detailed classifications for regular annual allowances for plant and machinery, the 3-year depreciation allowance is usedinour METR analysis.) Znventory accounting method Ordinary commercial and accountancy principles applicable to inventory valuation are accepted by the Comptroller. However, the LIFO method i s not acceptable. A change in the method of inventory accounting method may be made provided that the principles for ascertaining true income are followed to the satisfaction of the Comptroller. Loss carry-overs Business losses other than capital losses may be carriedforward indefinitely. Taxes on investment income Singapore had operated an imputation system regarding dividend taxation. From 1January 2003, the imputation system for corporate taxation was replaced with a one-tier system. As a result, the tax collected from corporate profits i s final and all Singapore dividends, whether from regular taxpaying firms or tax- exempt firms, are exempted from further taxation. Interest i s taxed as ordinary income. There is no withholding tax on interest received by residents. The withholding tax on non-residents i s 15 percent. There i s no capital gains tax. 2. Property Tax Property tax i s levied on all land and buildings.The basis of taxation i s the gross rent at which the property can reasonably be expected to be rented on an annual basis. The general property tax rate i s currently 10 percent, with a lower rate of 4 percent for owner-occupied residential properties. Singapore also levies estate duties (i.e. property transfer tax) at 5 percent for every dollar of the first S$12 million and 10 percent for 'O This information is provided by the Ministry of Finance in Singapore. Recent changes in both the personaland corporatetax structure appear to be inresponseto recommendationsby the Economic Review Committee. The Committee recommendedreductions in the corporate tax rate from 24.5% to 20% and the top personal tax rate from 26% to 20% within three years to enable Singapore to compete with other countries inthe region. (httdlwww. wto.ordenrrlish/tratop e l m elm229 e.htm 107  every dollar above S$12 million. Certain exemptions from payment of estate duties are provided such as that for dwellings usedsolely for residential purposes up to a value of S$9 million. 3. Transaction Taxes The goods and services tax (GST),a value-addedtax inSingapore, i s 5 percent on all taxable items except zero-rated items such as the export of goods, "international" services (mainly transportation and telecommunications services provided abroad), and trustee services to non-residents. Input tax credit applies generally except for exempted sectors such as the financial services sector. Import duty i s levied only on a limited number of goods listed by the Minister of Finance. Imported machinery and equipment are generally duty free unless they are "obtainable locally."71 4. Tax Incentives The main incentives are income tax relief through tax holidays and subsequent tax reductions granted to "pioneer companies." Tax reduction for a limited number of years i s also available to export manufacturing companies. The alternative to these "pioneer companies" incentives and export incentives i s the generous Investment Allowance. Pioneer companies A 5- to 10-year tax holiday i s available to a pioneer company. In computing exempt pioneer profits, deductions in respect of capital allowances on depreciable assets are mandatory. However, firms may ignore this mandatory claim for capital allowances if more than 50 percent of their company's paid up capital i s held by permanent residents in Singapore, or their fixed capital expenditure i s above S$150 million, or they receive an approval by the Minister. Where mandatory claim for capital allowances are ignored, the assets acquired during the pioneer period will be deemed to have been acquired on the day immediately following the end of the pioneer period. After the tax holiday ends, the "post-pioneer companies" may enjoy a reduced income tax rate of no less than 10 percent (usually 15 percent) for a period of 5 years. Export manufacturing companies Qualified export manufacturing companies72 may enjoy a tax holiday, inrespect 90 percent of export profit, for a maximum period of 5 years. If the company i s also enjoying a tax holiday as a pioneer company, the total tax relief period, inclusive of the 5-year pioneer period, i s 8 years. However, the total relief period can be extended to 15 years if the fixed capital expenditure i s above S$l,OOO million, or more than 50 percent of the paid-up capital i s heldby Singapore residents and the total capital expenditure i s above S$ 150million. "Singapore'scurrentMFNtariff, basedontheHS2002nomenclature,consistsof10,689linesattheHS-8 digit level. Since Singapore's previous Trade Policy Review, there have been no changes in its applied MFNtariff, which, with the exceptionof six items, is zero. The six items are stout andporter, beerandale, and medicated and non-medicated samsu. These six products subject to duty are both imported and manufacturedlocally. (httv://www. wto.ordendish/tratov e/tur e/tv229 e.htm) 72 To qualify, a manufacturing company's export sales must reach S$ 100,000 and account for over 20% of the total sales inthe given year when the tax relief commences. 108  Investment allowance The Investment Allowance, as an alternative to the above two incentives, is an additional allowance over and above the normal capital allowance and 100 percent of the fixed capital expenditure incurred i s eligible in calculating the allowance. That is, a 100 percent Investment Allowance, combined with the regular annual allowance, will provides a 200 percent write off of capital expenditure against taxable income. The most commonly granted Investment Allowance, however, i s 30 percent. Also note that, as under the above two incentives, dividends paid out of an "investment allowance account" are exempt from tax in the hands of the recipient shareholders. It is noteworthy that an official evaluation indicates that "it is not clear whether the tax incentives add anything to Singapore's already attractive status as a relatively stable economy and therefore a good destination for foreign inve~tment."~~Singapore seems to be moving towards a lower corporate tax rate (which would in itself suggest a lowering of the value of incentives to potential investors). In addition, some incentive programs have been discontinued as a result of a post implementation review system, Recent examples include: the Hotel Refurbishment Scheme (removed in July 2003) and Tax Deduction for Expenditure on Energy Conservation (removed in November 2003) under the Income Tax Act; Production for Export under the Economic Expansion Incentive Act is due to be removed sometime in 2004. It is unclear whether the overall number of programs and the revenue forgone has increased or decreased as a result. Singapore maintains that the objective of its tax incentive programs i s mainly to address market failures (e.g. in the area of promoting R&D) and to promote capability development incompanies." (Refer to http://www.wto.org/entZlish/tratop e/tpr e/tp229 e.htm) THAILAND 1. CapitalTax Provisions Thecorporate income tax rate The corporate income tax rate is 30 percent on net profit. However, reduced ratesof 20 percent and 25 percent are available to small and mediumenterprisesaccording to the size of taxable income. A 5 percent gross-receipts tax applies when the taxpayer fails to file a tax return for CIT or provide adequate financial/accounting documents. The tux depreciation rule The main feature of the depreciation allowance in Thailand is that "the rate adopted must be the same for accounting and income tax purposes. No deviation from or reconciliation between the depreciation recorded in the accounting books and the depreciation allowance for income tax purposes i s allowed." 73 Corporate tax revenue, as a percentage of total tax revenue is relatively high, in Singapore by international standards, some 32%, suggestingthat the incentives may not be as generous as they appear. 109  The official depreciation rate for tax purpose is 5 percent for buildings and 20 percent for machinery and equipment. These rates correspondto the official useful life of the depreciable assets. Accelerated depreciation allowances including investment allowances are available to small and mediumenterprises. Znventonl accounting method Either FIFO or LIFO may be used for inventory accounting as long as it i s followed consistently. Loss carry-overs Operating lossesmay be carried forward for five years. No carry-back i s allowed. Taxes on investment income Both dividends and interest incomes are taxed as ordinary income but a withholding tax applies to both incomes. The withholding tax rate on dividends i s 10percent and that on interest 15 percent. Capital gains are taxed as ordinary income at a rate of 30 percent. The property tax The property tax is 12.5 percent based on the annual (rental) value of the buildingstructures and land. There i s also a transfer tax on immovable property. The tax rate i s one percent of the sale price. 2. Transaction Taxes The general VAT rate i s 7 percent. Import duty varies by taxable item and the tariff system i s quite complicated. The overall simple applied MFN tariff average i s 14.7 percent and that for machinery and equipment i s under 10percent (http://www,wto.orden&h/tratop e/tpr e/t~223e.htm). 3. Tax Incentives The Board of Investment (BOI) is empowered to grant tax incentives to a promoted investment under the Investment Promotion Law (IPL).Major tax incentives include CIT exemption for from 3 to 8 years and import duty reduction or exemption. A "promoted investment" i s identified by its product, and the business location. For this purpose, the nation is divided into three Zones. Developments since 1999 in the BOI's provisions include the abolition in August 2000 of local-content requirements and export requirements. In August 2000, the BO1 abolished certain eligibility requirements for the reduction or exemption of import duties74.As a result, BOI-approved projects are granted either exemptions (for projects located in Zone 3) or a 50 percent reduction (for those in 74These requirements were that investment projects locate their factories in industrial estates or promoted industrial zones, or export at least 80% of total sales. 110  Zones 1or 2) of import duties on certain machinery, without any requirement to export or locate factories inindustrialestates or promotedindustrialzones.75 Corporate tax exemptions continue to be accorded depending on the zone in which an enterprise i s located.76A three-year corporate tax holiday (exemption) is granted for projects inindustrial estates or promotedindustrialzones inZone 1.InZone 2, the exemption i s for three years or five years (previously seven years), if projects are located in industrial estates or promoted industrial zones.77In Zone 3, an eight-year corporate tax holiday i s granted, regardless of the location, capped at 100 percent of investment capital.78Since December 2001, the amount of the corporate tax exemption has been capped at 100 percent of investment capital (excluding the costs of land and working capital).79 The BO1 selected five targeted industries: agri-business, automobiles, fashion, electronics, and high-valued services ; these industries appear to be eligible for exemptions of import duties on certain machinery and a corporate tax holiday for eight years, irrespective of the location of the project.8' Effective from 19 September 2002, zoning requirements for various promoted activities were relaxed: e.g. projects previously required to locate in either Zones 2 or 3 may relocate to zone 1. Certain export requirements tied to incentives granted to those located inEPZs were also eliminated on 1January 2003. 75 The measureapplies to machinery with an import tariff of 10%or above. 76 Zone 1 includes Bangkok and five provinces (e.g. Samut Prakan); Zone 2, twelve provinces (e.g. Ang Thong); andZone 3 (all 58 remaining provinces). InAugust 2000, two provinces were addedto Zone 2 and the 58 provinces in Zone 3 were divided into two areas based on development level of each province; incentives for the 40 more developed provinces are now more limited in scope than for the 18 less- developedprovinces inZone 3. 77 Projects submittedbefore end 2004 and located inindustrial estatesreceive a seven-yeartax holiday. 78 Extra incentives apply incertain Zone 3 provinces; a 50% tax holiday is available for five years after the eight-year exemption period; water, electricity, and transportation costs are deductible from taxable income for ten years from the date of first sales, meaning that these costs can be deducted twice; and 25% of the installation and construction costs of infrastructure facilities can be deductedinthe calculation of net profit. Since August 2001, projects with capital exceedingB 10 million must obtain IS0 or similar international standard certification within two years; otherwise the tax holiday is reduced by one year. 79 According to an amendmentto the Investment Promotion Act. 8oBO1Bulletin No.68/45, 10April 2002 "BO1Targets Five Industries For Proactive MarketingCampaign" Available at <http://www.boi.go.th/english/focus/boi_pr_target~industriesO2.html> These incentives also seem to apply to enterprisesengaged in other priority activities. Priority activities are: agriculture and agriculture products; technological and human resource development; public utilities and infrastructure; environmental protection; and targetedindustries. 111  ANNEX 2 ANNEX2 SECTION A: DATA STATISTICS AND It is useful as a reference point to study the descriptive statistics available from the data sources made available to us. This gives an appreciation of the issues as well as the constraints the data place on the depth of the possible analysis. We have two primary sources of data. The first is Datastream from Thomson International, which i s made available by the LSE. The second source of data i s the Census of Production Data (henceforth Census data) which was provided by the Malaysian Department of Statistics. 1. Datastream Figure A2.1: No of Firmsreporting by Years of Continuous Reportingon KLSE-Data from 4 5 6 7 8 9 10 11 12 13 No of years ofcontinousreporting Datastream only gives data for Zistedfirms on the KLSE. This data i s at Firmlevel and i s available in panel format for roughly 10 years till 2003. The industry classification follows the KLSE convention. Figure A2.1 describes the evolution over time of the reporting of R&D, both in absolute numbers and in the percent of firms on the KLSE main board. Both of these show an upward trend. This trend i s mirrored in more developed countries and indicates the growing importance of R&D activity inthe eyes of shareholders. 112  Figure A1 also shows that while growing, there are still a relatively small number of firms that report R&D. This proves to be a severe constraint on any time series econometric work, limiting much of the analysis to a simpler static cross sectional version. Any attempt to do time-series studies on this data run the risk of being driven by entry and exit of firms rather than true underlying changes in the behavior of the firms. This problem can be alleviated by using a balanced panel of firms, but again due to the small number that report continuously; this is not statistically robust. Figure A2.2: Number and Percentageof Firmson the KLSEreporting R&D 50 I 1 8 45 1 40 .-E35 / / //f 30- L 25 Noof Firms 0 .-c + 0 20 / - 15 3 2 Y lo 5 ~ Year The data does however lenditself to describing changes and gives a good idea of how R&D doing (or expensing) patterns change. We report the changes in the R&D intensity (R&D/ Sales) here, arguing that the R&D intensity rates are more comparable across time than absolute values of R~LD.*~ Figure A2.3 below shows this. The contraction of R&D intensity in 1998-2000 i s presumably due to the slowdown causedby the Asian Crisis. The change in the R&D intensity is less subject to bias due to entry and exit. For this reason changes in the R&D intensity are more likely to reflect the true macro behavior of R&D doing firms. We calculate the growth rate of R&D intensity as C(RDi/Salesi), - C(RDj/Salesj),l for t=1998-2003. iare all the firms in period t andj all firms inperiod t-1. 113  Figure A2.3: PercentageAverage ChangeinFirmR&DIntensity by PreviousYear on KLSE 006 1 0.04 0.02 2gP o 5 -0.02 ' -0.04 -0.08 -0.08 2. Census of Production Data (Department of Statistics) This is provided at the establishment level (Le. not firm) but only at a cross section format for the year 2000. It i s also limited to the manufacturing sector. The census data provided from the DOS has the advantage of sampling virtually the entire population of establishments in manufacturing in Malaysia. The industry classifications follow that of the Malaysian Industrial Classification 2000. Descriptions of the Census data are given typically at the 3-digit level but we have aggregated up further and describe at 2-digit level in our analysis. There are no identifiers which would enable us to aggregateup to firm level, or by region. Table A2.1 below gives an indication of the intensity of R&D, where intensity i s measured as (R&D Expenditure/ Sales). None of these figures are out of the ordinary, reflectingOECD figures by and large. 114  Table A2.1: R&Dintensity by Industry.NoteMIC (2 digit) definitionsare assignedby authors 0.146 Motor Vehicles 89822 Furniture,all other 5377 Additional descriptive statistics are provided from this data elsewhere in the report. It i s worth mentioning that this sort of data and the skill of the DOS are highly valuable and should be more actively used in monitoring and any future studies of this kind. 115  ANNEX2 SECTION B:METRMETHODOLOGY This section lays out the methodology used to calculate the impact of the tax system on the cost of capital - applied to R&D in this chapter, but applicable much more broadly. The methodology was developed by King and Fullerton (1984), and used in OECD (1991), Bloom, Chennells, Griffith and Van Reenen (2000) and Bloom, Griffiths and Van Reenen (2001). The idea is to derive a single quantitative measure to summarize the effects of a complex set of tax regulations of the costs of performingR&D. This i s the pre-tax real rate of return on the marginal investment project that i s required to earn a minimumrate of returnafter tax. To focus on tax effects we will ignore changes inother variables such as inflation and interest rates, but in reality clearly these will also play a role. The methodology considers investment funded in a variety of ways (debt, equity etc.) and by a variety of individuals (higher rate income tax payers, tax exempt institutions, internal cash flow etc.). Typically R&D, like physical capital, i s funded from retained earnings, and so i s equity financed, while the typical investor i s institutional and so pays no additional income tax. For this representative project consider an investment project with a present discounted value of V, where V=- P(1- t) r + d where p i s the rate of return, t i s the statutory tax rate, r i s the firm's cost of capital (discount rate) and d is the depreciation rate. The firms discount rate will depend on the form of finance used (debt or equity), with a standard measure being the weighted averagecost of capitals3. The initial cost of the investment, net of tax, i s unity but after tax i s defined by C, where C =1- A =1- Ad -AC where A is the present discounted value of deprecation allowances and tax credits on the asset. This will be a combination of the physical capital PDV of depreciation rates Ad and the R&D capital PDV of depreciation rates Ac. For physical capital investment this depends on the deprecation schedule of investment intax coding, so that where Dk i s the deprecation allowance in year k (Le. for a 5 year straight-line depreciation rate this would be 0.2 per year). For R&D capital all allowances inMalaysia are immediate so that Ac = Scxtx DC +(1 -Sc)XtXDk where Sc i s the rate of current R&D in total R&D (typically around 90 percent in OECD countries), Dc i s the rate of R&D deduction on current R&D (200 percent under Double 83This i s commonly called the WACC - see Brealey and Myers (2004). The WACC takes the averagedebt and equity discount rates and weights by the share of these in total firm financing. The finance directors in all large firms should know their firm's WACC since this number is required inthe planning and evaluation of all investmentprojects. 116  Deduction), t i s the tax rate as before and Dk is the rate of additional deduction for R&D capital spending (100percent for IAs). For a marginal investment project the value of the project will just equal the cost of the investment project, so that p(lWt)= l - A d - A c r + d which can be re-arrangedto p = ( r + d ) x1-Ad-Ac (1-1) Inwords, this says that the pre-tax rate of return (p) needs to equal the firms cost of capital (r)plus deprecation, multiplied by the ratio of depreciation allowances divided by (1-tax rate). Inthe absence of any taxes this equation for marginal investment would take the form p = ( r + d ) so the second term 1-Ad-Ac can be called the "tax wedge", or the impact on taxes on (1-t> firms required rate of return. If this i s greater than 1 (it i s about 1.05 in OECD countries for physical capital) then it puts up firms required rate of returns from any investment project, while if it i s less than 1 (it is about 0.9 for R&D in the UK) it puts down the required rate of return. Hence, the entire impact of the tax system can be captured in one number - the impact of taxes on firms pre-tax required rate of return (their pre-tax cost- of-capital). B y examining the impact of taxes on this tax wedge economists can evaluate the impact of differential tax systems on the marginal incentives to invest (see OECD 1991). Of course this assumes all firm activity is driven by marginal activity, and in smaller more open countries like Malaysia much FDI investment i s driven by average tax rates (because projects are all or nothing decisions). As Devereux, Griffith and Simpson(1999) have'shown this can be important for driving the location decisions of FDI, so in the Malaysian case average effective tax rates will also be important. However, noting this issue we shall focus on marginal effective tax rates. 117  ANNEX 2 SECTIONc:TAXEXHAUSTION, PROFITS,AND COSTING There can be two scenarios with regard to R&D and Profits. The first i s when Profits are greater than R&D expenditure and the second where Profits are less than R&D expenditure. These two cases produce very different results when considering the cost of the DD and also the incentive effects of the tax credit. We will consider each in turn and assume a straight volume credit i s used. 1. Profit > R&DExpenditure FigureA2.4: ShowingProfits>R&DExpenditure 1 7 Potential Incentive I Profits R&D Exp Figure A2.4 indicates the `usual' situation where a firm makes profits greater than its R&D expenditure. The key points to note are as follows. The Loss of Revenue to the government i s the same as the R&D expenditureg4.This i s because the DD can be totally offset against the Firm's profits. Inthis case the Firmis not tax exhaust. The second point to note is that, ifthe government increasedthe tax credit, say that it became a triple deduction, this would provide an incentive for the firm to do more R&Dg5.This is because it would still be able to claim the increased R&D expenditure against its profits. A subtle point i s revealed here, and that is, any increase in the tax credit only provides a greater incentive to firms if they are able to deduct it against profits. 84Strictly speaking the loss of revenue is the (R&D Expenditure) x (Corporation Tax). The diagram is drawn to highlight the idea we discussand for ease of exposition we drop the CT calculation. 85Assuming Profits remained greaterthan R&D Expenditure. 118  2. Profit c R&DExpenditure Figure A2.5: Showing Profits< R&DExpenditure No Incentive DD Cost I Profits . R&D Exp Figure A2.5 indicates the more unusual situation where a firm i s spending more on its R&D expenditures than it makesinprofit. This canbe the case for small start-up firms. The points to note inthis case as are follows. If the firm was doing a given amount of R&D activity such that Profits<R&D expenditure, the addition of the DD provides no extra incentive to the firm to carry out any additional R&D. However, it nonetheless means that the cost of the DD for the existing R&D i s borne by.the Government. This means the cost to the Government i s equal to the profitE6the firm makes but the DD provides no extra incentive. The entire DD expenditure i s therefore a dead weight loss in this case becausethe incentive has no effect at the margin. The second point in this case i s that if the Government raised the rate of tax credit, again say to a triple deduction, this would not change the amount of revenue lost to the Government and neither would it change the marginal incentive to do more R&D activity.87This i s because the R&D amount i s still greater than the profits the Firm makes. This means the Firmi s tax exhaust. What is however interesting i s that if the profit of the firm increased (but remained less than R&D expenditure), the revenue loss to the Government via the DD would increaseas Figure A2.6 below indicates. 86Hence claimable R&D is less than R&D expenditure. 87Assuming the R&D expenditure remained greater than profits. 119  Figure A2.6: Showing Profits< R&DExpenditure but with increased profits _- 3 No Incentive DD Cost Profits R&D E X ~ The important point here i s that as the firm profits increase, the cost of the DD to the government rises, even though no extra incentive i s being given to the firm to do more R&D activity. Only when Profit >R&D expenditure does the marginal incentive start. 120  ANNEX2SECTION D:R&DINCENTIVES IN EAST ASIA KOREA Korea i s the most successful country among the comparators, and among the most successful in the world, in terms of stimulating private sector spending on research and development. The success of its strategy, particularly in the case of high profile companies in the DRAM, CDMA, flat-panel display, automotive and petrochemical industries, is now globally evident both to the average consumer and to the observer of international patent filings. Korea does not, however, offer generous tax incentives relative to Malaysia specifically focused on increasingR&D expenditure. Korea has simplified its tax policies to minimize the number of discrete tax incentives. Following its opening to FDI, Korea offers generous incentives for firms locating in its targeted industrial'zones or in high technology sectors. Such firms are offered a seven year tax holiday, plus an additional three years at 50 percent of the normal tax rate. Research and development expenditures are depreciated on a straight- line basis over five years. Rather than fiscal incentives per se, a combination of sustained spending on tertiary education and research infrastructure appears to have contributed to the performance. There are over 4,800 research and development institutions in Korea, the third-highest mobile phone and highest broadband internet penetration inthe world. TAIWAN, CHINA Taiwan has been extremely successful in establishing research and development as a cornerstone of its industrial strategy, yet tax incentives have been a relatively minor part of its overall strategy. Taiwan maintains a corporate income tax of 25 percent, tax incentives that appear less generous than those of Singapore, Malaysia, and Thailand. However, it maintains significant grant-basedschemes that leverage a human capital base that has the highest concentration of researchers per capita, and the world's highest ranked student population in terms of science and math scores. Taiwan i s further noted by the quality of its public institutions, which draws upon international network of overseas Chineseto propel its technology strategy. Tax Incentives The Statute for Upgrading Industries grants investment incentives, accelerated depreciation, investment tax credits and tax holidays to firms that meet criteria. In addition to the Statute for Upgrading Industries, the Statute for Development of Small andMedium-Scale Enterprises grants similar incentives to SMEs." 88PriceWaterhouseCoopersinvestment guide. 121  0 Tax holidays. In a scheme operating similar to Malaysia's Pioneer Status, enterprises in emerging strategic industries are eligible for a five year tax exemption. 0 Deductibility of research and training. Companies may deduct 5-35 percent of the amount of their investment in R&D or personnel training from their income tax over a five-year period beginning with the year in which the investment i s incurred. To encourage increasing levels of spending, companies may deduct 50 percent of the amount of their investment in R&D or personnel training that exceed the average annual amount of their investment in R&D or personnel training for the previous two years from income tax. The total amount deducted from tax due per year under the previous two items may not exceed 50 percent of income tax per year. 0 Accelerated depreciation. Equipment and facilities used exclusively for R&D, experimentation, and quality control purposes are eligible for accelerated depreciation over two years. If there i s any residual post-depreciation service life remaining following the accelerated depreciation period, depreciation may be continued for one or several years within the service life of the assets as specified inthe IncomeTax Law untilthe assets are fully depreciated. Non-tax Fiscal Incentives Beyond tax incentives, Taiwan's subsidy programs are robust. It co-invests with the private sector up to 50 percent of the value of projects in strategic sectors, provides a number of science parks and institutions that support understanding and sourcing foreign technology, and interest free loans through the Executive Yuan Development Fund. 0 Industrial Development Bureau (IDB) provided direct subsidies from 1991 to 2001 to 451 R&D projects totaling NT$8.18 billion (US$233.71 million). 0 Venture capital. Taiwan also provides government funds to support private funds investing in venture capital and facilitates listing of such firms to list on stock markets, and supports technical services that support R&D. SINGAPORE Singapore offers a generous set of tax incentives, but increasingly, the most advantageous aspects of its environment for research and development i s its low corporate tax rate and its human capital stock. The corporate tax rate was reduced to 22 percent for the 2003 assessment, and will be reduced further to 20 percent for the 2005 tax year. With over two thousand scientists per million inhabitants, Singapore has a relatively high density of researchers capable of supporting a response by the private sector to the incentives. The incentives themselves cover both technology and human resourcedevelopment. 0 Double Deduction for R&D Expenses. Singapore offers a double deduction on R&D expenses similar to that of Malaysia. The DD can be used for in-house R&D expensesor R&D done by anR&D organization inSingapore. 122  R&D and Intellectual Property Management Hub scheme provides tax exemption for a fixed period on qualifying royalties or interest remitted to Singapore from foreign sources, so long as the money is used for further R&D expenses over the five years following the revenue, and so long as the IP on which royalties accruei s owned and managedby the applicant. Accelerated depreciation of intellectual property allows companies to write down its intellectual property (IP) acquisition costs and patents over five years on a straight-line basis, so long as the intellectual property has economic application. Depreciation of Cost Sharing allows companies to write down cost-sharing payments for economic R&D activities, which are otherwise not deductible, over one to five years. Cost-sharing agreement must be bona fide. Royalty payments to non-residents are exempt from withholding tax. To qualify, technology or know-how acquired must be more advanced than the average prevailinginSingapore. Innovation Development Scheme (IDS) i s a grant that provides partial funding of labor. IP and professional service costs to support activities that demonstrate innovation or leadto development of innovativecapabilities. The Patent Application Fund Plus (PAF Plus) scheme reduces the cost of applying for patents for inventions that are not already receiving government assistance. The Research Incentive Scheme for Companies (RISC) i s another grant-based programthat pays for training, equipment, IP, and professional services related to increasing the hiring and training of researchscientists and engineers. Start-up Enterprise Development Scheme (SEEDS) i s a matching grant program for start-up companies that have a third-party investor. Investors must have proven relevant managementhusiness experience. Technology for Enterprise Capability Upgrading (T-UP) program provides partial funding for firms who hire, on a secondment, research scientists and engineers from research institutes to local enterprises for two years. Open to Singaporeancompanies with minimal 30 per cent local equity. THAILAND Thailand has been the least successful in stimulating research and development spending of the comparator set, with an estimated R&D per capita of around 0.16 percent. This i s certainly not due to lack of incentives, as Thailand offers incentives as generous of those of Singapore and Malaysia. These incentives include Board of Investment tax holidays that appear similar to MIDA Pioneer Status, accelerated depreciation schedules, Double Deduction, and duty exemption on imported equipment. Tax Incentives e Board of Investment provides tax holidays for investment in R&D activities, scientific laboratories, calibration services, educational and vocational training institutes. These are typically negotiated with BO1within an indicative range. 123  0 The Ministry of Finance offers accelerated R&D machinery equipment depreciation (40 percent infirst year). 0 The Revenue Department offers 200 percent deduction of R&D expense for projects certified by the National Science and Technology Development Agency (NSTDA),includingthrough anR&D subsidiary or research that i s contracted out to independent, government, or university researchorganization. 0 Customs Department offers exemption of duties for imported R&D equipment. For example, scientific tools reduced from 30-40 percent to 5 percent, computer partsfrom 20-40 percent to one percent. 0 Non-tax Fiscal Incentives In addition to tax incentives defined above, all of which are not significantly utilized, there are a number of non-tax incentives. The Ministry of Science and Technology offers a revolving fund from which R&D soft loans are offered. NSTDA also operates an innovation fund, recently moved under the National Competitiveness Committee, as well as researchgrants and a Company Directed Technology Development Program. CHINA China provides a range of fiscal incentives and tax incentives, but these are typically determined at the local level. These include subsidized interest rates, direct support to enterprises, expenditure on technology innovation through grant and state-funded research, and a range of technology funds. According to local entrepreneurs, these programs also suffer from various shortcomings, because they are not transparent, information i s not publicized, many firms are unaware of the incentives. SMEs complain that the process of applying for support i s complicated and time- consuming and that approved and promised support may fail to materialize, since the application criteria are not standardized and official administrators often lack expertise. Furthermore, there i s little performance monitoring. Since 1979, Chinahas introduced a succession of programs to provide tax preferences in order to encourage business investment. Inmany cases, the intended beneficiaries are foreign-invested enterprises (FIEs). B y 2002, the authorities had designated 110 regions or zones for tax incentives. In many cases, corporate tax rates are reduced from 33 percent to 15 percent. As part of the program for Western region development, domestic and foreign enterprises in specified industries are reduced to 15 percent for the period 2001-2010. Inaddition, tax holidays are given to FIEs and domestic companies in certain industries andinethnic minority areas. 0 A reducedincome tax rate of 15percent may be possible for technology-intensive or knowledge-intensive projects; and enterprises with foreign investment recognized as new or high-technology enterprises established in the new and high-technology industrialdevelopment zones established by the State Council; 124  a enterprises with foreign investment recognized as new-technology enterprises established in the BeijingNew Technology Industrial Development Experimental Zone. a Ifestablished in one of the preferential zones, enterprises of a production nature that are engaged in technology- or know-how-intensive projects may enjoy a reduced enterprise income tax rate of 15 percent. At the state level incentives can vary widely. Some states offer more generous incentives for firms that derive income from technical transfer, R & D, consultancies, technical service of foreign-invested enterprises, R & D institutions subject to the approval of the StateTaxation Bureau. Some states offer a rolling incentive, in which income tax can be levied at 10 percent of the normal rate on the incomes of foreign enterprises from their provision of patent technology covering areas of R & D and anumber of strategic sectors. Incase that R & D expenses of foreign-invested enterprises increase by over 10 percent from the previous year, 50 percent of the actually incurred R & Dexpenses of a same year i s allowed to off- set the corporate income tax of the year. Tax administration in China is changing. China has separated local and central government taxes and tax administration in 1994such that the central tax bureau collects VAT and excise taxes, while the local tax bureaus are responsible for business income and other taxes. However, China is likely to overhaul its tax system over the next few years with a new division of labor that may not be as decentralized. 125  ANNEX2SECTION E:INTERVIEW SCHEDULE As part of the World Bank Fact Finding Mission, we spent three days interviewing various officials in companies, government organizations, think tanks and other relevant bodies. The table below gives a summary of the organizations interviewed. These interviews typically lasted 1-2hours each. Table A2.2: Summary of Interviews Sector Size Interviewee Comment Automotive Small CEO R&D Company Chemical Medium R&D R&D Company Manager, CEO Company Filter Listed R&D Manager In-house R&D Manufacturing 2515 Company Automotive Medium General Contract R&D Prototyping Manager 2515 Company Bio-Tech Listed Research R&D Comoanv Director Contract R&D Various Combined Kick-off inance+ lnlan meeting with MOF RevenueBoard and questioning IRB Deputy Directors Assistant Officer Deputy Directors, Unit Heads Deputy Director and others Senior and Visiting Fellows Directors and Partners Director and Partner 126  ANNEX 3 ANNEX 3 SECTION A: DESCRIPTIVESTATISTICS: INVESTMENT FUNCTION Table 1.DescriptiveStatistics All Firms Large-ScaleFirms Small- andMedium-ScaleFirms Numberof observations 563 169 394 Percentage 100 30.02 69.98 Numberof incentiveusers 316 116 200 Percentage 56.13 68.64 50.76 Numberof firms withinvestment>O 430 133 297 Percentage 76.38 78.70 75.38 Means Investment /Capitalt.I 0.176 0.164 0.182 (0.254) (0.228) (0.264) (ValueAdded - Value Addedt-l ) / 0.169 0.142 0.181 Capitalt-I (0.169) (0.944) (1.118) (Profit/Capital)t.I 0.169 1.292 1.750 (1.613) (2.081) (2.056) ln(Size)t-I 4.225 5.123 3.840 (1.283) (1.093) (1.161) FirmAge t-l 12.901 12.574 13.041 (1.283) (8.529) (8.082) (Long-termDebt/Capital),, 0.753 0.767 0.747 (0.950) (1.O14) (0.923) (Long-termDebt/Capital)zt-l 1.469 1.611 1.407 (3.518) (3.895) (3.347) Notes : Standard deviation are denotedinparentheses. A large-scale (small- and medium-scale)firm is defined as a firmwith total equity greater (less) than RM2.5 million. Valueadded i s calculatedas total sales minus direct materialcost. Capitali s given by net book value of new machinery and equipment. Size i s equal to the total number of permanent employees. Long-term debt is equal to domestic cunency denominated long-termliabilities. * denotes significance at 10%level; I,denotes significance at 5% level; **.denotes significance at 1%level. 127  ANNEX 3SECTIONB: PROPENSITYSCOREMATCHINGS9 1. CausalInference through StatisticalMatching Let Yil be the outcome in the "treated" state and Yio be the outcome in the "untreated" state for firm i. Inthe application presentedinthis chapter, Yil corresponds to the potential investment rate associated with using the reinvestment allowance/accelerated capital allowance and Yiocorresponds to the potential investment rate associated with not using the reinvestment allowance/accelerated capital allowance. The outcomes are called potential investment rates because only one of (Yil, Yio) is observed for each firm, given the cross-sectional nature of the PICS. Let Di = 1indicate that a firm used the reinvestment allowance/accelerated capital allowance and Di = 0 indicate that a firm did not use the reinvestment allowance. For an firm i, the actually observed investment rate is therefore Yj = Yjo +Dj(Yj1 - Yio).Finally, let Xj be a vector of observed covariates affecting both the choice of using the reinvestment allowance/accelerated capital allowance and the investment rate of firm i. The parameter of interest - the average efSect of the treatment on the treated - i s the mean effect of using the reinvestment allowance/accelerated capital allowance as opposed to not using the reinvestment allowance/accelerated capital allowance. In terms of the above notation, the parameterof interest is: AT =E(Y,-Y,, IDi =l)=E(Y,,IDi =l)-E(Y, IDi =1). [All This expression measuresthe change inthe investment rate of users that is causedby the fact that the firm used the reinvestment allowance/accelerated capital allowance (Di = 1). The last term inEquationA1 describes the hypothetical average wage ifusers insteaddid not use the reinvestment allowance/acceleratedcapital allowance. Of course, this term is not observable and has to be estimated using a control group of nonusers. However, the average investment rate of nonusers i s typically not suitable since users and nonusers may differ incharacteristics that affect their investment rates, that is: E(Y,, Di =1)f E(Y,, Di =0) I I [A21 Equation 2 states that using the investment rate of nonusers as an estimate for the hypothetical situation in which a user instead did not use the reinvestment allowance i s generally not valid, since both groups differ due to observable and unobservable characteristics, giving rise to selection bias. The treated firms (Le., users) are not a random sample of the population, but may select themselves or be selected on the basis of characteristics that also influence their outcome (Le., their investment rate). In order to make this point clear, suppose that the investment rate Yj is a linear function of a set of observed characteristics Xj and unobservedcharacteristics represented by 4. Di is the binary variable indicating whether the firm i belongs to the treatment Adaptedfrom Black and Smith (2004) andSmithandTodd (forthcoming). 128  group (reinvestment allowance/acceleratedcapital allowance used; Di = 1) or belongs to the control group (reinvestment allowance/accelerated capital allowance not used; Di = 0),that is yi =pzi +mi+ E i . ~ 4 3 1 When selection into the usingand non-using groups i s non-random, selection bias in the estimation of Scan arise becauseof dependencebetween Di and 4. The selection bias i s present if E(& I Di,Xi) # 0, which implies that ordinary least squares regression of Yi on Xi and Di does not yield a consistent estimate of the causal effect Sof Di since E(Yi I Di,Xi) f Xis+ SDi (see Heckman et al., 1997). The decision to use the reinvestment allowance/accelerated capital allowance can be described in terms of an index function framework. Assume that Dt*isan unobserved index of the net benefits of using the reinvestment allowance/accelerated capital allowance and that Di*i s a function of observed variables Xi as well as unobservables variables Ui.The vector Zi may include all of the variables inXi. Assuming further that this function is of the form D* = ziy ui , + the decision to be in the treatment group (reinvestment allowance/accelerated capital allowance user) i s determined by the rule Di = I, if D~*>o 0, otherwise. Selection bias (dependency between Di and 4) arises because of dependency between Zi and 4 (selection on obsewables) and dependency between 4 and Ui (selection on unobsewables). Both cross-sectional matching methods and standard linear regression analyses estimate the impact of a "treatment" under the assumption of selection on observables. That is, both approaches assume that conditioning on an available set of covariates removes all systematic differences inoutcomes inthe "untreated" state between users and non-users. The literature formalizes the selection on observables assumption that justifies matching as the ConditionalIndependenceAssumption (CIA), given by This assumption states that the outcome in the base state (investment rate under nonuse) i s independent of the treatment (reinvestment allowance/accelerated capital allowance use), conditional on some set of observed covariates X. That is, within subgroups defined by X, reinvestment allowance/accelerated capital allowance use is unrelated to the investment rate in the absence of the reinvestment allowance/accelerated capital allowance. It i s important to emphasize that the CIA i s just that, an assumption. 129  Inany given context, it need not hold for any particular set of X available in the data, and it may not hold for any set of X variables inthe available data. Moreover, cross- sectional matching does nothing to account for selection on unobservables, and can even act to increase the bias relative to not matching for certain configurations of the unobservables. The key difference between matching and linear regression is that regression makes the additional assumption that simply conditioning linearly on X suffices to eliminate selection bias. Moreover, matching methods, but not linear regression, rely on the "common support" assumption, that i s Pr(D= 1 X) < 1for all X. I [A61 The support condition states that, for each X satisfying the CIA, there must be some individuals who do not get treated; in the context of this chapter, for each X, there must be some firms who do not use the reinvestment allowance/accelerated capital allowance. If there are X for which all firms use the reinvestment allowance/accelerated capital allowance, then there i s no way in a matching context to construct the counterfactual outcome for these observations. 2. Propensity Score Matching and BalancingTest Matching on X when X i s of high dimension raises the problem of empty cells, that i s the number of distinct vector values becomes very large, and many of the treated observation will have no corresponding untreated observation with exactly the same values of X. One responseto this i s to reduce the dimension of X by reducing the number of matching variables, but this will also reduce the plausibility of the CIA. Instead, Rosenbaum and Rubin (1983) show that the assumption that justify matching on X also justify matching on the probability of treatment, Pr(D = 1 I X),which the literature calls the "propensity score." The intuition behind the propensity score matching i s that subgroups with values of X that imply the same probability of treatment can be combined because they will always appear inthe treatment and (matched) comparison groups in the same proportion. As a result, any differences between subgroups with different X but the same propensity score balance out when constructing the estimates. The theorem presentedby Rosenbaum and Rubin(1983) does not aid inchoosing what variables to include inX, but motivates a specification test for Pr(D = 1 I X),that i s the so-called "balancing test." The general idea i s to test whether or not there are differences in X between the D = 1 and D =O groups after conditioning on propensity score, The test has been implement several ways in the literature. The approach used in the chapter follows Dehejia and Wahba (1999, 2002) and divides the observations into strata based on the estimated propensity scores. These strata are chosen so that there i s not a statistically significant difference in the mean of the estimated propensity scores between the experimental andcomparison group observations within each strata. 130  3. Implementation of the Propensity Score MatchingEstimators For each firm iin the group of users, a (group of) comparable nonusers has to be found. Matches are constructed on the basis of a neighborhood N(ei), where ei i s the propensity score for firm i.It i s assumed that NOi s the number of observations of nonusers and N1is the number of observations of users. Thus, those nonusers who are neighbors to firm i are firms j for whom e, ~N(ej),i.e. the set of workers Aj = I ej E H(ej)}. The effect of treatment (reinvestment allowance/accelerated capital allowance use) for each observation i in the group of users is estimated by subtracting the weighted average of the outcome of nonusers from the outcome of the treatment observation i (see Heckman et al., 1999): j=l Matching estimators differ, especially in the weights w(i,j) E [0,1] with ]=lw(i,j) for the members of the comparison group. 3.1 Nearest-Neighbor Matching Nearest-neighbor matching defines a neighborhoodAi of user i such that only nonuserj i s selectedas a control that i s closest to ei in some metric: where 11 11 is a metric measuring the distance. Equation 14 states that nonuser j with the value of ejthat i s nearest to ei i s selected as a match and i s defined as a control for user i. This selectednonuser is attachedwith the weight w(i,j)= 1. Nearest-neighbor matching can be executed with or without replacement. "With replacement" means that the non-treated individuals can be used more than once. This can improve the matching quality, but it increases the related standard error of the estimated effect. 131  3.2. Calliper Matching Calliper matching (also called radius matching) may be used in order to reduce the risk of "bad matches" (see Cochran and Rubin, 1973). For a pre-specified level of tolerance Y >0, user i i s matchedto nonuserj such that: The corresponding neighborhood i s Ai ={X Illei -e 11) Ifthere arenononuserswithin the intervalY arounduser i, useri is left unmatched andi s not usedfor estimation. 3.3 Mahalanobis Metric Matching Usingadditional variables (a subset of variables includedinX that are assumedto be important) besides the propensity score may decreasethe selection bias and may be an additional protection against any impact due to inconsistent estimation of the propensity score. The propensity score in combination with the additional variables i s called the balancing score b(Z) (see Rosenbaum and Rubin 1983). The matching on the balancing score i s performed by the Mahalanobis distance (see Rubin 1980): where bi and bj are the balancing scores andZ i s the covariance matrix formed from the C = 1sample. 3.4. Kernel Matching Recently developed nonparametric matching estimators construct a match for each program participant using a kernel weighted average over multiple observations in the control group. For example, consider the kernel matching estimator described in Heckmanet al. (1997, 1998) andHeckman, et al. (1998), given by 132  where G(e)i s a kernel function and a, i s a bandwidth parameter. Interms of equation A7, the weighting function, w(i, j), equals . The neighborhood H(ej) depends on the specific kernel function chosen for the analysis. Under standard C YOjC[") jeIo conditions on the bandwidth and kernel, an i s a consistent estimator of E(Y0 ID =1, eJ. 133  ANNEX3 SECTIONc:MANUFACTURINGINCENTIVESFORTHE CONTENT OFMALAYSIA'S SECTOR 1. Explanation of Incentivesfor the Manufacturing Sector from the Malaysian Industrial DevelopmentAuthority (MIDA) In Malaysia, tax incentives, both direct and indirect, are provided for in the Promotionof Investments Act 1986, IncomeTax Act 1967, Customs Act 1967, Sales Tax Act 1972, Excise Act 1976 and Free Zones Act 1990. These Acts cover investments in the manufacturing, agriculture, tourism (including hotel) and approved services sectors as well as R&D, training and environmental protection activities. The direct tax incentives grant partial or total relief from income tax payment for a specified period, while indirect tax incentives come in the form of exemptions from import duty, sales tax andexcise duty. INCENTIVES FOR THE MANUFACTURING SECTOR 1.1MainIncentives for Manufacturing Companies The major tax incentives for companies investing in the manufacturing sector are the Pioneer Status or the Investment Tax Allowance. Eligibility for Pioneer Status or Investment Tax Allowance is based on certain priorities, including the levels of value-added, technology used and industrial linkages. Such eligible projects are termed as "promoted activities" or "promoted products" (please refer to the List of Promoted Activities andProducts - General) (i) Status Pioneer A company granted Pioneer Status enjoys a 5-year partial exemption from the payment of income tax. It pays tax on 30 percent of its statutory incomeg0, with the exemption periodcommencing from its Production Day (defined as the day its production level reaches30 percent of its capacity). To encourage investment in the promoted areas, i.e. the States of Sabah and Sarawak and the designated "Eastern Corridor"'' of Peninsular Malaysia, applications received from 13 September 2003 from companies located inthese areas will enjoy a 100 percent tax exemption on their statutory income during their 5-year exemption period, Companies which have been granted approval for this incentive but have not commenced commercial production, or applications under consideration, are also eligible. All project applications received by 31December 2005 will be eligible for this enhancedincentive. 9oStatutory Income is derived after deducting revenueexpenditure and capital allowances from the gross income. The "EasternCorridor" of PeninsularMalaysia covers the States of Kelantan, Terengganu andPahang, and the district of Mersing in the State of Johore. 134  Applications for Pioneer Status should be submitted to the Malaysian Industrial Development Authority (MIDA). (ii) Investment Tax Allowance As an alternative to Pioneer Status, a company may apply for an Investment Allowance (IA). A company grantedan IA gets an allowance of 60 percent on its qualifying capital expenditure (such as factory, plant, machinery or other equipment used for the approved project) incurred within five years from the date on which the first qualifying capital expenditure is incurred. The company can offset this allowance against 70 percent of its statutory income for each year of assessment. Any unutilized allowance can be carried forward to subsequent years until fully utilized. The remaining 30 percent of its statutory income will be taxed at the prevailingcompany tax rate. To encourage investment in the promoted areas i.e. the States of Sabah and Sarawak and the designated "Eastern Corridor" of Peninsular Malaysia, applications received from 13 September 2003 from companies located in these areas will enjoy an allowance of 100 percent on the qualifying capital expenditure incurred within a period of five years. The allowance can be utilized to offset against 100 percent of the statutory income for each year of assessment. Companies which have been granted approval for this incentive but have not commenced commercial production, or applications under consideration, are also eligible. All project applications received by 31December 2005 will be eligible for this enhancedincentive. Applications should be submitted to MIDA. 1.2 Incentives for HighTechnology Companies A high technology company is a company engaged in promoted activities or in the production of promoted products in areas of new and emerging technologies (please refer to the List of Promoted Activities and Products - High Technology Companies). A hightechnology company qualifies for: (i) Pioneer Status with a tax exemption of 100percent of the statutory income for a periodof five years; or (ii)Investment Tax Allowance of 60 percent on the qualifying capital expenditure incurred within five years from the date the first qualifying capital expenditure i s incurred. Any unutilized allowance can be carried forward to subsequent years until the whole amount has been fully utilized. The allowance can be utilized to offset against 100percent of the statutory income for each year of assessment. 135  Applications should be submittedto MIDA. The hightechnology company must fulfill the following criteria: (i) percentageoflocalR&Dexpendituretogrosssalesshouldbeatleastone The percent on an annual basis. The company has three years from its date of operation or commencement of business to comply with this requirement. (ii)Scientific and technical staff having degrees or diplomas with a minimum of 5 years experience in related fields should comprise at least 7 percent of the company's total workforce. 1.3 Incentives For Strategic Projects Strategic projects involve products or activities of national importance. They generally involve heavy capital investments with long gestation periods, have high levels of technology, and are integrated, generate extensive linkages, and have significant impact on the economy. Such projects qualify for: (i) Statuswithataxexemptionof100percentofthestatutoryincomefor Pioneer a period of 10 years; or (ii)Investment Tax Allowance of 100 percent on the qualifying capital expenditure incurred within five years from the date the first qualifying capital expenditure i s incurred. This allowance can be offset against 100 percent of the statutory income for each year of assessment. Applications should be submitted to MIDA. 1.4 Incentives for Small- andMedium- Scale Companies Effective from the year of assessment 2003, small- and medium-scale companies with a paid-up capital of RM2.5 million and below are eligible for a reduced corporate tax of 20 percent on the chargeable income of up to RM100,OOO. The tax rate on the remaining chargeable income i s maintained at 28 percent. Dividends distributed will be given a tax credit of 20 percent inthe hands of the shareholders. The threshold for chargeable income eligible for the reduced corporate tax rate of 20 percent i s increased from RM100,OOO to RM500,OOO effective from the year of assessment 2004. Small-scale manufacturing companies incorporated in Malaysia with shareholders' funds not exceeding RM500,OOO and having at least 60 percent Malaysian equity are eligible for the following incentives: 136  (i)Pioneer Status with an income tax exemption of 100 percent of the statutory income for a periodof five years; or (ii)Investment Tax Allowance of 60 percent on the qualifying capital expenditure incurred within five years. This allowance can be offset against 100 percent of the statutory income for each year of assessment. A sole proprietorship or partnership is eligible to apply for this incentive provided a new private limitedlimited company i s formed to take over the existing productiodactivities. To qualify for the incentive, a small-scale company has to comply with any one of the following criteria: (i) valueaddedmustbeatleast15percent;or The (ii) projectcontributestowardsthesocio-economicdevelopmentoftherural The population. The company shall carry out the manufacturing of products or participate in activities listed as promoted products and activities for small-scale companies (please refer to the List of Promoted Activities andProducts - Small Scale Companies). Applications should be submitted to MIDA. 1.5 Incentives to Strengthen Industrial Linkages To encourage large companies to participate in an Industrial Linkage Program (ILP), expenditure incurred in the training of employees, product development and testing, and factory auditing to ensure the quality of vendors' products, will be allowedas a deduction in the computation of income tax. Vendors, including small and medium-scale companies that propose to manufacture promoted products or participate in promoted activities in an ILP (please refer to the List of Promoted Activities and Products - Industrial Linkage Program (ILP), are eligible for the following incentives: (i)Pioneer Status with a tax exemption of 100percent of the statutory income for a periodof five years; or (ii) Investment Tax Allowance of 100 percent on the qualifying capital expenditure incurred within five years from the date the first qualifying capital expenditure is incurred. This allowance can be offset against 100percent of the statutory income for each year of assessment. 137  To encourage vendors to manufacture promoted products or participate in activities for the international market, vendors in an approved ILP who are capable of achieving world-class standards in terms of price, quality and capacity, will be eligible for the following incentives: (i) Pioneer Status with a tax exemption of 100 percent of the statutory income for a period of 10 years; or (ii)Investment Tax Allowance of 100 percent on the qualifying capital expenditure incurred within a period of five years which the company can offset against 100 percent of the statutory income for each year of assessment. Applications should be submittedto MIDA. 1.6 Incentives for the Machinery andEquipment 1.6.1 Incentives for the Production of Specialized Machinery andEquipment Companies undertaking activities in the production of specialized machinery and equipment, namely, machine tools, plastic injection machines, plastic extrusion machinery, material handling equipment, packaging machinery, robotics and factory automation equipment, specialized /process machinery or equipment for specific industries, and parts and components of the mentioned machinery and equipment, are eligible for: (i) Statuswithataxexemptionof100percentofthestatutoryincomefor Pioneer a period of 10 years; or (ii)Investment Tax Allowance of 100 percent on the qualifying capital expenditure incurred within five years from the date on which the first qualifying capital expenditure i s incurred. This allowance can be offset against 100percent of the statutory income for each year of assessment. Applications should be submittedto MIDA. 1.6.2 Additional Incentives for the Production of Heavy Machinery Applications received from 13 September 2003 from existing locally-owned companies that reinvest in the production of heavy machinery such as cranes, quarry machinery, batching plant and port material handling equipment, are eligible for the following incentives: (i) Pioneer Status with a tax exemption of 70 percent (100 percent for promoted areas) on the increasedstatutory income arising from the reinvestment for a period of five years; or 138  (ii)Investment Tax Allowance of 60 percent (100 percent for promoted areas) on the additional qualifying capital expenditure incurred within a period of five years. The allowance can be offset against 70 percent (100 percent for promoted areas) of the statutory income for each year of assessment. Applications should be submittedto MIDA. 1.6.3 Additional Incentives for the Production of Machinery andEquipment Applications received from 13 September 2003 from existing locally-owned companies that reinvest in the production of machinery and equipment, including specialized machinery and equipment and machine tools, are eligible for the following incentives: (i) Statuswithataxexemptionof70percent(100percentforpromoted Pioneer areas) on the increasedstatutory income arising from the reinvestment for a periodof five years; or (ii)Investment Tax Allowance of 60 percent (100percent for promoted areas) on the additional qualifying capital expenditure incurred within a period of five years. The allowance can be offset against 70 percent (100 percent for promoted areas) of the statutory income for each year of assessment. Applications should be submitted to MIDA. 1.7 Incentives for Automotive Component Modules New and existing companies that undertake design, R&D and production of qualifying automotive component modules or systems are eligible for: (i)Pioneer Status with a tax exemption of 100 percent of the statutory income for a periodof five years; or (ii)Investment Tax Allowance of 60 percent on the qualifying capital expenditure incurred within five years from the date the first capital expenditure i s incurred. The allowance can be offset against 100 percent of the statutory income for each year of assessment. The qualifying modules or systems are front corner modules, rear corner modules, instrument panel modules, struts and absorbers and spring assembly modules, bumper modules, front cross member modules, function integrated door modules, fuel tank modules, seat modules, pedal modules, door trim modules, floor console modules, tyre and wheel modules, brake systems, wiper systems, exhaust systems, audio systems, heater ventilation air-conditioning systems, air bag systems, power and signal distribution systems, alarm systems, seat belt systems, exterior lighting systems, body in white 139  modules, engine management systems, safety systems, telematics, navigational systems, engine fuel injection systems, and vehicle intelligence systems. This incentiveis for applications receivedby MIDA from 21 September 2002. 1.8 Enhanced Incentives for the Utilization of Oil PalmBiomass Applications received from 13 September 2003 from companies that utilize oil palm biomass to produce value-added products such as particleboard, medium density fiberboard, plywood, pulp andpaper are eligible for the following incentives: (i) Companies New (a) Pioneer Status with a tax exemption of 100percent of the statutory income for a period of 10 years; or (b) Investment Tax Allowance of 100 percent on the qualifying capital expenditure incurred within a period of five years. The allowance can be used to offset against 100percent of the statutory income for each year of assessment. However, companies which have been granted approval for these incentives but have not commenced commercial production, or applications under consideration, are also eligible for these incentives. (ii)ExistingCompanies that Reinvest (a) Pioneer Status with a tax exemption of 100 percent on the increased statutory income arising from the reinvestment for a period of 10years; or (b) Investment Tax Allowance of 100 percent on the additional qualifying capital expenditure incurred within a period of five years. The allowance can be used to offset against 100percent of the statutory income for each year of assessment. 1.9 Additional Incentives for the ManufacturingSector (i)Reinvestment Allowance A manufacturing company that has been in operation for at least 12 months and incurs qualifying capital expenditure to expand, modernize or automate its existing business or diversify its existing business into any related products within the same industry can apply for Reinvestment Allowance (RA). The RA is given at the rate of 60 percent on the qualifying capital expenditure incurred by the company, and can be offset against 70 percent of its statutory income for the year of assessment. Any unutilized allowance can be carried forward to subsequent years untilfully utilized. 140  A company can offset the RA against 100 percent of its statutory income for the year of assessmenti f - The company undertakes reinvestment projects in the promoted areas i.e. the States of Sabah, Sarawak and the designated "Eastern Corridor" of Peninsular Malaysia; or - The company attains a productivity level exceeding the level determined by the Ministry of Finance. For further details on the prescribed productivity level for each subsector, please contact the Inland Revenue Board. (see Useful Addresses - Relevant Organizations) The RA will be given for a period of 15 consecutive years beginning from the year the first reinvestment i s made. Companies can only claim the RA upon the completion of the qualifying project, i.e. after the building i s completed or when the plandmachinery is put to operational use. Assets acquired for the reinvestment cannot be disposed of within a periodof two years from the time of the reinvestment. Effective from 21 September 2002, a company that intends to reinvest before the expiry of its Pioneer Status can surrender its Pioneer Status for cancellation and be eligible for RA. Applications for RA should be submitted to the Inland Revenue Board (IRB), while applications for the surrender of Pioneer Status for RA should be submitted to MIDA. (ii)Accelerated Capital Allowance After the 15-year period of eligibility for RA, companies that reinvest in the manufacture of promoted products are eligible to apply for an Accelerated Capital Allowance (ACA). The ACA on capital expenditure i s to be utilized within three years, i.e. an initial allowance of 40 percent and an annual allowance of 20 percent. Applications should be submitted to the IRB accompanied by a letter from MIDA certifyingthat the companies are manufacturing promotedproducts. (iii) ExemptionontheValueofIncreasedExports Tax To promote exports, manufacturing companies inMalaysiaqualify for: - A tax exemption on the statutory income equivalent to 10 percent of the value of increased exports, provided that the goods exported attain at least 30 percent value- added; or 141  - A tax exemption on the statutory income equivalent to 15 percent of the value of increased exports, provided that the goods exported attain at least 50 percent value- added. Claims should be submittedto the IRB. To further encourage the export of Malaysian goods, a locally-owned manufacturing company with Malaysian equity of at least 60 percent i s eligible for: - A tax exemption on the statutory income equivalent to 30 percent of the value of increasedexports, providedthe company achieves a significant increase in exports; - A tax exemption on the statutory income equivalent to 50 percent of the value of increasedexports, providedthe company succeeds inpenetrating new markets; - A full tax exemption on the value of increased exports, provided the company achieves the highestincreaseinexport inits category. These incentives are effective from the year of assessment 2003. Note: Pleaserefer to Section 17 for other incentives related to the manufacturing sector 2. List of Promoted Manufacturing Activities and Products that are Eligible for Consideration of Pioneer Status and Investment Tax Allowance under The Promotion of Investments Act 1986 I.Processing;of agricultural produce 1. Chocolate andchocolate confectionery 2. Cocoa and cocoa products 3. Fruits 4. Vegetables, tubers or roots 5. Essential oils 6. Livestock products 7. Agricultural waste or agricultural by-products 8. Aquatic products 9. Aquaculture feed 10. Plant extracts for pharmaceutical, perfumery, cosmetic or food industries 11.Highfructose syrup 12. Illipe products* 13. Coconut products except copra or crude coconut oil* 14. Starch products* 15. Cocoapaste, butter andpowder* 11.Manufacture of nibberproducts 1. Earthmover tyres, agricultural tyres, industrial tyres, commercial vehicle tyres, motorcycle tyres, aircraft tyres or solid tyres 142  2. Precuredtread liners 3. Retreading of aircraft tyres 4. Latex products: a. Surgical gloves b. Safetyhpecial function gloves c. Condoms d. Catheters e. Rubber (elastomeric) specialty coatings f. Rubberisedfabrics 5. Dry rubber products: a. Beltings b. Hoses, pipes andtubings c. Rubber profiles d. Inflatablerubberproducts e. Industrialand office equipment rollers f. Seals, gaskets, washers, packings andrings g. Anti-vibration, damping and soundinsulation products h.Rubber linings i.Rubberfloorings j.Rubbermoulds k.Modifiednatural rubber 6. Reclaimed rubber 7. Latex products:* a. Carpet underlay b. Swimmingcaps c. Balloons d.Finger cots e. Toys f.Latex thread 111.Manufacture of palmandpalmkerneloilproducts andtheir derivatives 1. Oleochemicals or oleochemical derivatives or preparations 2. Margarine, vanaspati, shortening or other manufactured fat products 3. Cocoa butter replacers, cocoa butter substitutes, cocoa butter equivalent, palm midfraction or special olein* 4. Crude palm kernel oil or palm kernel meal* 5. Refiningof palm oil or palmkernel oil* IV.Manufacture of chemicals andpetrochemicals 1. Chemical derivatives from organic or inorganic sources 2. Finechemicals 3. Basic manufacture of pesticides 4. Petrochemical products 5. Epoxy encapsulation mouldingcompounds 6. Cable compounds (excluding PVC cable compound) 7. Titanium dioxide pigment 143  8. Barium sulphate pigment 9. Irondioxide pigment 10.Metallic pigment 11.Recyclingof chemicals 12. Anti-tack solutions 13. Inkjet inks 14. Cleaning preparations, cosmeticsor toilet preparations" 15. Wax products* 16. Specialized paints or coatings* V. Manufacture of pharmaceutical andrelatedproducts 1. Pharmaceuticals goods 2. Clinical diagnostic reagents 3. Gelatine or gelatine products 4. Intravenous, dialysis or irrigating solutions 5. Vaccines 6. Medicaments VI. Manufacture of wood andwood products 1. Reconstituted wood-basedpanel boardsor products 2. Wooden solid or other specialized function doors or wooden solid windows 3. Multi-ply parquet 4. Wooden furniture or parts 5. Insulationfor cryogenic vessels 6. All wooden products except sawn timber, veneer andplain plywood* VII. Manufacture of pulp, paper andpaperboard 1. Pulp 2. Newsprint 3. Security paper 4. Resin impregnatedpaper andproducts thereof 5. Printingand writing paper 6. Corrugated mediumpaper, testliner or kraftliner 7. Kraft paper 8. Paperboard 9. Mouldedpaper 10.Specialty paper 11.All types of paperproducts from pulp* VIII. Manufacture of textiles andtextile products 1, Naturalor man-madefibres 2. Yarn of natural or man-made fibres 3. Woven fabrics 4. Finished knitted fabrics 5. Finishingof fabrics such as bleaching, dyeing andprinting 6. Knitwear 144  7. Skiwear or winter outerwear 8. Non-woven products 9. Elastic webbings 10.Textile hose piping IX.Manufacture of clav-based, sand-basedandother non-metallic mineralproducts 1. Highaluminaor basic refractories 2. Kilnfurniture 3. Laboratory, chemical or industrial wares 4. Artware, ornaments or articles for adornment of ceramic or glass 5. Glassware 6. Hightension electrical glass insulators 7. Glass components or parts for electrical, electronic or industrialuse 8. Glass fibre in all forms producedfrom basic raw materials 9. Finished woven fabrics of glass fibre 10.Optical glass blanks 11.Alumino-silicate ceramics fibres 12. Ceramic components or partsfor electrical, electronic or industrialuses 13. Fritz, zirconium silicate powder, glaze or glaze stains 14. Silicon dioxide fillers 15. Rockwool 16. Synthetic industrialdiamonds 17. Processedball clay 18. Articles of pressedor moulded glass such as bricks, tiles, slabs, pellets, paving blocks and squares 19. Tableware 20. Coated glass 21. Integrated cement projects 22. Absorbent mineralclay 23. Marble and granite products 24. Gypsumplaster board 25. Quicklime and hydratedlime 26. Processedkaolin* 27. Ceramic wall or floor tiles* 28. Vitrified clay pipes* 29. Calciumcarbonatepowder* 30. Coatedor uncoated talc or barium sulphatepowders (average particle size less than 5 microns)* 31. Highgrade silica sand or powder* 32. Panels, boards, tiles, blocks or similar articles of vegetable fibre, wood fibre, straw, wood shavings or wood wastes, agglomerated with cement, plaster or other mineral bindingsubstance* 33. Clay roofing tiles* X. Manufacture of iron andsteel 1. Blooms or slabs of steel 145  2. Shapes or sections of steel of height more than 200 mm 3. Plates, sheets, coils, hoops or strips of steel: a. Hot rolled plates, sheets, coils, hoops or strips b. Coldrolledcoldreducedplates, sheets, coils, hoops or strips 4. Seamless steel pipes 5. Seamless highpressure gas cylinders 6. Steel tyre cord and highpressurereinforced hose wire 7. Ferromanganese,silicon manganeseor ferrosilicon 8. Electrolytic galvanized steel sheet incoil 9. Welded steel pipes or pipe fittings* 10. Bars or wire rods (except those of mildsteel), angles, shapes or sections of all grades of steel either hot-rolled, cold-rolled or cold-finished* 11.Wires or wire products of ironor steel* 12. Steel fabricated products* XI. Manufacture of non-ferrous metalsandtheir products 1. Dressing andor smeltingof non-ferrousmetals other than tinmetals 2. Primary ingots, billets or slabs of non-ferrous metals 3. Bars, rods, shapes or sections of non-ferrous metals except EC copper rods 4. Plates, sheets, coils, hoopsor strips of non-ferrous metals 5. Pipes or tubes of non-ferrous metals 6. Copper clad laminates andproducts from in-house copper clad laminates 7. Powder, cream or paste of non-ferrous metals 8. Wire or wire products of non-ferrous metals 9. Fabricated products of non-ferrous metals * * XI1. Manufacture of machinery andmachinery components 1,Specializedlprocess machineryor equipment associatedwith specific industry including: a. Agricultural machinery or equipment b.Miningor mineralextraction/processing machinery or equipment c. Construction machinery or equipment d. Waste waterhewage treatment equipment e. Industrialsewing machines 2. Supporting services machinery or equipment including power generating machinery or equipment 3. Materialhandling machinery or equipment includingelevators or escalators 4. Handtools or power tools 5. Machinery andindustrialpartshomponents including: a. Printingrolls or embossingrolls b.Dicingblades, accessoriesfor silicon wafers or ceramic substrates c. Offset printingplates d. Industrialseals or sealmaterials 6. Machine tools (metalworking, woodworking and others) including weldingholdering equipment 7. Packagingmachinery 146  8. Machinery or equipment for the services sector including: a. Fire fighting equipment b.Handlabellers 9. Reconditioningof heavy machinery and equipment a. Automobile air conditioning compressors 10. Servicing andupgradingof machinery andequipment XIII.Manufactureof transport eauipment, componentsandaccessories 1. Bicycles 2. Bicycles parts: a. Drive set (chain wheel andcrank) b. Brake set c. Speedchange set d. Hub 3. Specialty cars 4. Engines 5. Engineparts: a. Cylinder block, cylinder head, rocker cover, flywheel or pulley b.Crank shaft, connecting rod, cam shaft, rocker, rocker shaft, engine valve, sprocket, piston pinor piston ring c. Intakemanifoldor exhaustmanifold d. Oil pan, oil pump,oil pumpgear shaft, fuel pump, water pump or oil seal e. Timingbelt, timingchain, carburettor, ignition coil or distributor f.Fuelinjectionmechanism(injector, pump,tubing, valves, regulator, sensors, electronic control modules) g. Hightension cables h.Enginebracket i.Magneto j.Capacitor dischargeunit 6. Transmissions 7. Transmission parts: a. Transmission shift lever andfork b. Transmission control linkages c. Speedometerpinion d. Clutch e. Torque converter f. Drive shaft 8. Axle, wheel, wheel hub or knuckle 9. Disc brake, drumbrake, brake cylinder, brake master cylinder, brake booster, anti-lock brakingmechanism, clutch master cyclinder or clutch operating cyclinder. 10.Steering wheel, steering column, steering gear box, power steering pump, steering linkages, tie rod or constant velocityjoints 11.Stabilizer bar, suspensionarmor suspensionarm shaft andmember 12. Body panels, chassis frame, fuel tank, window regulator, locks and keys or hinges 13. Head lights, indicatindsignallinglights, meters, gauges, switches or horns 14.Weather strips, control cables, speedometer cables, metalic tubings or hoses  15. Catalytic converter 16. Vehicle safety air bag 17. Navigational system 18. Automotive electronic modulekomponent or sensor 19. Seat mechanism including seat adjuster or lockingmechanism or seat recliner 20. System integrator: a. Front corner module b. Rear corner module c. Instrumentpanel module d. Strut and absorbers andspringassemblymodule e. Bumper assembly f.Frontcrossmembermodule g. Functionintegrated door module h.Fueltank module i.Seatassembly j.Pedalassembly k.Door trimassembly 1. Floor console assembly m.Tyre andwheel assembly n.Brakesystem 0. Wiper system p.Exhaust system q. Audio system r.W A C (Heater Ventilation Air-conditioning system) s. Airbag system t. Power and signal distribution system u. Alarm system v. Seat belt system w. Exterior lighting system x. Body inwhite assembly 21. Gear 22. Coolingequipment, air-inlet equipment or exhaust equipment, compressor and expansion valve for automotive air-conditioning 23. Aerospace industry: a. Manufacture and assembly or aircraft b. Manufacture of aircraft equipment, components, accessories or parts thereof c. Ground support equipment for aerospaceindustry 24. Pleasurecrafts, hydrofoils or hovercrafts 25. Maintenance, repair, overhaul or service of aircraft, aircraft components or accessoriesor testing and repairing of avionics 26. Manufacture of train andrelatedequipment: a. Construction of locomotive, railcar b. Coach, wagon, bogie c. Electric multiple unit andpower generating car d. Railway signallingand communication system 27. Electrical or electronics systems instrumentation* 148  28. Shipbuilding* 29. Shiprepair* XIV. Supportingproductdactivities 1. Metal castings 2. Metal forgings 3. Metal surface treatmentlfinishing 4. Machining 5. Moulds, tools and dies 6. Powder metallurgicalparts(sintering of metalparts) 7. Heat treatment 8. Mouldtexturing 9. Metal stamping * 10. Galvanizing, shearing or slittingof metal sheets or other relatedengineering services* 11,Overhaul, repair, reconditioning, modificationor servicing andtesting of turbine engines, components or sub-assemblies 12. Advanced composite materials 13. Irradiationservice 14. Gas sterilization service XV. Manufactureof electrical andelectronic products andcomponents andparts thereof 1.Digitaltelevision receivers 2. Color televisionreceiver parts: a. Cathode ray tubes b.Electronguns c. Polished glass panels or glass funnels for color picturetubes 3. Digital audio video recorders/players and parts: a. Digital audio video recorders/players b.Digital tape mechanisms c. Digital disc mechanisms d. Opticalpick-upunits e. Magnetic heads 4. Computers, parts and peripherals: a. Computers (excluding detachedperipherals not manufactured in-house) b. Monitors c. Computer printers (including printer mechanism) d. Printerheads e. Computer scanners f.Drive units g. Headgimbal assemblieslheadcarriage assemblies h.Headstack assemblies i.Computermagneticheads j.Datastoragemedia k.Voice coilmotors 1. Actuators m.Electronic games equipment includingphoto detectorjoysticks 149  n.Disksubstratesor diskblanks 0. Re-manufacturing of computer drives 5. Electronic components: a. Quartz crystals b.Motors c. Printedcircuit boards (excluding rigidsingle sided circuit boards) d. Cables or wires for electronic devicesincludingflat cables e. Hermetic seals f. Electrical/electronic componentsmoulded with magnets g. Thermistors h.Connectors with or without wires or cables i.Bondingwires j. Lead-frames k.Magnets or ferrite cores 1. Displays-electroluminescent, plasma or liquidcrystal m. Membrane switches n.Surfacemountcomponents 0. Optical fibres or optical fibre products p. SMTchipholders on lead-frames q.Magnetron r.Fabricationof lightemittingdiodes(LED) 6. Recorded and unrecorded media: a. Compact discs b. Magnetic webs or pancakes 7.Electronic machines andequipmentldevices: a. Teller machine b. Office equipment c. Alarmequipmenthystems or devices d. Ultrasonic cleaners e. Computing scales f.Cashregisters g. Demagnetizers h.Industrialcontrollers i.ComputerAidedDesign(CAD),ComputerAidedManufacturing(CAM) or Computer Aided Engineering (CAE) equipment j.Robotsor robotics k.Multimediaintegratedcontroller 8. Wafer fabrication: a. Semiconductor wafer fabrication b.Reclaimed silicon wafers 9. Electricalproducts: a. Unintemptible power supplies b. Batteries excluding manganesedioxide, dry cells andlead acidbatteries c. Solar cells and solar panels d.Discharge tubes e. Heat shrinkable cablejoints andterminations 150  10. Telecommunication: a. Telecommunication equipment includingmultifeature mobile phones but excluding fixed line telephone sets b.Antennae for communication equipment c. Voice/pattern/vision recognition or synthesis equipment d.Dataterminal displays e. Global positioning system f. Electronicnavigational aid g. Electronic tracking aid 11.Software development andproduction 12. Discharge tubes andproducts thereof 13. Transformers or coils* 14. Automatic gate mechanisms* 15. Consumer electronic products; parts, sub-assembliesor accessoriesthereof* 16. Industrial electronic products; parts, sub-assembliesand accessoriesthereof* 17. Electricalhousehold appliances and partsthereof* 18. Electrical industrial equipment or parts thereof* XVI. Manufacture of professional, medical, scientific andmeasuring devicedparts 1. Medical, surgical, dental or veterinary devices/equipment 2. Gauges or measuring apparatus 3. Surveying, hydrographic, navigational, meteorological, hydrological or geophysical instruments 4. Testingequipment 5. Clocks or watches 6. Stainless steel cannulae or tubes for needles XVII. Manufacture of photographic, cinematographic, video and optical goods 1. Cameras 2. Lenses 3. Binoculars, telescopes, magnifyingglasses or microscopes 4. Cinematographic or video equipment XVIII. Manufacture of plastic products 1. Inflatable plastic products 2. Specialized plastic filmdsheets 3. Geosystemsproducts [Cellular Confinement System (CCS) and Porous Pavement System (PPS)] 4. Plastic products for engineering use 5. Precision engineering plastic products 6. Multiwall pipes 7. Expandedpolystyrene foam* XIX. Miscellaneous 1. Musical instruments 2. Furniturehardware 151  3. Souvenirs, handicrafts or giftware 4. ElectronicToys 5. Sports goods or equipment 6. Spectacles or spectacle frames 7. Accessories for the textile industry 8. Cutlery 9. Lock sets or lock cylinder mechanisms 10. Jewellery of precious metal 11.Costumejewellery 12. Designing and printingof decorative surfaces for commercial applications 13. Integrated exhibits 14. Microbials and probiotics 15. Bank notes 16. Thermi containers andparts thereof 17.Biodegradable disposablepackaging andhousehold wares 18. Toys (excluding electronic toys)* 19. Art and design apparatus - alltypes* 20. Enamelledhouseholdware* 21. Cooker or barbequesets* 22. Fall protectionequipment 23. Ball pentips 24. Bio-ceramics embeddedtextile products * Additional promoted activities and products for promoted areas (other than the Federal Territory of Labuan) i.e. Sabah, Sarawak, Kelantan, Terengganu, Pahang and the district of Mersing inJohor. 3. List of Promoted Activities and Products for High Technology Companies under the Promotionof Investments Act 1986 I. electronics Advanced 1. Design, development andmanufacture o f a. computer or peripherals b.microprocessor application 2. Development andproduction of communication equipment 3. Designand production of integratedcircuits (ICs) 4. Design, development and production of cameras including,lens units, lens barrel units andview finder units 5. Development and production of cathode ray tubes and advance displays 6. Design, development and manufacturer of printer heads, head gimbaldhead carriages, headstacks,magnetic heads, voice coil motors and actuators 7. Development and production of advancedconnectors 11.Equipment/ Instrumentation 1. Design,development andmanufacture o f a. medical equipment 152  b. medicalimplant or devices c. scientific equipment d. cyclonic separationequipment 2. Development andproduction of highpressurewater cuttingequipment 111.Biotechnology 1. Development, testingandproduction o f a. pharmaceuticals b. fine chemicals c. food or food ingredients d. feed or feed supplements e. biodiagnostics 2. Development and production o f a. cell cultures b. biopolymers c. biomaterials 3. Development andproduction of biotechnology processesfor waste treatment IV. Automation andflexible manufacturing:systems 1. Development andproduction of a. computer processcontrol systems/equipment b. processinstrumentation c. robotic equipment d. computer numerical control (CNC) machine tools V. Electro-optics andnon-linear optics 1. Development andproduction o f a. optical lenses b. laser application equipment c. fibre-optic communication equipment VI. Advanced materials 1, Application or production o f a. polymers or biopolymers b. superconductors c. fine ceramics or advancedceramics d. highstrength composites VII. Optoelectronics 1. Development andproduction o f a. optoelectronics systems components b. optical systems components c. photo-couplers d. semiconductorslasers 153  VIII. Software engineering 1. Development andproduction o f a. neuralnetworks b.patternrecognition systems c. machine vision d. fuzzy logic systems IX.Alternative energy sources 1. Development andproduction o f a. fuel cells b. polymer batteries c. solar cells d.renewableenergy X. Aerospace 1. Designor development andproduction or assembly o f a. aircraft b.aircraft equipment, components or accessoriesor partsof aircraft 2. Modificationand/or conversion of aircraft 3. Refurbishment or re-manufacture of aircraft equipment, components or accessories or partsof aircraft XI. Foodproduction and foodprocessing 1, Foodproduction usingemerging technologies andadvancedfarming systems. 2. Development, testing and manufacturing of food products usingemerging technologies and advancedmanufacturing systems XII. Engineering support industrieshervices 1. Designor development andmanufacture o f a. trim and form dies b. semiconductor cavity/encapsulation moulds c. suspension tooling for harddisk drive parts d. progressive tooling for leadframes e. fibre optic connection tooling 2. Design, development and manufacture of advancedtoolings and equipment for the production of precision components/parts for industrialapplications 3. Development and production of precision machined and die cast parts using advanced manufacturing systems 154  4. List of PromotedActivities andProducts for small scale companiesfor consideration of Pioneer Status under The Promotion of Investments Act 1986 I.Processing of Agricultural Produce 1. Cocoa andcocoaproducts 2. Coffee 3. Tea 4. Coconut products except copra and crude coconut oil 5. Fruits 6. Vegetables 7. Cereal Products 8. Starch andproteins 9. Herbs or spices 10.Essential oils 11.Fodder or other animal feedingredients 12. Tobacco 13. Flowers or ornamental foliages 14. Honey 15. Meat 16. Livestock products 17. Aquatic products, including seaweed 18. Agricultural waste and by-products 19. Aquaculture feed 20. Sugar and confectionary products 11.Forestry Products 1. Rattanproducts (excluding rattanpoles, peel andsplits) 2. Bambooproducts 3. Other forestry products 111.Manufacture of RubberProducts 1. Mouldedrubberproducts 2. Extrudedrubber products 3. General rubber goods 4. Foamrubberproducts 5. Inflatable rubber products 6. Engineering components of rubber (e.g. buildingmounts and anti-vibration mounts) 7. Rubber adhesive/ rubber sealant IV.Manufacture of PalmOil andPalmKernelOil Products andtheir Derivatives 1. Margarine,vanaspati, shortening andother manufactured fat products V. Manufacture of Chemicals andPharmaceuticals 1. Pigmentpreparation anddispersion 2. Desiccant 155  VI. Manufacture of Wood andWoodProducts 1. Fancy plywood, prefinishedandprinted plywood 2. Timber mouldings 3. Builders carpentry andjoinery 4. Products derived from utilization of wood waste (e.g. activated charcoal, wooden briquettes, wood wool) 5. Wooden household and office articles 6. Wooden furniture parts and components VII. Manufacture of Textiles andTextileProducts 1. Batik 2. Accessories for the textile industry VIII. Manufacture of Clay-Based and Sand-Based Products and other Non-Metallic MineralProducts 1. Artware, ornaments andarticles of ceramic and glass 2. Glass fittings for lightingpurposes 3. Panels, boards, tile blocks and similar articles of vegetable fibre, straw, wood shavings or wood wastes, agglomerated with cement plaster or with other mineral bindingsubstances 4. Abrasive products for grinding, polishing and sharpening IX.Manufacture of IronandSteel Products 1. Wire andwireproducts of iron and steel 2. Steel fabricated products X. Manufacture of Non-Ferrous Metals andtheir Products 1. Wire andwire products of non-ferrous metals XI. SupportingProducts andServices 1. Metal forgings 2. Machining 3. Stamping 4. Plating 5. Moulds, tools and dies 6. Industrialseals and seal materials XII. Manufacture of Handtools 1. Handtools, all types (such as axes, pliers, spanners, screwdrivers, wrenches, hammers, pincers, riveting tools and other handtools) XIII. Manufacture of Motor Vehicles, Components and Accessories 1. All automotive components and accessories 156  XIV. Assembly and Manufacture of Electricaland Electronic Products and Components andPartsThereof 1. Decorativelights 2. Antennae 3. Capacitors 4. Disc card players 5. Energy-saving lamps 6. Resistors 7. Power supplies 8. Invertors 9. Key pads andkey switches 10. Printed Circuit BoardAssemblies using surface mount technology 11.Electronic ballast XV. Manufacture of Kitchenware 1. Kitchenware, all types 2. Tableware XVI. Manufacture of Furniture 1. Furniture XVII. Manufacture of Toys 1. Toys XVIII. Manufacture of Souvenirs 1. Souvenirs, handicrafts, giftwares anddecorative wares XIX. Manufacture of Sports Goods andEquipment 1. Sports goods andequipment, all types XX. Manufacture of Jewellery andRelatedProducts 1. Jewellery 2. Processedgems XXI. Miscellaneous 1. Wax products XXII. Manufacture of Plastic Products 1. Decorative panels andornamentsof plastic 157  5. List of promoted products and activities in an Industrial Linkage Programme (ILP) for consideration of Pioneer Status or Investment Tax Allowance under the Promotion of Investments Act 1986 I.Manufacture of rubber products 1. Mouldedrubber products 2. Conveyor belts, transmission belts, V-type belts or rubber beltings 11.Manufactureof plastic products 1. Plastic products for engineering use 111.Manufactureof clav-based, sand-basedandother non-metallic mineralproducts 1. Ceramic components or partsfor electrical, electronic or industrialuses 2. Glass envelopes 3. Glass fittings 4. Advanced composite materials or products IV. Manufacture of textiles andtextile products 1. Elastic webbings V. Manufacture of ironandsteel 1. Wire or wire products of ironandsteel 2. Steel fabricated products VI. Manufacture of non-ferrous metals andtheir products 1. Copper cladlaminates andproducts thereof 2. Wire or wire products of non-ferrous metals 3. Fabricatedproducts of non-ferrous metals VII. Supporting productshervices 1. Metalcastings 2. Metal forgings 3. Plating 4. Machining 5. Moulds, tools or dies 6. Heat treatment 7. Mouldtexturing 8. Metal stamping 9. Industrial seals or seal materials 10.Powder metallurgical parts(sinteringof metalparts) 11.Maintenance, repair, overhaul, modification, servicing or testing of turbine engines, components or sub-assemblies 12. Maintenance, repair, overhaul, modification, servicing or testing of aircraft, aircraft components or accessories 13. Maintenance, repair, overhaul, modification, servicing or testing of ship components or accessories 158  VIII. Manufacture of transport equipment, componentsandaccessories 1. Parts andcomponentsfor bicycles ortricycles 2. Parts andcomponents for pleasurecrafts, hydrofoils or hovercrafts 3. Parts, components or accessoriesfor motor vehicles 4. Aircraft equipment, components, accessoriesor partsthereof IX.Manufacture of machinery andmachinery components 1. Machinery components X. Manufacture of electrical andelectronic products andcomponents andparts thereof 1. Computer peripherals: a. Drive units b. Keyboards 2. Alarm equipmenthystem or devices 3. Parts, sub-assembliesor accessories of consumer or industrial electronic products. 159  ANNEX 4: CONDUCTINGPOLICY CONSULTATIONS When considering an R&D incentive, a full consultation exercise should exhaustively cover the issues raisedinFigureA4.1. Figure A4.1: Key Questions inDesigningan R&DIncentive "Leakages" o f the subsidy to foreigners A H o w much Firm compliance would have costs? been done anyway? A 1 I s there rent seeking b y R&D fiscal , firms? H o w much incentives Derive bang for R&D i s induced? H o w much H o w much does did it cost inducement - change with to induce? What is the subsidies? subsidy rate? The role o f adjustment costs? costs overall What are subsidy the spillovers? What i s the Are private cost o f raising returns 4 finance? reduced? D o some firms respond more? The role o f the investment + horizon Source: Lattimore (1997). 160  6.24 Design principles that should be borne in mind when an incentive i s structured. are summarized inFigure A4.2. Figure A4.2: Design Principles for an R&Dincentive 1 Target the source of the problem 2 Selectionissues(adverseselectionand uptake) 3 Inducement 4 Right scaleof resources 5 Timeliness 6 The rightduration 7 Avoiding strategicbehavior/distortions 8 Avoidingunforeseengovernmentliabilities 9 Encouragingadministrativeefficiencyand ease 10 Accountabilityandtransparency 11 Policylearning 12 Robustness 13 Avoiding badinteractionstencouraginggoodones 14 Cost effectivenessand net benefit 15 Superiority(greatestnet benefits) 16 Evaluation,monitoringand reporting Source: Lattimore (1997). 6.25 Our analysis underlying this report has raised the following questions, which ought to be covered ina consultation exercise: 0 Providingincentives for all Malaysian R&D o How much R&D undertaken in Malaysia is for foreign firms and is therefore not eligible for an R&D double-deduction? o Could the current system be changed to ensure that the Government provides an R&D tax credit for this R&D, Le. an incentive for R&D financed from abroad? o Do you think a scheme in which Malaysian firms could claim the R&D credit themselves if they were doing contract R&D for foreign firms would be operable? Do you see any potential for tax abuse under this scheme? 0 Differential rates of provisionon similar R&D o How could the government harmonize these three systems into one without raising the potential for tax abuse?For example, the Government i s considering moving all systems to a system in which firms can obtain 161  either a 100percent IA or Pioneer Status (as in a contract R&D Company) or a Double Deduction. This i s more generous than the current "R&D Company" provision, less generous than the "Contract R&D Company", and different from the "In house research" provisions. What are your views on this system? o What is the relative generosity of the MlDA schemes (IA and Pioneer) versus the IRB schemes (double deduction)? What i s your experience of them? 0 Operation o What is your experience with applying for and receiving double- deduction, IA and pioneer tax incentives? Was this transparent, quick and easy to do? o Ifnot, what problems didyou encounter?Do you have any suggestionsfor how the Government could address these? 0 0 Checks on R&D activities o I s there potential underthe current systemfor afirm to become certified as R&D doing but then changeits activities and still claim the credit?Do you know of any instanceswhere this has happened? o What should the Government do to address this? 0 Transfer Pricing o Do you think there is much scope for abuse of the R&D tax incentives through transfer pricing?If so have you seen any evidence of this? o I s there any way in which the Government could improve its current regime on transfer pricing?For example would: (a) more guidance; andor (b) transfer pricing legislation help? 0 Checking of Overseas R&D o How rigorously does the IRB check R&D expenditure claimed against double-deduction which i s carried out overseas? o Do you think it would be possible for firms to claim R&D which was not being carried out or was not really R&D? o Why should the Government subsidize R&D undertaken overseas given it's explicitly objectives to promote R&D within Malaysia? 0 R&D Capital Resale o Would it be possible for a firm to buy a capital asset, claim the IA allowance on this, and then resell this asset without returning the IA tax saving? Inyour experience has this ever happened? o How could the Government counter this threat without removing the tax incentives on capital assets? 162  0 False Company "rebirth" o I s it possible for companies to keep claiming the IA or Pioneer status by transferring all their activities to a newly incorporated subsidiary in the same group? Are you aware of this happening? o What measurescouldthe Government take to prevent this happening? 0 Firmsobtaining credits on grants o I s it possible to claim aDDon revenues that have come from MOSTI? o Have you ever considered doing this?Have you ever come across anyone that has managed to obtain grants and tax credits on the same R&D project? 0 INPioneerStatus o Have you ever triedto obtain aMIDA IA andor Pioneer scheme?Didyou succeedand what was your experience of this? Consultation Documents Two consultation documents may provide useful comparisons. For R&D tax credits, a document was producedby the UK Treasury as a summary of the consultation exercise for the R&DTax Incentivethat was carried out inthe UK in2001. It can be found at: http://www .hm-treasury.gov.uk/media/F2A9F/ACF525.pdf For Annuities, a consultation document was produced by the UKDepartment of Works andPensions andthe InlandRevenue. Although the topic of the consultation i s different, the approach and the style of the consultation i s similar to what we are advocating here. It can be found at: http://www.inlandrevenue.gov.uk/pdfs/mod annuities.pdf 163