Promoting growth in Sri Lanka : lessons from East Asia

A)P5 IH7v POLICY RESEARCH WORKING PAPER 1478 Promoting Growth Why - despite solid progress in human development and in Sri Lanka in the reduction of consumption poverty - has Lessons from East Asia Sr Lanka's per capita income fallen far behind the dynamic East Asian economies? Sri Sadiq Ahmed Lanka's weaker growth Priya Ran jan performance was the result of several factors, including Sri Lanka's lower investment and (especially) domestic savings rates, its lower average productivity of investment (return on capital), its weak agricultural performance, and its poor export growth. The World Bank South Asia Country Department I Office of the Director June 1995  | PoucY RESEARCH WORKINCG PAPER 1478 Summary findings Sri lIanka's weak economic performance, although tariff structure, and, possibly, reform customs (to reduce compoiunded by the civil war and budgetary imbalance, leakage and abuse). Iargely reflects: * Rationalize employment, exit, and bankruptcy * A stop-and-go pattern of policv reform, because of regulations and procedures. political constraints - even thougIl the results of reform nImprove communications between government and were generally positive. the private sector. * Weak economic management, resulting in high * Make the financial sector more competitive by inflation and a high fiscal and balance of payments legislating banking reform, giving state-owned banks deficit. more autonomy and putting private commercial banks * Poor management of public spending. on an equal footing with the two state banks, with the * Mixed performance in exchange-rate management. ultimate goal of privatizing the state banks. with periods of substantial overvaluation. * Strengthen the supervision of banking. * Financial policies that (despite recenit * Privatize insurance and pension funds to strengthen improvements) hamper efficient financial intermediation. the capital market. * Prolonged trade protection, followed by selective * In the agriculture sector: (1) privatize the estate trade liberalization. plantations, perhaps through long-term management * Continued distortion in agricultural policies. contracts and the gradual sale of shares in assets, (2) * Inflexible labor markets and, despite Sri Lanka's reduce trade protection (especially on rice, wheat, outstandinig track record on humiian development, potatoes, chilies, and onions), (3) implement land problems with the quality of the labor force. reform, (4) strengthen agricultural support (for example, To address a substantially unfinished policy agenda, Sri irrigation, research and extension, and rural Lanka needs to: infrastructure), and (5) possibly support rural financing * Intensify efforts to peacefuilly resolve civil conflict. institutions. * Squarely address its macroeconomic imbalanices: (1) * End government controls on hiring, firing, and wage sharpiv redtice the fiscal deficit; (2) cur back on public setting, and rationalize (depoliticize) civil service spending (including defense spending, with peace) and employment decisions. redefine spending priorities; (3) improve cost recoverv * Make the changes in education needed to improve for public services; and (4) continue improving the the quality of the labor force (for example, improve maniagement of the exchange rate. teacher training, courses in science and English as a * In trade policy, eliminate most quantitative second language courses, and skills training). restrictions, further reduce tariff protection, simplify the This paper - a product of the Office of the Director, South Asia Country Department I - is part of a larger effort in the region to help identify policies for supporting higher growth in Sri Lanka. Copies of the paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Ann Bhalla, room Dl0-071,extension 82168 (34 pages). June 1995. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. 'I'he papers carry the na,pnes of the authors and should be used and cited accordingly. The findings, interpretations, and conclusions are the authors' own and should tot be attributed to the World Bank, its Executive Board of Directors, or any of its nmetnber countries. I'roduceCl by the l'olicy Research Dissemination Center  PROMOTING ECONOMIC GROWTH IN SRI LANKA--LESSONS FROM EAST ASIA BY Sadiq Ahmed and Priya Ranjan Sadiq Ahmed is Lead Economist in the World Bank's South Asia Region. Priya Ranjan is a graduate student at the Columbia University and worked on this paper as a summer intern. We are grateful to Paul Isenman, Mark Baird and Mohan Munasinghe for helpful comments. The paper also benefitted from a seminar at the Sri Lanka Central Bank. We would like to express our gratitude to the participants (too many to mention individually by name).   SUThARY Sri Lanka has made very impressive progress since independence in reducing the incidence of poverty. This is reflected in outstanding human development indicators as well as when poverty is measured in consumption terms. Despite this progress, Sri Lanka remains a poor country with substantial poverty. Although Sri Lanka's per capita growth rate of 2.5% per annum over the past thirty years or so compares favorably with most of the developing world, it falls much short of the growth rate achieved by the high-performing East Asian countries. Thus, in 1960 Sri Lanka's per capita income in nominal dollars was 276% higher than in Indonesia, 46% higher than in Thailand, about the same as in South Korea and about 50% lower than in Malaysia. Some 30 years later, Sri Lanka's per capita income is now less than a 12th of that in Korea, only a fifth of that in Malaysia, less than a third of Thailand and 24% below that in Indonesia. A stronger growth performance along the lines of the high-performing East Asian economies would have allowed much deeper reduction in Sri Lanka's poverty. What factors explain why, despite solid human development, Sri Lanka's per capita income has fallen far behind the dynamic East Asian economies? Using an aggregative time series growth model as well as a more disaggregated and somewhat heuristic approach the paper shows that Sri Lanka' s weaker growth performance can be explained by differences in the underlying policy framework. Like the East Asian economies, Sri Lanka put a lot of emphasis on human development and equitable distribution of income. These policies did pay off handsomely in terms of strong human development and reduction in consumption poverty. But Sri Lanka was unable to fully exploit its good human capital base to increase its income as rapidly as the high-performing East Asian economies for a number of reasons. First, Sri Lanka's investment and saving rates were lower than in the dynamic East Asian economies; in particular the domestic saving rate was very low. Thus, between 1960-92 Sri Lanka invested at an average rate of 20% of GDP while the investment rates were 27% in Korea, 25% in Thailand, 24% in Malaysia and 21% in Indonesia. More strikingly, as compared with domestic saving rates of 30% of GDP in Malaysia, 25% in Korea, 24% in Thailand and 25% in Indonesia, Sri Lanka's average saving rate was only 13%. Second, the average productivity of investment in Sri Lanka has also been much lower; the average return on capital in Sri Lanka was about 21% as compared with 33% in Korea, 31% in Thailand, 27% in Malaysia and 26% in Indonesia. Third, the performance of agriculture, particularly over the past two decades, has been weak. For example, over the 1970-92 period, Sri Lanka's agriculture sector grew at about 2.3% p.a. as compared with 4.2% in Thailand, 4.1% in Malaysia and 3.5% in Indonesia. Finally, Sri Lanka's overall export growth over the past thirty years has been rather poor. Thus, the volume of exports has expanded at an average rate of only 1.6% between 1960 and 1992, as compared with 20% in Korea, 11% in Thailand and 10% in Indonesia. Although manufactured export performance has improved substantially in response to liberalization policies since 1977, the export base remains fragile and narrowly concentrated on a small number of commodities. These weaknesses in economic performance largely reflect the effects of policy problems in a number of critical areas, although the ongoing civil war has further compounded the difficulties and added to the budgetary imbalance. The policy problems include: weak macroeconomic management, reflected in relatively high fiscal and balance of payments deficit and high inflation; weak public expenditure management; mixed performance on exchange rate management, with periods of substantial overvaluation; a long period of substantial trade protection, followed by trade liberalization on a rather selective basis (agriculture incentives remain distorted); financial sector policies that continue to hamper efficient financial intermediation despite some recent improvements; agriculture policies that have been least reformed over the past 30 years and continue to constrain agriculture performance; and policy distortions that limit the flexibility of the labor market. Even in the area of human development, where Sri Lanka has an outstanding track record, there are concerns about the quality of the labor force.  - ii - Reform efforts in the past sought to address many of the above policy problems with varying degree of success. The main difficulty has been the stop- go pattern of reforms due to political constraints, even though the results of the reforms were generally positive. With a new government voted into power in 1994, there is a real opportunity for taking control and ensuring sustained progress with the critical policy reforms. In addition to intensifying efforts to peacefully resolve the civil conflict, Sri Lanka would need to strengthen and sustain efforts to address the substantial unfinished policy agenda. The nature of these challenges and implications for future reforms are discussed below. Macroeconomic Management: The task of securing and maintaining a stable macroeconomic environment has proven to be the Achilles Heel of Sri Lanka's economic management throughout the period. Fiscal deficits have generally been high for the entire period. The balance of payments deficits were relatively low up to the late 1970s, but have expanded since 1979. To a large extent, these deficits in the balance of payments and the fiscal account reflect the surge in concessional foreign aid since the late 1970s. Nevertheless, large macroeconomic imbalances, particularly fiscal deficits, have contributed to a rapid growth in domestic and international indebtedness, fueled a relatively high rate of inflation, and complicated monetary policy management in Sri Lanka's open economy. Although the growth in international debt did not create any significant debt servicing problem partly because of the high concessionality of the debt flows, the growth in domestic debt has generated some elements of dynamic instability. Thus, in 1992, the government's interest payment obligations (mostly on domestic debt) ate up almost 6% of GDP or a third of total tax revenue. Clearly, a major challenge for Sri Lanka in the coming years will be to squarely address its macroeconomic imbalances. Despite good progress in recent years to bring down the fiscal deficit from double to a single digit number, a fiscal deficit of 8-9% of GDP is unsustainably large. Given the already large stock of public debt, Sri Lanka cannot continue to run a deficit in its primary budget balance for any substantial period in the future without running into a financial crisis. So, there is a need to convert the primary deficit into surplus to maintain internal balance. The magnitude of the fiscal effort is in the range of 4-5% of GDP. In view of Sri Lanka's already good tax effort (19% of GDP), much of the fiscal deficit reduction will need to come from further expenditure cutbacks and redefining public expenditure priorities. Rapid progress with the peace initiative will help in this task by reducing defense spending. However, there is some further scope for additional revenue mobilization through better cost recovery policies for public services. Lower fiscal deficit will help raise domestic saving and reduce inflation through reduction in excess demand resulting from excess money supply. It will also help contain the pressure on domestic interest rates. The management of the exchange rate has improved considerably in the 1980s with the unification of the exchange rate and the move to a crawling peg adjustment mechanism. But a combination of large government spending and large capital flows, especially official aid, has generated pressure on the real exchange rate (RER) to appreciate. While exports have continued to do well, following the trade and investment reforms of the 1977-82 and 1990-92 periods, there is a substantial risk that the continued RER appreciation will hurt exports. A reduction in government spending and lower fiscal deficits will again help. Careful management of foreign aid and private capital flows will be necessary to avoid large demand pressure in the non-traded sector. Financial deepening through further financial sector deregulation will also assist in avoiding an appreciation of the RER. Trade and Industrial Policies: The reform of trade and other policies to promote industrial development has been the subject of debate in Sri Lanka and outside. Sri Lanka's own experience with trade protection has shown the pitfalls of relying on government intervention for industrial development. Inward looking  - iii - policies of the 1970-77 period, for example, seriously hurt exports and sharply reduced the growth of the manufacturing sector. On the other hand, the positive response of exports and the manufacturing sector to trade and other liberalization policies since 1977 suggest the potential benefits of further liberalization. So, it would seem that Sri Lanka's development objectives would be better served through reliance on neutral incentives. Notwithstanding good progress in liberalizing trade policies, there is substantial scope for further reducing the bias against exports induced by the trade policy regime. While many of the quantitative restrictions have been eliminated, a number of restrictions affecting particularly agriculture remain in place. Moreover, a combination of specific duties, surcharges, exemptions and waivers has resulted in a fairly complex tariff regime. These have continued to impart a bias against exports and have generally reduced investment efficiency. Future trade policy reforms should concentrate on eliminating most quantitative restrictions, further reducing tariff protection and simplifying the tariff structure. To ensure proper implementation of these reforms, a reform of the customs department may also be needed. Options here include use of external assessment firms (such as the SGS) for preshipment inspection. Other countries, such as Indonesia, have successfully used this option to minimize leakages and abuse. Sri Lanka continues to rely on a number of fiscal incentives (tax holidays and import duty exemptions) to promote private investment, especially in the manufacturing sector. Particularly, since 1990, Sri Lanka has offered generous tax incentives to promote foreign investment and exports. Overall, these incentives distorted resource allocation and contributed to macroeconomic imbalance by reducing budgetary revenues. In light of these problems, a better policy option would be to replace tax holidays with: (i) a low flat corporate rate of taxation consistent with international standards; and (ii) limited use on investment tax credit aimed at new investments which are consistent with Sri Lanka's long-term development needs. Sri Lanka has come a long way in deregulating its investment licensing regulations. However, a number of regulatory restrictions in the area of employment, exit and bankruptcy procedures continue to impede investment in manufacturing and services. The flexibility of labor market should be increased (discussed below) while the Companies Act and the Bankruptcy Law should be simplified and made more flexible to facilitate winding up of business when needed. A notable feature of the East Asia experience is good communication between the government and business which contributed to the development of a business- friendly environment. In addition to a generally favorable business environment, formal mechanisms were established in a number of countries to have regular consultation between government and business on relevant policies. The focus was on improving competition and sharing information, rather than collusive behavior for rent seeking. Similar consultative arrangement between private sector and the Sri Lankan government might be very useful in disseminating information, developing right policies and building consensus, and proper implementation of policies. Financial Sector Policies: Past reforms in Sri Lanka's financial sector have helped to deepen the financial sector, as indicated by the growth in the ratio of bank deposits to GDP. This has contributed positively to increasing private saving and helped in reducing the adverse effects of high government expenditure and large aid inflows on the RER. Yet, a substantial unfinished agenda remains. The financial sector remains much less developed than in the East Asian high-performing economies. For example, even with past progress, the deposit to GDP ratio is only 24% in Sri Lanka as compared with 38% in Korea, 48% in Indonesia, 57% in Malaysia and 78% in Thailand. Sri Lanka has moved well in deregulating interest rates. However, a major constraint that remains is the  - iv - dominance of inefficient, state-owned banks. Despite poor financial performance, these banks have continued to dominate the financial sector because of various government support including privileged access to deposit sources. The resultant lack of competitive pressure has reduced both the level and quality of financial intermediation. Quality has suffered due to two reasons. First, the state- owned banks have been used as instruments of government policy to direct credit to politically determined beneficiaries, often without regard to financial or economic viability. Second, these banks have played the role of market-makers in determining domestic lending rates. Consequently, while inefficiency resulting from over-staffing and bad portfolio has raised the cost of doing business, the effect of these inefficiencies on cost has often been passed on to the borrowers in terms of higher loan rates. As a result, the spread between average deposit and lending rate has been very high. Recent steps to increase competition has included efforts to legislate a banking reform act that would provide greater autonomy to the state-owned banks and place the private commercial banks on the same footing as the two state-owned banks. A speedy implementation of this measure will be an important first step. Nevertheless, this by itself will not be adequate. The main policy change that is needed to enhance the dynamism of the financial sector is to privatize the two state banks. Along with this key reform, steps will also need to be taken to strengthen supervision capacity of the Central Bank to ensure the soundness of the banking sector in a more competitive environment. Sri Lanka's small but growing capital market has shown dynamism in recent years as direct foreign investment has been growing in response to the various deregulation measures. Nevertheless, the state's control over the use of insurance and pensions fund has tended to reduce the effectiveness of the capital market. Not surprisingly, most of resources mobilized in these institutions have been channelled toward the purchase of T-bills to reduce the cost of the budget deficit. The privatization of insurance companies has been under consideration for sometimes as a way of increasing competition and efficiency in insurance. Implementation of this reform will also benefit capital market development by allowing more flexible use of insurance funds. Similarly, an independently managed public pensions fund is also likely to help the capital market as investment decisions are likely to be guided by profitability considerations rather than dictated by the need to finance the budget deficit at lower cost (i.e., purchase of T-bills). Agriculture Policies: Unlike the East Asian economies, Sri Lanka's agriculture sector performance has generally been weak. Value-added in the tree crops sector has virtually stagnated and paddy experienced mixed performance. Other crops (chillies, onions, potatoes) have done relatively better; but there are serious efficiency concerns in many of these crops. Factors contributing to the sector's poor performance include nationalization of estate plantations, trade policy distortions, absence of a well-functioning land market and weak support services. In the tree crops sector, the main challenge is to privatize the estate plantations. Some progress has been made in this regard since 1992. The reform effort, however, has been slow and hesitant. Given political sensitivity, the implementation of privatization could be phased over a longer period through a system of long-term management contracts (50 years or more) and gradual sale of assets in the form of shares. At the same time, the Government should remove all remaining restrictions on marketing, refrain from involvement in wage setting and restructure tree crop research institutes to allow greater private sector involvement. In the non-plantation agriculture sector, reforms are needed in the areas of trade, land markets and support services. The most important trade policy issue affecting agriculture concerns the control of rice trade. There are also restrictions on import of wheat, potatoes, chillies and onions. Except in the case of wheat, the main aim of trade controls has been to protect farmer incentives. As a result of these policies, a fairly distorted production pattern has emerged. As in the case of the manufacturing sector, neutral incentives would probably be better for boosting agriculture performance than a strategy of  guiding specific production activities through trade protection. Even in agriculture, a small economy like Sri Lanka will likely benefit more from opening up to trade. Reconciling objectives of food security and protection of farmer income with production efficiency considerations could involve some trade-offs, but these can be made consistent with outward orientation. Thus, while lower trade protection for rice, onions, potatoes and chillies could hurt producers in the short-term, the medium-term gains from diversification toward other, more efficient crops would likely improve total income from agriculture. In addition, rice and other farmers facing lower protection need not lose over the medium term if better support services are provided to increase productivity. Changes in trade policy alone will not lead to greater efficiency and higher growth in agriculture. A major constraint that needs to be addressed as well is the lack of a well-functioning land market due to the state's role as the dominant landlord, restrictions on land sale and use, and inadequate land registry. The Government has been working on a legislation to facilitate land titling and to speedily resolve land disputes. This effort should be translated into action as soon as possible. In addition, measures are needed to give clear land titles to settlers in colonies, effectively implement Agrarian Services Amendment Act of 1991 (by which all restrictions on land use were removed) and prepare a program to divest to the private sector all government-owned land presently devoted to agriculture. Third, there is a need to boost the availability of efficient support services. This entails (i) making more efficient use of irrigation water through better O&M, supported by cost recovery and farmer involvement in canal maintenance; (ii) strengthening research and extension by allocating more budgetary resources, greater outreach effort to disseminate research findings, involving users in determining the research agenda, and using private sector research facilities; (iii) making a careful assessment of rural infrastructure needs, particularly rural roads, and support priority needs through public expenditures. A careful review of rural financing services is also needed; based on this review there may be a need to support the development of rural financing institutions. The nature of this support should be based on lessons of experience from other countries (e.g. the Grameen Bank in Bangladesh, the Kuppedes in Indonesia). Policies for Enhancing Labor Market Flexibility and Labor Quality: An important feature of the East Asian experience has been the flexible working of the labor market, with employment and remuneration decisions largely based on market factors. The growth in employment and wages came from rapid economic growth and not through non-market government interventions. Labor market in Sri Lanka has not worked as efficiently. An estimated third of total employment, belonging to the public sector and large private enterprises, has been subject to various types of non-market interventions. As a result, wages in the regulated sector have been higher than their market clearing levels and there is an excess supply of labor in this market. Since employment in this regulated sector is associated with a much larger share of total investment and value-added than the employment share, the economy-wide ramifications of the distortions for economic growth and employment have been quite serious. The adverse effects have been compounded by the relatively low quality of Sri Lanka's labor force, notwithstanding Sri Lanka's overall strong achievements in human development. Policies for enhancing the flexibility of labor market would need to eliminate the distortions in the employment and remuneration decisions in the regulated market. The main reforms include removing government involvement in the hiring and wage setting decisions in the state plantations and other corporations, deciding civil service employment on the basis of needs rather than political expediency, and eliminating the TWA (termination of workers act which restricts labor mobility in large public enterprises) . Admittedly, implementing these reforms will pose a major challenge in view of political sensitivities. On the other hand, a failure to address these distortions might seriously constrain Sri Lanka's ability to attain the NIC status. The situation calls for a sensitive handling of these reforms but not an indefinite postponement.  - vi - Reforms for improving the quality of labor would need to address the concerns in both general education and vocational training. There is a need to improve the composition of public spending on primary and secondary education. At present most resources go to finance teacher salaries, leaving very little for materials and supplies. Teachers are generally in excess supply in the cities. So, a reduction in the number of teachers in the cities will reduce the salary bill and release resources for essential school supplies and materials. Greater emphasis on science and English as a second language will improve the ability of students to respond better to the needs of the job market. More generally, a careful review of the curriculum and appropriate revisions in light of Sri Lanka's development needs and market demand will help improve the quality of human capital. Efforts to enhance the quality of trainers--the teachers--are also needed. Sri Lanka needs to rethink its skill development program. International evidence shows that training programs are most likely to succeed when sponsored and implemented in collaboration with prospective employers. There are various ways how Sri Lanka could reorganize its training programs to make it demand based. The main need is to ensure a strong involvement of the private sector (employers) in the design and implementation of the training programs.  I. INTRODUCTION 1. Sri Lanka has made very impressive progress since independence in reducing the incidence of poverty. This is reflected in terms of human development as well as when poverty is measured in consumption terms. Indeed, Sri Lanka's progress with human development stands out when compared with other countries at similar income levels. The progress with poverty reduction was made possible by a combination of two factors: (i) a strong push on human development and equitable income distribution through public expenditures; and (ii) a relatively good growth performance--3% per capita GNP growth rate over the past 40 years, exceeding the rate achieved in South Asia, Latin America and low income countries in general. 2. Despite this progress, Sri Lanka' s poverty problem remains substantial. While the improvement in human development has been outstanding, Sri Lanka remains a low income economy. Although Sri LankaIs per capita growth rate compares favorably with most of the developing world, it falls much short of the growth rate achieved by the high-performing East Asian countries (see Table 1). Thus, in 1960 Sri Lanka's per capita income in nominal dollars was 276% higher than in Indonesia, 46% higher than in Thailand, about the same as in South Korea and about 509 lower than in Malaysia. Some 30 years later, Sri Lanka's per capita income is now less than a 12th of that in Korea, only a fifth of that in Malaysia, less than a third of Thailand and 24% below that in Indonesia. A stronger growth performance along the lines of the high-performing East Asian economies would have allowed much deeper reduction in Sri Lanka's poverty. Table 1: Per Capita GDP in Sri Lanka and Comparators, 1960-92 (US Dollars, Current Prices) At Nominal Exchange Rate Per Capita Real GDP Growth Rate 1960 1992 (1960-92, % p.a.) Sri Lanka 142 540 2.5 Korea 156 6790 6.5 Malaysia 275 2790 4.2 Thailand 97 1840 5.2 Indonesia 51 670 3.4 Source: WDR (various issues). 3. What factors explain why, despite solid human development, Sri Lanka has fallen far behind the dynamic East Asian economies in raising its per capita income? What are the key lessons for the future, especially in view of Sri Lanka's aspirations to achieve the status of a newly industrializing country (NIC) around the turn of this century? These are some fundamental questions the Sri Lankan policy makers need to ask in preparing to address the development challenges it faces in the mid-1990s and beyond. 4. The objective of this paper is to analyze growth performance in Sri Lanka over the past three decades with a view to identifying the key factors that explain past performance and drawing policy lessons for the future. The analysis hopefully will provide a useful input to the Sri Lanka's own search for the right policy mix aimed at improving the living standards of its population. The paper is organized as follows. Section II contains a brief review of the cross-country growth analysis literature. Based on this review, an aggregative growth model for Sri Lanka that links policies with the growth outcome is developed and estimated using time series data. Given the inherent limitations of an aggregative time series growth model, in Section III the discussion focusses on  - 2 - a more detailed, heuristic review of the key policy parameters that underlie Sri Lanka's performance. The analysis here uses the East Asian experience as a paradigm. In particular, specific attention is given to the experience of Korea, Malaysia, Thailand and Indonesia as comparators. Finally, the implications for future policy are indicated in Section IV. II. HOW POLICIES INFLUENCE ECONOMIC GROWTH--STATISTICAL ANALYSIS 5. In the traditional neo-classical framework a la Solow (1956), the rate of economic growth is determined by the rate of accumulation of capital and labor and by exogenous technical progress. In this framework, policies have only a transitory effect. The steady state per capita growth rate is dnly a function of exogenous technical change, unrelated to policies. Since the seminal work of Solow (1956), there has been considerable interest in examining whether long-term economic growth is exogenous (unrelated to policy) or endogenous in the sense that technical change is influenced by policies and therefore becomes an endogenous variable. Interesting modifications along these lines include Romer (1986) and Lucas (1988) . Romer and Lucas endogenize technical progress by allowing for externalities in the accumulation of physical and human capital respectively. Thus, policies that encourage or discourage physical or human capital will affect long-term growth accordingly. Much empirical work has been devoted in recent years to identifying policies that influence physical and human capital accumulation and the efficiency with which these are used. 6. Sunmary of Cross Country Evidence: A brief examination of the empirical growth literature suggest a role for the following determinants of growth. The results reviewed and reported here are only indicative and are not meant to be exhaustive or conclusive. o Initial conditions: The role of initial conditions is expected to pick up the influence of two factors. First, in endogenous growth models the initial endowment of human capital will allow a faster rate of technological progress; so countries with better equipped initial human capital stock are expected to grow faster, ceteris paribus. Second, in the neoclassical model, poorer countries with low initial capital-labor ratios are expected to grow faster because of higher marginal product of capital. Using a sample of 98 countries for the period 1960-85, Barro (1991) finds that for a given starting value of per capita GDP, a country's subsequent growth rate is positively related to the initial level of human capital as measured by enrollment ratios. As well, holding constant a set of variables including human capital, initial per capita GDP is negatively related to per capita income growth. A similar result is also reported by the World Bank's East Asian Miracle Study (1993). o Fiscal policy: A number of fiscal variables has been used in empirical work. Three commonly used variables are the share of government consumption in GDP, the share of public infrastructure investment in GDP, and the GDP share of public spending on health and education. Government consumption is expected to lower growth through distortionary taxation or expenditure programs. Public investment in infrastructure is expected to raise the productivity of private investment and growth. Public spending on education and health is expected to raise growth by increasing labor productivity. Barro (1991) finds a negative association between government consumption (excluding defense and education) as a share of GDP and per capita GDP. Easterly and Rebelo (1993), in a cross-country study for the period 1970-88, find that public investment in transport and communication is positively correlated with growth. Similar evidence on the positive impact of public infrastructure investment on growth was also reported by Baffes and Shah (1993).  o Outward orientation: Greater outward orientation is expected to have a positive impact on growth, resulting from better access to markets, finance and technology. Many studies have found a positive influence of outward orientation on growth. These include Havrylyshyn (1985), Edwards (1989) and Dollar (1990). In a more recent study Dollar (1992) obtains results which show that trade liberalization and appropriate exchange rate management improves growth performance. o Monetary and financial policies: Empirical investigations of the role of monetary and credit policies, including the impact of inflation, on growth have yielded mixed results in a cross-country framework. The East Asian Miracle Study (1993) finds a significant negative relationship between inflation and growth in a cross country framework. But time series investigation for Korea, Taiwan and China, however, do not yield significant coefficients. Regarding financial development and growth, Goldsmith (1969), Mckinon (1973) and Gelb (1989) find a positive relationship between growth of domestic financial market and economic growth. However, one unresolved issue is the direction of causality. Also, Gelb's finding of a significant positive statistical relationship between growth and real interest rate has been called in question by the East Asian Miracle Study. o Inequality and growth: Income distribution can affect growth through different channels: increased inequality causes greater conflict over distributional issues, thereby encouraging greater intervention in the economy; also, greater inequality can hinder human capital formation. Alesina and Rodrik (1991) found that higher inequality tends to lower growth in the subsample of democracies but is insignificant in nondemocratic countries. Persson and Tabellini (1991) find similar results. Clark (1992) found that growth was negatively affected by inequality. 7. Implications for Sri Lanka: The above review suggests that it is possible to investigate Sri Lanka's growth outcome using a statistical framework that seeks to explain this growth in terms of underlying policies--fiscal, monetary, financial, trade and income distribution (see also Box I) 8. The statistical investigation of the determinants of growth in Sri Lanka is premised on the hypothesis that per capita growth rate depends on the level and the efficiency of use of investment. The investment rate and its efficiency in turn depend upon a number of policy variables. These include: fiscal policy, monetary policy, financial policy and trade policy. Fiscal policy is measured in terms of two variables: share of government consumption in GDP and public spending on health and education; monetary and financial policies are approximated by the rate of inflation and the real interest rate; trade and external payments policies are measured in terms of two variables: black market premium and the rate of growth of exports. The role of income inequality could not be investigated due to lack of time series data. The regression was run using OLSQ on annual time series data over the period 1970 to 1992. The data was not extended to 1960 because of concern with the reliability of the data base for some of the earlier years. A slope dummy for the investment rate variable was included to capture the impact of the civil conflict over the 1983-88 period.  -4- Box 1: Role of Initial Conditions What are the possible effects of initial conditions? That is to what extent, if any, Sri Lanka was left behind by Korea, Malaysia, Thailand and Indonesia because it had poorer initial endowment of human. capital or because Sri Lanka was richer and therefore experienced diminishing marginal productivity? Let us first look at the initial level of human capital (see Tables 2 and 3). School enrollment ratios in 1960 were roughly comparable for Sri Lanka, Malaysia and Korea, while Thailand and Indonesia lagged behind. Indeed, Sri Lanka had an edge over the comparator East Asian economies in terms of other indicators of human development. One cannot, therefore, argue that the East Asian economies grew faster because they had better access to human capital. What about the initial level of GNP per capita? Here, it is hard to find any systematic relationship. In nominal dollars, Sri Lanka's per capita income was roughly comparable with Korea, about half of that of Malaysia and much higher than in Thailand and Malaysia. In PPP- adjusted terms, Sri Lanka's income was about the same as in Thailand, lower than in Malaysia and higher than in Korea and Indonesia. So, it appears implausible that these other countries grew faster than Sri Lanka partly because their initial per capita incomes were lower. Table 2: School Enrollment Ratios in Sri Lanka and Comparators (% of Age Group Enrolled) Primary Secondary Tertiary 1960 1991 1960 1991 1960 1991 Sri Lanka 95 108 27 74 1 5 Korea 94 107 27 88 5 40 Malaysia 96 93 19 58 1 7 Thailand 83 113 12 33 2 16 Indonesia 67 116 6 45 1 10 Source: WDR (various issues). Table 3: Other Social Indicators for Sri Lanka and Comparators Infant Life Expectancy Adult Mortality Rate la at Birth Literacy Rate 1960 1992 1960 1992 1960 1992 Sri Lanka 71 18 63 72 75 89 Korea 85 13 56 71 71 97 Malaysia 73 14 57 71 48 80 Thailand 103 26 56 69 68 94 Indonesia 139 66 41 60 39 84 /a Per 1,000 live birth. Source: WDR and World Tables (various issues).  9. The initial regression was run with its full specification (see Box 2) The coefficients of inflation and interest rate variables came out insignificant and the inflation variable had the wrong sign. The results are not very surprising. As noted, the empirical evidence cited in the literature regarding the relationship between inflation, real interest rate and growth is inconclusive. Moreover, it has also been argued that the relationship between growth and inflation may be positive up to a certain rate of inflation and then turn negative for higher rates of inflation [see Dorrance (1966) and Thirlwall (1978)]. Indeed, dropping these variables considerably improved the explanatory power of the regression. Share of government expenditure on human capital came up with the wrong sign and was statistically insignificant. This possibly is a poor proxy for measuring the effect of human capital on Sri Lankan per capita growth. Alternatively, since the investment rate is supposed to pick up the effects of spending on both physical and human capital, the separate influence of public spending on human capital did not show up. This variable was also dropped. Results of the final regression are shown in Table 4. Box 2: Determinants of Per Capita Growth--Model specification The basic model expresses the growth of per capita GDP (GPGDP) as a function of the fixed investment rate (INVR), share of government consumption (GOVCON), public expenditure on human capital (PEHK), the inflation rate (INV), the real interest rate (RINT) the black market exchange premium (BLPREM) and the growth of exports (GEXP). Thus, GDGP=F(INVR, GCON, PEHK, INF, RINT, BLPREM, GEXP) ............. (1) Linearizing: GDGP= ao + a, INVR + a2 GCON + a3 PEHK + a4 INF + a5 RINT + a6 BLPREM +a7 GEXP ................. (2) The expected signs of the coefficients are: a1>0; a2<0; a3>0; a4<0; a5>0; a6<0; a7>0 10. The statistical findings reported in Table 4 show that the coefficient of the black-market premium is significant at the 95% level. The coefficient of the share of government consumption in GDP is also significant at the 95% level. The coefficient of the investment rate is significant at the 90% level. Also, the civil conflict adverse effect dummy is significant at the 90% level. However, the coefficient of the growth rate of exports is only significant at the 82 % level. Overall, the regression performs reasonably well and explains about 43% of the variance in Sri Lanka's per capita growth. More importantly, the economic interpretation of the results are very meaningful.  -6- Table 4: Sri Lanka: Determinants of Per Capita Economic Growth-- Regression Results (1970-92) Variable Constant INVR GCON BLPREM GEXP INVDUM Coefficient value 2.824 0.120 -0.198 -0.015 0.026 -0.051 t-ratios (1.35) (1.71) (-1.73) (-1.96) (1.25) (-1.79) Adjusted R2 = 0.43 D.W. = 1.81 F = 4.37 Note: INVDUM = Slope dummy for the INVR with 1983-88 period taking a value of 1, and 0 otherwise. 11. The main implications of this statistical investigation can be summarized as follows: o Consistent with the findings of the cross-country regression analysis, per capita GDP growth in Sri Lanka is significantly related to the investment rate. Causality test (Granger Causality test) confirms that the causality is from investment rate to per capita growth and not the other way round. O Surprisingly, however, the significance of the coefficient is at the 90t level and not higher. This can be explained by the relatively low productivity of the investment effort. Indeed, total factor productivity estimates show that Sri Lanka experienced negative factor productivity over a substantial part of the 1970-92 period. O As expected, the civil conflict reduced the efficiency of investment. O There is significant evidence that larger government consumption adversely affected growth in Sri Lanka. This is consistent with Barro's findings from cross-country analysis [Barro (1991)] . The result is quite appealing because the share of public consumption has been going up in Sri Lanka, with a large part being diverted to non- productive uses. Much of the revenues raised to finance government consumption has relied on instruments with relatively high efficiency costs--international trade and domestic sales taxes. o The highly significant negative coefficient for the black market premium variable suggests the negative role played by foreign exchange controls in Sri Lanka. More broadly interpreted, this variable is also a proxy measure for macroeconomic instability. The significant negative correlation with growth is consistent with the findings of the East Asian Miracle Study which shows that the high-performing East Asian economies had a relatively stable macroeconomic environment. o There is some support for the positive role of export performance for growth. However, the low statistical significance of the coefficient casts doubt on the reliability of the relationship. This probably reflects the fact that trade liberalization policies have been implemented seriously only after 1977 and, moreover, a substantial unfinished agenda remains.  IV. POLICY DETERMINANTS OF SRI LANKA'S GROWTH--A DISAGGREGATED APPROACH 12. In the previous section we looked at the statistical relationship between Sri Lanka's growth performance, its investment rate and proxies for a number of policy variables. In addition to the positive impact of investment rate on growth, we found supportive evidence for the role of fiscal policy (share of government consumption in GDP), and overall macroeconomic management (proxied by the foreign exchange black market premium). These results are intuitively appealing and provide useful guidance to policy. However, the constraints imposed by data in terms of quality and quantity, including finding the right proxy, limit the possibility of probing econometrically the role of various other policies in explaining growth in the context of an aggregative time series model. So, below a more disaggregated and somewhat heuristic approach to the policy determinants of economic growth is used to better inform the policy debate. The experience of the East Asian economies is particularly relevant in this regard. The East Asia Paradigm 13. Between 1965 and 1990 average real per capita GNP in the East Asia and Pacific Region expanded by 5.3 e per annum, highest for any region including high-income economies, and much higher than the world average (1.5%). Most of this exemplary growth performance is explained by the performance of eight countries--Japan, the Republic of Korea, China, Taiwan, Singapore, Malaysia, Hongkong, Thailand and Indonesia. Many studies have been undertaken to understand better the determinants of economic performance in these dynamic economies and the relevant lessons for other countries. In particular, a recent study by the World Bank--The East Asian Miracle--has probed quite deeply the factors explaining growth in these economies and the implications for policies. The study identifies the following common characteristics shared by these fast growing economies: o Overall high rates of productivity growth; o More rapid output and productivity growth in agriculture; o Higher rates of growth of manufactured exports; o Higher growth rates of physical capital, supported by higher rates of domestic savings; o Higher initial levels and growth rates of human capital; o Earlier and steeper decline in fertility; and o Declining income inequality and reduced poverty. 14. To what extent Sri Lanka shared these features? Let us start with the favorable ones. As noted earlier, Sri Lanka stands out as a country with very impressive human development indicators. This is both a reflection of favorable initial human capital endowment as well as sustained progress based on Sri Lanka's firm commitment to human development (see Box 3). Similarly, Sri Lanka has done very well in population management. The total fertility rate has been declining steadily from 4.3 in 1970 to 2.5 in 1992. The 1992 rate is close to the East Asia and Pacific average of 2.3. Its current population growth rate of 1.1% per annum compares favorably with the 1.2% for the East Asia and Pacific region.  -8- Box 3. Exolainina Sri Lanka's Proaress with Human Development. : : x~~~n