Revenue erosion through exemption and evasion in poor countries

/3-D ?4 Regional Program on Enterprise Development _ rsHo~~'I 'ion Papers .~V Directed by The World Bank Africa Technical Department Private Sector Development and Economics Division Sponsored by Belgian Administration for Development Cooperation Canadian International Development Agency Danish International Development Agency Finnish International Development Agency Ministry of Cooperation, France Ministry of Foreign Affairs, France Ministry of Foreign Affairs, Italy Ministry of Foreign Affairs, The Netherlands Norwegian Agency for Development Cooperation Swiss Development Cooperation Swedish International Enterprise Development Cooperation  PRcoonal Program on lEntcrprlsc Development Discussion Papers Revenue Erosion Through Exemption and Evasion in Poor Countries Bernard Gauthier and Mark Gersovitz October 1995 The views and interpretations expressed in this study are solely those of the author. They do not necessarily represent the views of the World Bank or its member countries and should not be attributed to the World Bank or its affiliated organizations.  October 1, 1995 Revenue Erosion throuqh Exemption and Evasion in Poor countries by Bernard Gauthier Ecoles des Hautes Etudes Commerciales Montreal, Quebec and Mark Gersovitz12 The Johns Hopkins University Baltimore, MD Abstract The administration of tax statutes materially determines their economic effects. A survey of Cameroonian businesses, a country seemingly typical of its economic level, demonstrates widespread erosion of the tax bases paid by businesses through exemptions and evasion. Businesses seem willing to provide information on these potentially confidential activities. The survey provides information on the number of businesses evading taxes or having exemptions, on the characteristics of these businesses relative to others, on whether businesses maintain information of use to the tax administration, on the contacts between the administration and businesses affecting evasion, and on penalties imposed by the administration. ' Corresponding author: Professor Mark Gersovitz, Department of Economics, The Johns Hopkins University, Baltimore, MD 21218. 2We would like to thank James R. Hines, Howard Pack, Christina H. Paxson and Matthew D. Shapiro for comments and Michel Sylvain for assistance. We acknowledge support from the Regional Program on Enterprise Development (RPED) of the World Bank and the Centre d'Etudes en Administration Internationale (CETAI) of the Ecoles des Hautes Etudes Commerciales (HEC) at the Universite de Montr6al.  1. Introduction Perhaps the greatest difference in how rich and poor countries raise government revenues is the determination of the tax base. In rich countries, tax statutes and associated administrative and judicial interpretations establish rules that determine to a first approximation what is taxed, how, and how much. Some taxpayers certainly do not comply with these rules, but such noncompliance is secondary. Furthermore, taxation is largely anonymous; the tax base is determined without reference to the identities of individuals and businesses. In poor countries, however, statutes alone inadequately approximate the tax base. Instead, two mechanisms powerfully mediate the economic effects of the statutes in poor countries: First, special programs give government officials discretion to override the statutes and favor particular businesses. The many tax holiday schemes (Mintz, 1990) are typically not automatic but depend on application and discretion. Pritchett and Sethi (1994) document discretionary reductions in tariffs. Second, tax administrations in poor countries fail to ensure general compliance with the statutes, so evasion is commonplace if not pervasive.3 Willful evasion is only one form of noncompliance but we do refer to evasion rather than 3 Bird (1992, p.190) and Radian (1980. p.76) share this opinion. Not all observers, however, agree that compliance is of second-order importance in rich countries. Rice (1992, p. 125) refers to IRS estimates suggesting that almost 20% of US corporate tax payments may be evaded. 1  noncompliance.4 We do so partially based on preconceptions about what is happening. More importantly, however, our empirical definition of evasion is based on declarations by the businesses of evading behavior and on indicators of noncompliance that seem to leave little room for anything other than evasion. The information on erosion through exemption and evasion comes from a survey of about 200 businesses in Cameroon on their activities in 1993, apparently the first systematic information on erosion provided by businesses themselves (Rice, 1992). Our work shows that statistics on these sensitive matters can be collected, contrary to others' skepticism (Radian, 1980, p.197). our analysis is motivated by, and is relevant to, the large theoretical literature on tax administration and evasion begun by Allingham and Sandmo (1972) and Srinivasan (1973). These authors assume rationally calculating taxpayers. They then look at the taxpayer's trade-off between successfully evading taxes and being detected and punished by the administration and at the administration's problem of affecting this trade-off through knowledge and enforcement power. We provide information on the fraction of evading businesses, on their characteristics relative to compliers, on whether these businesses maintain information of use to the administration, on the contacts between the administration and businesses relevant to compliance, and on the 4 Taxpayers may genuinely be confused about what to pay, even when experts are unanimous. In addition, the expert community may find the statutes ambiguous until precedent accumulates. Furthermore, some payments, evaded initially, may be paid subsequently as the administration goes to work. 2  penalties that the administration imposes. We also look at the exemptions reported by these businesses. The findings should help condition future theoretical modeling and formal econometric investigation of revenue erosion, methods of analysis that are beyond the scope of this paper. Our goal is to establish modalities for behaviors on which available information has so far been anecdotal and to suggest some associations that bear further investigation. Section 2 discusses the most important Cameroonian taxes, section 3 describes the opportunities for exemption, and section 4 provides information on the collection of the data themselves. Section 5 then defines evasion. Section 6 considers the relation between characteristics of the businesses and erosion and section 7 describes the interaction between these businesses and the tax administration. Section 8 offers some conclusions. 2. The Basic Cameroonian Tax Statutes We focus on three of the most important taxes paid by Cameroonian businesses: contribution des patentes (the patente), impot sur les benef ices industriels et commerciaux (the IBC), and imp6t sur le chiffre d'affaires interieur (the ICAI).5 One or another part of the national Ministry of Finance administers all three. The businesses in the sample were asked whether they paid 5 For information on the Cameroonian tax statutes, we rely on C.I.E. (n.d.) and FIDEAFRICA in addition to interviews in Cameroon with private accountants and Ministry of Finance officials. For the IBC, we relied especially on C.I.E. (n.d.), articles 24, 26, 41, 50 and 108 and on FIDEAFRICA, p. 48. 3  each of these taxes, and if so, how much, and for the ICAI, what the rate was on their principal products. The patente is an annual fee for a business license which must be displayed in each place of business. It is based on broad indicators of business activity, such as inputs of intermediate goods, outputs and the number of employees. The statute specifies that all businesses must pay this tax. The IBC is meant to be a profits tax. It is, however, a complicated tax and the payment is the largest of three minima depending on the juridical status and age of the business. Incorporated businesses must pay at least 35% of their business profits. They cannot, however, owe less than the larger of two other amounts, the imp6t minimum de perception (the IMP) equal to 1% of sales or the imp6t minimum forfaitaire (the IMF) of 600,000 CFA francs.6 (Businesses in their first two years of operations are exempt from the IMP but not from the IMF although the IMF is only due at 50% of its normal value. Businesses in their third and fourth years of operation are liable for only 50% of the normal values of these minima .7) Unincorporated businesses 6During 1993, 50 CFA francs, the Cameroonian currency, equaled one French franc. 7Provisions that benefit young businesses raise the question of the tax definition of age. Businesses have an incentive to declare themselves as young. There appear to be no provisions to stop such behavior and some tax officials whom we interviewed claimed that it does occur while others disagreed. Presumably large visible businesses that maintain the same premises and employees and engage in the same line of business are stopped from doing so. As for small businesses, at least some identify themselves as old and appear to comply with the tax consequences. 4  must pay the maximum of the IBC at 22% and the IMP unless their annual sales are below 60,000,000 FCFA. In the latter case, instead of the IMF, their payment is negotiated on a presumptive basis and paid with their patente. The ICAI is imposed on a business's sales of its domestic production (and on its purchase of inputs).S There are several different rates, depending on the product: the general rate of 9%, a reduced rate of 4.5%, and a rate of 2.25% (for bakeries). The Cameroonian system is advantageous for studying evasion because these taxes normally require minimum payments. The patente must be paid regardless of the business's circumstances if it operates at all. Usually a profits tax does not require payments by businesses without taxable profits, but in Cameroon minimum IBC payments are due regardless of profitability unless the business has a special exemption (section 3). Finally, it is relatively easy to establish that businesses have positive sales and therefore owe the ICAI. Consequently, asking if the business pays the tax at all along with information on its exemptions establishes a necessary condition for the business to be in compliance with the statutes. Without this property, we would have to reconstruct the business's tax filing from survey questions which is probably impractical. 8 Businesses were expected to answer about the amount of tax they owed the government on their sales, and not on what other businesses were collecting on the purchases by the responding business of its inputs. 5  3. Programs of Tax Reduction and Exemption As in many poor countries, Cameroonian businesses have access to a rich array of tax reduction and exemption programs, but they must apply for approval by specific authorities based on various criteria. Some statutory requirements at least implicitly favor larger businesses. We do not, however, have information on applications that have been accepted, or rejected, or that would have been rejected had they been submitted so we cannot provide direct evidence on the granting of these privileges. Businesses can qualify for either of two regimes of preferential indirect taxes, taxe unique (the TU) and taxe interieure a la production (the TIP). These regimes replace the ICAI and tariffs with favorable rates negotiated on a case by case basis. The rates apply to sales by these businesses (and to their purchase of inputs whether domestic or imported). There is an annual fee of 200,000 CFA francs. The many regimes of the investment codes provide other reductions in taxes. The statutory requirements for gaining these benefits are not especially restrictive, although some of the more generous regimes require significant exports, use of domestic natural resources, or domestic value added. The 1990 code requires an annual payment of 100,000 CFA francs if the business is not already paying a TU/TIP fee. Businesses in the sample operate under a regime of either the 1984 or 1990 investment codes. During the first 3 to 12 6  years of operation (depending on the regime), the 1990 code reduces the ICAI and customs duty to 15% and waives the IMF part of the IBC. During the subsequent 8 to 17 years, the code waives the IMP and reduces the IBC rate by 50%, as well as some minor taxes. Some regimes of the 1990 code also reduce taxable income under the IBC if the business meets requirements for the employment of Cameroonian workers. Regimes A and B of the 1984 code offered only reductions in the ICAI and tariffs, but to a very low 5%, while regime C added complete exemption from the IBC some years. The zone franche (the zone) exempts businesses from all taxes for ten years and thereafter requires payment of only the IBC at a much reduced rate of 15%. Businesses must export all output and must not adversely affect the natural environment. Under the convention speciale (the convention), a business may reach any agreement whatsoever about tax obligations. Only these two privileged regimes can exempt a business from the patente. The survey asked each business if it operated under either the TU or TIP regimes, and if so, how much it paid, and what the rate was on the principal products involved. It also asked if the business operated under either the 1984 or 1990 investment codes, and if so, under which regime as well as if the business operated in a zone or under a convention. 7  4. Nature of the Sample Before analyzing the businesses' responses, it is critical to clarify the nature of the sample which importantly affects the interpretation of the data. Table 1 presents some basic characteristics of the sample, such as average size and age. Businesses were interviewed about activities in calendar 1993 as part of a multi-purpose manufacturing survey. The survey was conducted without the involvement of the Cameroonian government by joint teams of university-based interviewers from Cameroon and abroad.9 Participation by businesses was voluntary. Respondents were told about the sponsorship of the survey and assured that responses were strictly confidential. The survey organizers believe that these procedures helped to gain the frank co-operation of respondents. The sample is not drawn randomly from a well-defined and enumerated population. There is no recent official census of businesses. Instead, businesses were contacted from sources such as the 1989 Directory of the Chamber of Commerce, business associations and co-operatives, and searches in the field. The 201 businesses are mostly located in the main economic center, Douala (141), and the capital, Yaounde (57), although a few are from two provincial cities, Edea (2) and Limbe (1). The 9 It was administered in June and July of 1994 by the Centre d'Etudes en Administration Internationale (CETAI) of the Ecoles des Hautes Etudes Commerciales (HEC) at the Universite de Montreal in collaboration with Ecole Superieure des Sciences Economiques et Comptables (ESSEC) of Douala under the auspices of the Regional Program on Enterprise Development (RPED) of the World Bank. 6 8  businesses were restricted to four sectors: food and beverages (56), wood and furniture (43), textiles and clothing (38), and metals and machinery (64). These sectors are generally believed responsible for the vast bulk of manufacturing in Cameroon. To be included a business had to have at least one employee or apprentice in addition to the owner. As far as size and sophistication of the businesses, we believe the sample includes most large- and medium-sized businesses in these sectors. By contrast, we suspect it is missing a large but indeterminate fraction of small and informal businesses. This shortcoming is relevant to a discussion of revenue erosion which depends on business visibility as influenced by size and age among other characteristics. 5. Definition and Interpretation of Evasion This section develops two definitions of evasion, an indirect one based on calculations and a direct one based on the business's own declarations. Under the first, those businesses that neither pay a group of taxes nor report an exemption are evaders. The second uses the answers to why businesses do or do not fear being closed by the tax administration; some answers are admissions or denials of evasion. We constructed the first definition as follows: Each business was asked if they pay: (1) the business license (the patente), denoted P; (2) the business profits tax (the IBC or its minima, the IMF and the IMP), denoted B; and (3) the sales taxes 9  (the ICAI, the TIP or the TU), denoted S.10 If they do not report a payment under one or another of headings (I)-(3), we turned to their exemptions. The zone and the convention regimes may, although the latter need not, fully exempt a business from all the taxes under each of the three headings. Without further information, businesses with these privileges were assumed exempt if they did not pay particular taxes. In addition, a business in its first two years and under the 1990 investment code is exempt from both minima of the IBC (the IMF and the IMP) and so could owe nothing under the IBC if it had non-positive taxable profits. We considered it not to have had to pay any IBC if it did not pay any. Finally, businesses operating under regime C of the 1984 investment code are exempt from the IBC. All other businesses are evaders of any of the three headings of taxes that they do not pay. Figure 1, a Venn diagram, shows the results of these classifications. Each circle gives the number of businesses that pay taxes under one of the three headings (P, B or S) and the overlap with those businesses that pay other taxes. Each area also indicates whether the businesses that do not pay the unpaid taxes are exempt (x) or evading (v). When they do not pay more than one tax, the figure gives the cross-tabulation between evasion and exemption across unpaid categories. Finally, entirely outside the circles are 15 businesses that pay none of '