Electronic copy available at: http://ssrn.com/abstract=1393991

Tax Challenges Facing Developing Countries:

A Perspective from Outside the Policy Arena

Richard M. Bird

International Tax Program

Joseph L. Rotman School of Management

University of Toronto

March 2007

DISCUSSION DRAFT: COMMENTS WELCOMED

Contact rbird@rotman.utoronto.ca

Electronic copy available at: http://ssrn.com/abstract=1393991

2

Preface

The structure and purpose of this paper reflect both its origins and my limitations. As part of a broader study of “Taxation for Effective Governance and Shared Growth” being carried out for the U.K. Department for International Development (DFID), I was asked to prepare a background paper from a tax perspective. The project as a whole is intended to link macroeconomic modelling to the challenges facing developing countries with respect to tax effort, political accountability and sustainable economic growth. Other papers for the project are to consider macroeconomic and political science perspectives. The terms of reference asked authors to present the latest evidence and best knowledge in their areas, to evaluate the findings and suggest both detailed policy options and areas for improved research as well as “practical operational guidance” for the international community and partner governments. In the tax paper, I was asked to consider challenges with respect to tax policy and administration in meeting the domestic revenue mobilization objectives of the Monterrey Consensus,with specific reference to how to achieve politically and administrative sustainable pro-growth revenue collection without unduly burdening the formal economy -- for instance, through more effective and transparent administration and broadening the tax base. The challenge thus set is formidable. This paper is my attempt to meet it. Inevitably, it falls short in some respects, notably when it comes to providing ‘practical operational guidance’ for the international community. This lapse is not accidental. Developing countries, like the rest of the world, exist in a complex and changing international environment, and tax outcomes in many of them have been affected, sometimes adversely, by globalization. My focus here, however, is not on what the world has done to, or should do for, developing countries but rather on what countries should and can do to help themselves, irrespective of how the world treats them. Throughout the paper I suggest many ways in which countries can, if they wish, improve their tax systems in terms of structure, administration, and (at least from my perspective) outcomes. In the course of doing so, I attempt to respond to all the points noted above as best as my reading of the evidence and my own experience enable me to do. In the end, however, I do not suggest many ways in which the rest of the world can intervene and directly help developing countries to improve their tax systems. I do not do so primarily because, as I explain in the paper, one of the most important lessons from the half-century or more that fiscal experts have been in the ‘tax advisory’ business is that one cannot sell good tax policy or administration to those who are unwilling to buy them. If the international community wants developing countries to do more to meet fiscal challenges, its main contribution, I suggest, should be to improve the capacity of the domestic policy communities in those countries to cope with such challenges in their own way, not to tell them in detail how they should do it.1

1 There are, of course, a number of things most developed countries could do to their own tax systems to make the lives of policy-makers in developing countries simpler, and there is also much that could be said about the lack of international fiscal forums directly responsive to the needs of developing countries, but these issues – which I and many others have discussed at length elsewhere -- are not the focus of this paper.

3

Contents

1. Introduction………………………………………………………………………………..….5 2. Increasing Tax Effort………………………………………..………………………………..8

Determinants of tax ratios…………………………………................................................9 Assessing tax performance………………………..……………………………………..15

3. Broadening Tax Bases…………………………………..………………………………….16 The tax mix………………………………………..……………………………………..16 Tax base policy…………………………………..………………………………………20 Taxation and growth………………………..……………………………………21 Tax incentives………………………………..…………………………………..23

4. Lowering Tax Rates……………………………………..…………………………………26 The costs of taxation…………………………………..…………………………………27 Administrative costs…………………………..………………………………….28 Compliance costs………………………………..……………………………….28 Efficiency costs………………………………….……………………………….29 Progressivity and fairness………………………………..………………………………31 Progressive income taxes……………………….………………………………..32 Regressive consumption taxes…………………..……………………………….33

Can VAT be more progressive than an income tax?.............................................33 Tax mix and tax structure………………………………………………………………..33

5. Improving Tax Administration…………………………………………………………….35 Assessing tax administration…………………………………………………………….35 How to fix tax administration……………………………………………………………37 Keep it simple……………………………………………………………………39 The taxpayer as client……………………………………………………………39 Salvation through reorganization?.........................................................................41 The IT solution…………………………………………………………………..43 An example: VAT as a self-assessed tax………………………………………………..44

6. The Politics of Taxation……………………………………………....................................47 Accountability and visibility……………………………………………………………..47 Tax awareness: good or bad?.................................................................................47 What’s in a name?..................................................................................................48 Earmarking and the wicksellian connection……………………………………………..49 Getting prices right………………………………………………………………………53 Decentralization as a partial solution…………………………………………………….54

7. What Should Be Done?.........................................................................................................56 Institutionalize tax reform………………………………………………………………..56 Play the right game………………………………………………………………………58

References……………………………………………………………………………………..64

4

List of Figures

1. Tax revenue as a percent of GDP by GDP/per capita category………………………………10 2. Tax structure by region, percentage of total tax revenue, 1975-2002………………………...18 3. Incentives in Central American and Caribbean countries……………………………………..24 4. Income tax rates in Latin America, 19852002……………………………….……………..…27 5. The distributive effects of VAT……………………………………………………………….33 6. Prospects for VAT success: a subjective appraisal…………………………...……………….45 7. Varieties of earmarking……………………………………………………………………..…50

5

Tax Challenges Facing Developing Countries: A Perspective from Outside the Policy Arena

1. Introduction

Most developing countries face substantial fiscal challenges: spend more; spend better ; tax more; tax better. The development community expects them to meet such challenges while at the same time paying close attention to the implications of the fiscal policies adopted for many topics of interest – the list varies from time with changing intellectual fashion but would likely today include such issues as the effects on informality, poverty relief, and gender balance. Moreover, not only must those in power make decisions on all these matters with an eye to economic sustainability -- large uncontrolled deficits are definitely out of style --but they must also, of course, be concerned with their own political survival. A recent IMF (2005) assessment set out a revenue-to-GDP ratio of 15-20 percent as a reasonable minimum “threshold” for developing countries.2 Since the average tax ratio for developing countries as a whole is already in this range (see Figure 1 below), this may not seem to present much of a challenge. However, many countries fall below this cut-off point. In West Africa in 2003, for example, Guinea, The Gambia, Liberia, and Togo and the Democratic Republic of Congo (DRC) all had tax ratios below this threshold (IMF 2005a). In a recent review drawing on IMF and other data, Fox and Gurley (2005) found that 44 out of 168 countries examined had tax ratios less than 15 percent in the 1990s, with 18 of those that failed the threshold test being in sub-Saharan Africa. UN Millennium Project (2005) has also set the tax bar high, suggesting that most developing countries should mobilize up to an additional 4 percent of GDP through domestic revenue mobilization efforts in the near future, presumably in a sustainable fashion. Of course, there is nothing new about setting such targets: Half a century ago, for example, Martin and Lewis (1956) aimed at a ratio of 17 to 19 percent, while Kaldor (1963) was even more ambitious, arguing that if a country wished to become ‘developed’ it needed to collect in taxes an amount closer to 25-30 percent of GDP. Unfortunately, then as now many developing countries have failed to meet such targets, no matter how often they are assured that such “…expansion is not only necessary – it is achievable through using broad-based revenue sources, such as a value added tax, [and] strengthening tax collection….(UN Millennium Project 2005, 245).” Even among the fast-growing countries of east and southeast Asia, for example, only Korea managed to increase its tax take by the prescribed 4 percent of GDP over the last 15 years or so, while in a number of countries in Latin America and elsewhere the tax ratio actually declined over this period. Experience everywhere suggests further that the hope so often expressed that developing countries can and should achieve this goal largely simply through more vigorous collection efforts is particularly optimistic.3 There is more to understanding how countries may improve their tax effort than simple exhortation to try harder. I discuss this question further in Section 2.

2 ‘Revenue’ (even current revenue) is not identical to (or limited to) tax revenue. It is clear from the context, however, that this source treats the two terms as identical, so I follow the same practice in what follows. (Bird, Martinez-Vazquez and Torgler (2006) analyze both revenue and tax ratios in developing countries and find no great differences in most instances.) 3 One of the best documented cases in which better administration increased revenues markedly in a short time was Argentina’s rapid expansion (from 13 to 23 percent of GDP) over the the 1989-92 period. Morrisset and Izquierdo (1993) estimated that about two-thirds of this increase was attributable to improved administrative effort. As in othr

6

Many who have posed tax ratio targets have, like UN Millennium Project (2005) gone on

tell developing countries how they should meet such targets. Earlier analysts, drawing largely on experience in developed countries during World War II, took it almost for granted that a highly progressive personal income tax (sometimes with marginal rates ranging up to 60 or 70 percent) buttressed by a substantial corporate income tax (often at 50 percent or so) constituted something close to an ideal tax system. Consumption taxes were grudgingly accepted as necessary for revenue purposes, but the feeling one gets from reading most documents of the period is that the sooner such levies were replaced by decent income taxes the better.4 No one talked about local taxes, since all the action was at the central government level. Nor did anyone worry much about the international context since tax policy was considered a domestic affair. Both revenue and redistribution goals, it was often argued, could be achieved largely by imposing high effective tax rates on income, essentially because the depressing effects of taxes on investment and saving were considered to be small.5 In short, to exaggerate only a bit, the conventional wisdom at the time was essentially that all developing countries needed to do to solve their fiscal problems was, as UN Millenium Project (2005) seems also to assume, essentially to “learn to tax” (Kaldor 1963) -- which then, if not necessarily now (see below) usually meant to tax in a properly progressive fashion.

Both ideas and realities have changed over the last half century. Most economists and

many policy-makers now think that high tax rates not only discourage and distort economic activity but are also ineffective in redistributing income and wealth. High direct tax rates are out of fashion: “lower rates on broader bases” has for some years been the common mantra of experts (World Bank 1991). In the more competitive international environment of recent decades, income tax rates on both persons and corporations have been sharply reduced. In Latin America, for example, the average tax rate on corporations fell from 41 percent in 1985 to 29 percent in 2003 and the top rate on personal income from 51 to 28 percent (Lora and Cardenas 2006). Over this period, collections from direct taxes in Latin America increased by only 5 percent (from 4.0 to 4.2 percent). Since trade taxes also declined, the tax share of GDP in the region would actually have declined had it not been for a substantial (70 percent) increase in VAT revenues. Reflecting – indeed, to some extent leading – world-wide trends, VAT has become the mainstay of the revenue system in Latin America owing both to rate increases – the regional average VAT rate rose from 11 to 15 percent in the 1985-2003 period -- and to broader bases and improved administration.6 The combination of declining taxes on international trade as a result of import liberalization (and WTO adherence) and increased competition for foreign investment has motivated similar changes in tax structures in most countries around the world in recent years.

cases, however, subsequent experience in Argentina as in Mexico (Martinez-Vazquez 2001) and elsewhere demonstrate how difficult it is to sustain such increases over time (Bergman 2003). 4 Kaldor (1956) famously proposed an expenditure tax for India, but he did so not because he was against taxing ‘ability to pay’ but because he thought the expenditure base came closer to measuring ‘spending power’ than did income as conventionally defined for tax purposes. 5 Indeed, an extra bonus of high rates was considered to be that they made it easier to lead balky private investors by the very visible hand of well-designed fiscal incentives into developmentally productive channels (Bird 2000). 6 The effect of base and administrative changes is evident because ‘VAT productivity,’defined as VAT revenues as a percentage of GDP divided by the (standard) VAT rate rose substantially in the region over the period. More refined calculations approximating the VAT base more closely show similar results (Bird and Gendron 2007).

7

A major connecting link between the level of taxation and the structure of taxation is the breadth of the tax base, as emphasized in IMF (2005). Indeed, perhaps the only recommendation more frequent in reports on taxation in developing countries than base broadening is that to improve tax administration. As already noted, most experts are almost equally keen to lower tax rates -- although some policy-makers been reluctant to accept this recommendation (like the related recommendation often made by travelling economists to place even more emphasis on domestic consumption taxes, especially VAT) largely because such suggestions are seen by important segments of the public to be little more than code for “increase taxes on the poor.” (I return to this concern in Section 4.) Nonetheless, almost regardless of political stance, the consensus of most fiscal experts these days is that the best ways for developing countries to respond to the tax challenges they face – to expand their “fiscal space” on the revenue axis (IMF 2006) --is threefold: (1) broaden tax bases (especially of consumption taxes), (2) reduce tax rates, and (3) improve tax administration.7 It therefore seems appropriate to organize much of this paper under these three headings. Section 3 therefore considers some aspects of broadening tax bases. To some extent countries are, as it were, passive recipients of base changes both in the long run (for example, as a result of such major economic shifts as urbanization) and in the short run (owing, for instance, to exogenous shifts in commodity prices). How should tax systems be structured to cope with these challenges? On the other hand, countries may, through clever design and better administration, be able to ‘grow’ their effective tax bases to at least some extent, say, by reducing the ‘cost’ of taxation or by influencing the way in which the economy evolves through, for example, encouraging and facilitating the expansion of the formal sector. Section 4 focuses on two key points with respect to tax rates. First, as just mentioned, there may be economic gains from lowering the cost of taxation and an obvious way to do so is by reducing rates. Secondly, tax rates may have (or at least be thought to have) important effects on the distribution not only of income and wealth but also of economic and political power. Although in the end the state of debate on these issues in most countries turns more on subjective perceptions than on the limited amount of objective evidence we can bring to bear, I try in this section to focus on evidence-based (or at least experience-based) arguments. The tax ‘system’ of any country comprises a ‘tax structure’ – a set of laws and regulations that set out what a country presumably wishes to do through its tax policy – combined with the reality of how that structure is actually applied. “Tax administration is tax policy” said an leading IMF expert (Casanegra de Jantscher1990) some years ago. Anyone with experience in the fiscal trenches knows exactly what she meant. However, despite the critical role played by tax administration in shaping policy outcomes only in the last few years has much much serious work been done on this subject. Consequently, much of what is said about improving tax administration in Section 5 is inevitably based on (and no doubt biased by) my own limited experience in a few dozen countries over the years.

7 For three recent surveys, from different perspectives, that basically reach this conclusion, see Toye (2000), Moore (2004), and Heady (2004).

8

Of course, what countries actually do when it comes to reforming either tax structure or tax administration as a rule depends more on the politics of taxation than on the economics of taxation. While this paper certainly cannot consider all aspects of this complex subject, in Section 6 I review briefly three particular ways in which politics and economics interact with respect to taxation -- decentralization, earmarking, and visibility—here, as elsewhere, again drawing to a considerable extent on earlier work carried out on these topics in particular countries. Of course, it is not only with respect to tax politics that we know surprisingly little about many critical issues that may affect the ability and willingness of particular countries to mobilize a ‘target’ level of revenues, let alone how they might choose to do so. In the concluding section of the paper, Section 7, I suggest that the best international aid agencies can likely do to help developing countries meet the many tax challenges they face is by assisting them to create appropriate human and institutional capacity that will help them to make good fiscal choices. Such general advice may not, perhaps, be considered to provide much practical operational guidance. For too many years, however, fiscal experts and others have landed at airports around the world carrying ‘one-size-fits-all’ solutions as if in response to a cargo cult approach to fiscal reform. To change the metaphor, however, such ‘silver bullets’ as revenue authorities and IT on the administrative side and VAT and flat taxes on the policy side have missed the target so often that it seems more than time for sometime to begin biting the real bullet – the need to help countries build capacity to find their own solutions in their own ways. Of course, there may still be an important role in many instances for short-term outside assistance on this or that on narrower issue (such as coordinating customs and internal tax administration or integrating – or not – corporate and personal income taxes). But such advice should be delivered not from on high but only on request. Lenders may of course still wish to impose fiscal conditions of various sorts and those seeking money may choose to accept them, but one should not confuse such transactions with the development and implementation of sustainable fiscal reforms.8 No doubt such games will always be playe. My point here is simply that the real task is to try to raise the competence and knowledge base with which developing countries play in this arena, as well as the transparency and understanding with which fiscal issues are discussed by all relevant players, both domestic and international, in the great tax game.

2. Increasing Tax Effort

Taxes matter. Not only are they necessary to pay for public services but people talk about them, complain about them, and try to dodge them when they can. How people react in turn affects the level and structure of taxation. For the most part those who design and implement tax systems, like those who try to avoid being burdened by them, consider themselves to be eminently ‘practical’ people. However, as Keynes (1936, 383-84) famously said, “practical men, who believe themselves to be quite free from any intellectual influences, are usually the slaves of some defunct economist…..soon or late, it is ideas, not vested interests, which are

8 Hirschman and Bird (1968) made this point in a more general context many years ago.

9

dangerous for good or evil.” Practical tax policy is shaped in all countries not only by both ideas and vested interests but also by the specific political institutions that reflect them.9 Developing countries are no different from others: ideas, interests, and institutions play a central role in shaping tax policy. Since circumstances differ in different countries, no single tax structure can possibly meet the requirements of every country. The ‘best’ tax system for any country can thus be designed only by taking into account not only such factors as its economic structure, its capacity to administer taxes, its public service needs, and its access to other sources of revenue (whether aid or oil) but also such more nebulous but perhaps more important factors as its ‘tax morale’ (Frey 2002), its ‘tax culture’ (Edling and Nguyen-Thahn 2006), and, above all, the level of ‘trust’ existing between people and their government (Bergman 2002). Determinants of Tax Ratios A common way to begin discussions of taxation in developing countries is by taking a look at what taxes exist around the world. One recent survey, for example, looked at data for some recent years for 168 countries, representing every region of the world (Fox and Gurley 2005).10

On average, the tax ratio -- taxes as a share of GDP -- was 18.8 percent for the 168 countries in the sample.11 Tax ratios ranged from well under 10 percent in a few countries, most of which are small and all of which are low income -- for example, Myanmar, Chad, Guatemala, and Central African Republic -- to well over 40 percent in a few high-income countries in western Europe such as France and Sweden. However, some lower-income countries had high ratios such as Belarus, Ukraine, Algeria, and Sudan. Similarly, some higher-income countries, such as the United States, had considerably lower tax ratios than others, with Hong Kong being at the extreme.

This global overview suggests that both opportunity and choice appear to affect tax levels. Countries with access to rich natural resource revenues, such as Venezuela and Azerbaijan, tend to have higher tax ratios than otherwise comparable countries, though such revenues may also be highly volatile, reflecting commodity price changes. Tax ratios in higher income countries appear to reflect more choice than chance. Some, such as Sweden and the Netherlands, have large and centralized governments and others, such as the United States and Switzerland, have smaller and more decentralized governments. Broadly, however, tax ratios vary by income levels. The countries in the sample for which GDP data were available were

9 See, for examples, Daunton (2001, 2002) on the U.K., Gillispie (1991) on Canada, Steinmo (1993) on Sweden, Lieberman (2003) on South Africa and Brazil, and IDB (2006) on Latin America more generally. 10 As Lieberman (2002) notes, there are many problems in assembling such data . For example, data coverage in the study cited varies: for 55 countries, only central government is included, while for 69 countries, general government, including regional and local government, is covered. For 17 countries, it is not clear which governments are included. Data were collected (mainly from IMF and OECD sources) for a period averaging about six years, usually in the mid-1990’s.. 11 This is a simple average, treating each country as a single observation, so that a small island such as St. Lucia receives the same weight as the United States or China. If social insurance contributions are included, the average tax ratio rises to 21.7 percent. Although such contributions raise some interesting points (see the later discussion of earmarking), they are not further discussed here in part because it is not always clear whether they are included in the tax data for some countries: for an interesting discussion in the context of OECD countries, see Messere, de Kam, and Heady (2003).

10

divided into three groups based on per capita GDP.12 As earlier studies (Tanzi 1987) found, taxes tend to rise as per capita incomes rise. The tax ratio rises from about 17 percent in the low-income group, to 22 percent in the medium-income group, and 27 percent in the high-income group. Figure 1, which is based on a somewhat different data set, shows much the same picture: the average tax-to-GDP ratio for low-income countries (in this sample, those with per capita GDP less than USD5,000) is 18.3 percent , for medium-income countries (per capita GDP between USD5,000 and USD20,000 US) it is 22.5 percent, and for high-income countries (per capita GDP greater than USD20,000) it is 29.4 percent.

Figure 1. Tax Revenue as a Percentage of GDP by GDP/Capita Category

18.3%

22.5%

29.4%

23.2%

0%

5%

10%

15%

20%

25%

30%

35%

Low Middle High Total$0 - $4999 $5000 - $19999 $20000 +

Source: Bird and Zolt (2005).

Many factors might explain this relationship. The demand for public services may rise

faster than income (that is, the income elasticity for public services is greater than one), particularly in lower-income countries. For example, urbanization tends to rise with income, the demand for public services is generally higher in urban areas, and it is usually easier to collect taxes in urbanized areas. More generally, the capacity of countries to collect taxes appears to rise as income levels increase although more detailed analysis suggests that the relationship between rising income levels and higher taxes is significant only for the poorest countries.13 As

12 Eighty-nine countries in which per capita GDP was less than USD1,000 in 1999 were classified as low-income, 51 countries where per capita GDP was between USD1,000 and USD17,000 were classified as medium-income, and the 24 countries with per capita GDP greater than USD17,000 were classified as high-income 13 Fox and Gurley (2005) find a simple regression of per capita taxes on per capita GDP has an elasticity of only 0.61, indicating that taxes grow more slowly than income, but a more appropriate quadratic regression on income shows that taxes tend to rise with income but more slowly as income rises. Indeed, after per capita income reaches about USD35,000, the tax ratio actually declines. Linear regressions estimated separately for each income group yielded a significant coefficient on per capita GDP only for the low-income countries.

11

incomes rise in poor countries, the size of the public sector almost invariably becomes relatively larger. After some point, however, this “income determinism” of the tax level declines and the relationship between income and tax levels largely disappears. Rich countries have more choices, but even the poorest countries, while obviously more constrained, appear to have considerable discretion as to how much they raise in taxation.

Taking a longer perspective, Tanzi (1987) reported for the late 1970s an average tax ratio

of 17.8 percent of GDP for the 86 developing countries in his sample. The comparable ratio for the 75 countries for which overlapping data are available in Fox and Gurley (2005) is 18.6 percent, again suggesting a slight increase over time in tax ratios. Revenue growth over time may be summarized in terms of “tax buoyancy,” that is, the percent change in tax revenue divided by the percent change in GDP.14 In the last two decades of the 20th century, the overall buoyancy for the world as a whole was only 1.04, with most broad regional and income groupings being close to this figure. There have been many tax reforms around the world. but tax ratios in recent decades have grown only modestly. Many factors in addition to per capita GDP affect tax ratios. As Tridimas and Winer (2003) note, studies attempting to explain the behaviour of this variable essentially group the possible explanatory factors into ‘demand’ factors , ‘supply’ factors and--albeit often more implicitly than explicitly--‘political’ factors that affect the way in which changes in demand and supply variables enter into and shape policy decisions. Tridimas and Winer (2003) set out an interesting integrative model incorporating all groups of factors, but it has not as yet been applied in a developing country context. However, in a less ambitious recent paper, Bird, Martinez-Vazquez and Torgler [hereafter BMT] (2006), first review ten previous empirical studies of the traditional supply-side (‘tax handle’) variables, and then make new estimates with broadly similar results: per capita GDP and the non-agricultural share of GDP are major factors explaining the size of public revenues in different countries. Like Baunsgaard and Keen (2005), BMT (2006) find that openness is no longer as significant an explanatory factor as in most earlier studies, presumably as a result of the substantial trade liberalization of recent years. If this were the whole story, then from one perspective the situation seems hopeless. It is no great surprise that, say, the availability of as an oil sector is important in explaining how much a country raises in revenue. However, telling a country that wants to raise its tax to GDP ratio to find oil is not very helpful. Supply-side studies make the problem facing most developing countries look more like a dilemma than a challenge: (1) Poor countries tax less because they have less to tax. (2) But to develop their economy (and tax base) poor countries need to spend more on public infrastructure, education, and so on so. (3) Therefore they need to tax more. One way out of this dilemma is obvious: one can argue (as did Kaldor 1963) that the real reason countries do not tax more not so much because nature makes it impossible but because it is not in the interest of those who dominate their political institutions to increase taxes even up to the extent ‘nature’ (and the world economy) permits. Howver, if this were the story,

14 Tax elasticity refers to revenue growth in the absence of any tax policy changes, while tax buoyancy refers to growth including the effects of such changes. In principle, elasticity is a better measure of the growth potential of the tax structure. However, data limitations often force analysts to rely on buoyancy data, and in any case the latter, which incorporate the effects of policy and administrative changes, provides a more policy-relevant measure.

12

economists – who seem professionally reluctant to leap to the revolutionary barricades – would appear to find it difficult to suggest an alternative solution. BMT (2006) offer a somewhat more hopeful version of this story by constructing several new ‘demand-side’ variables (such as quality of governance, inequality, size of informal sector, and tax morale). While crude, their estimates suggest that to a significant extent tax levels reflect people’s perception of the quality and responsiveness of the state. Kaldor (1963) was thus right in the important sense that countries that wish to tax more need to ensure their governing institutions facilitate the achievement of this goal. Enhancing the rule of law, reducing corruption and the shadow economy and improving tax morale are not simple or easy tasks. But perhaps progess along these lines may be more feasible in at least some developing countries than ‘engineering’ fiscal gains by altering the relative share of the non-agriculture sector in the economy or the weight of imports and exports in GDP. In addition to such hard-to-alter facts as the the structure of the economy and the extent of adequate and effective political support, the rate at which revenues increase in any country of course depends directly upon its tax structure, the quality of tax administration, and the pace and nature of economic growth. The income elasticity of a tax system -- defined as the percentage change in tax revenues divided by the percentage change in GDP (or potential tax base, such as personal income) --measures how fast revenues grow relative to the economy. The overall elasticity of the tax system is simply the average of the elasticity of individual taxes, weighted by the percentage of total taxes raised by the tax. With administration of a given quality, the elasticity of any tax depends on the specific characteristics of its structure. For example:

• The elasticity of personal income taxes generally reflects the progressivity of their rate structure and, most importantly, the level of personal exemptions (the zero bracket) relative to average income levels.

• Consumption taxes are more elastic if they cover more rapidly growing goods and services and not just more slowly growing traditional (excise) goods and if they are levied as a percentage of the price (like VAT) rather than on the specific number of units purchased (as with many excises).

• Property tax revenue increases more rapidly when reappraisals occur on a regular basis and when property is fully valued.

In principle, a sensible target might be for revenues to expand in the long run at the same rate as desired expenditures (that is, the income-elasticity for revenues and expenditures should be the same).15 If country currently collecting, say, 10 percent of GDP wants to meet the UN

15 Revenue growth is seldom uniform over time: it generally slows during recessions and accelerates during expansions. Revenue elasticity also tends to rise in expansions and fall in recessions, thus exacerbating the volatility of revenue flows. The elasticity of the corporate income tax is particularly volatile because in a recession corporate profits fall much more precipitously than overall economic growth. Countries that depend heavily on taxation of natural resources such as oil or minerals are especially vulnerable to cyclical swings, with wide swings in commodity prices changing the level of tax revenues. A balanced set of tax instruments rather than a single revenue source will have lower tax revenue volatility, just as an individual investor can reduce the volatility of her investment portfolio by adopting a diversified investment strategy.

13

Millennium Challenge target and increase its tax ratio by 4 percentage points to a (still low) 14 percent, it has to sustain increased revenue performance at least at the world average rate of 4% growth a years for nine years. To expand its revenue ‘fiscal space’ to the targets suggested in IMF (2005) would obviously take even longer. Over the years, it is obvious that many developing countries have had great difficulty in achieving even the much less ambitious target of simply financing their existing level of spending sustainably from domestic resources. Indeed, in order to so many countries have had one tax ‘reform’ after another, virtually always with the main aim of closing short-term revenue gaps. Unfortunately, policies enacted in such economically and politically difficult circumstances have often failed to resolve the underlying basic problem of inadequate revenue elasticity. Of course, not everyone may agree with the argument that a tax system that is more than proportionally responsive to growth (and inflation, although this aspect is not discussed further here) is an essential ingredient of sound tax strategy in an emerging economy. Ideas and interests may differ widely with respect to the appropriate size of government, the appropriate structure of expenditures, and the effects of taxation on growth. For instance:

• If one thinks that the best government is that which governs least, then obviously policies that will, over time, tend to maintain or increase the size of government will not be welcome.

• Similarly, if one thinks that public expenditures less necessary, desirable or sustainable outlays but rather mainly rent-seeking and short-run political interests, cuts in such expenditure may be welcome.

• Finally, even if both the level and the structure of expenditures are acceptable, if one is nonetheless convinced that higher taxes impinge undesirably on growth, then the price of sustaining a given public sector may be considered too high.

Such concerns may be politically relevant to varying degrees in different countries at different times. Of course, exactly the same is true with respect to the opposite beliefs – that more government is better, that public expenditures are well spent, and that taxes do not have very harmful effects. Generalizations about the desirability of improving the income-elasticity of tax systems are thus as suspect as most other generalizations in this diverse world. Nonetheless, barring very extreme cases no sensible argument can possibly support the relatively inelastic tax systems found in many developing countries today. Such countries often face the unpleasant choice of either running unsustainable deficits or neglecting important developmentally and politicaly important unsatisfied public needs. In the circumstances, UN Millennium Project (2005) and others, while making it all look too mechanical and easy, are nonetheless quite right in arguing that measures to improve the tax system and increase revenue elasticity must almost invariably constitute an essential part of sound development policy. The tax policy question facing governments in most developing countries is thus not whether revenues should be increased, but how they can be increased. Tax systems that fail to meet obvious revenue needs clearly damage both their growth prospects and the well-being of their people. Although most major tax changes are crisis-driven

14

(Bird 2004), tax reforms should as a rule focus on long-term rather than short-term objectives. Tax structures should not normally be altered on a temporary basis to meet anticipated current year shortfalls.16 Frequent changes in taxes increase compliance costs and, to the extent businesses make production and location decisions on the basis of a particular tax structure, perhaps also efficiency costs. To provide a stable environment for economic activity, a good tax system should thus yield an appropriate revenue growth path. Although the precise rate at which revenues should grow will differ from country to country in the light of the demand for public services, a useful base line might be that the growth rate of revenues should be at least the same as the overall economic growth rate, unless a country wants to increase (or reduce) the size of its government. Even this very modest goal has not been met in recent years by many developing countries, however. In these circumstances, macroeconomic concerns understandably often tend to dominate all other factors in determining the appropriate level and structure of taxation. When, as just mentioned, the existing revenue system is inelastic but expenditures expand more quickly,

continued tax reforms - reforms that all too often amount to little more than ad hoc changes in tax structure and tax administration intended simply to bring in more money to deal with immediate fiscal needs – are required simply to maintain the current level of government spending (relative to the economy). In an environment of repeated fiscal crises, the changes made to generate additional revenues are almost inevitably those that are most politically expedient, with relatively little attention being paid to good tax policy. Over time the result of this process is likely to be a tax system that resembles a patchwork quilt and deviates substantially from the system that would best achieve the country’s policy goals. Observers of fiscal changes in countries such as Mexico, for example, have sometimes referred to the “fiscal constant, ” by which they mean the observed fact that, despite frequent tax changes, the government’s share of national income has on the whole remained remarkably constant for decades (Martinez-Vazquez 2001). In transition countries such as Ukraine, despite frequent reforms, the tax ratio has even decreased substantially (World Bank 2003). Such tax structure instability may not increase the tax ratio but it inevitably creates uncertainty and probably reduces both private investment and growth. Furthermore, the haphazard tax structure that tends to result from this process tends to increase economic distortions (and thus reduce any social gain from increasing expenditures) as discussed further in Section 4 below.

Crises will continue to occur, and revenue measures are often needed to cope with them. It is easy to say that short-term measures adopted in crisis conditions should be consistent with long-term objectives, but it may be hard to do this in countries in which the only quick way to increase revenue may often be to increase, say, the rates of taxes on trade or excise taxes on particular commodities or to impose taxes on all financial transactions, as has lately become popular in Latin America (Arbeláez, Burman and Zuluaga 2005). Whenever possible, it is much better to respond to short-term fiscal pressures by moving towards a sounder long-term tax structure -- one that both reduces the likelihood of fiscal crises and provides a better basis for any discretionary rate adjustments that may be required in the future. As Section 3 discusses, a

16 The immediate revenue effects of a tax policy change may not reflect its long-term effects, owing both to transitional aspects of the new structure and to the fact that taxpayers often change their behaviour temporarily to take advantage of higher or lower tax burdens in the years before or after the changes.

15

broader tax base is usually the way to go from either an economic or an administrative perspective.

Assessing Tax Performance

International observers (and aid agencies) frequently ask for assessments of the tax performance of developing countries. As a rule, what those who ask for such things seem to be looking for are indicators of performance and output that are comparable over time and across countries and are preferably simple and quantifiable. With respect to tax ratios, for example, the aim is usually to provide a basis for assessing which countries are making a real effort to improve and which are not. If a reliable standard of the level of performance that should be expected from countries at various stages of development can be established, then the actual performance of a country, its "effort", may be assessed in terms of how closely it approaches the standard. Such a standard with regard to tax performance may be established in two quite different ways:

• First, following a great tradition from Kaldor (1963) to IMF (2005) and beyond, one might postulate an ideal (minimum) tax level that any developing country that is "serious" about development, poverty relief, or whatever else is of interest, ought to strive to achieve. However, even when it can be shown that other countries, similar in some ways, have attained the specified goal, it is not clear how much persuasive merit such normative standards possess when applied to any particular country.

• Second, a standard might be established by using the average of ‘comparable’ countries as a benchmark measure of performance. Doing so may be useful and interesting from several perspectives: again, however, using such numbers to assess the performance of any particular country is somewhat suspect. It is difficult, for example, to specify a model of (usable) taxable capacity in any very convincing way, so defining ‘tax effort’ as the ratio of the tax ratio to the ‘potential’ ratio predicted on a basis of a taxable capacity equation, let alone interpreting the difference between the two numbers as showing how much more ‘effort’ a country can reasonably be expected to make, is suspect.17

Either approach may of course provide a useful point of reference but one must be careful not to read too much into benchmarking exercises (Gallagher 2004). There is no easy short-cut to assessing tax performance in any specific country: numbers are necessary, and can help a lot, but they cannot do the whole job (Bird and Banta 2000). The difference between predicted (or target) and actual values in any particular country does not, for example, measure in any meaningful way either the scope for change in that country or the gap that can (or should) be closed through additional effort. Reality is too complex and particularistic to be captured by such mechanistic approaches. Consider, for

17 To illustrate such an interpretaton, one early study stated that, if a country has a low index, "...the main impediment to a higher tax ratio is the unwillingness of the Government to raise taxes (Bahl, 1971, 572).”

16

example, the case of Nicaragua: in the first five years of the revolutionary Sandinista regime the tax ratio rose sharply, from 18 to 32 percent. By the early 1980s the average tax ratio in Nicaragua was 23.6 percent, well above the average not only in Central America but in Latin America as whole. Even after a conservative regime returned to power, Nicaragua’s tax ratio remained well above regional averages. Nothing that happened in economic terms over this period explains either why Nicaragua’s taxes went up so sharply or why they remained so high. Assessing the potential fiscal performance of a country with models that usually leave out critical political (and administrative) realities leaves much to be desired. Well-done quantitative international comparisons are interesting and often suggestive, but they can never be conclusive. On one hand, as Tanzi (1973) noted "If we believe, as we all seem to do, that the tax structures of most developing countries are far from what they should be and that they should be changed, why should we use as our reference point the average of all those distortions?” On the other hand, unless one has more faith in the direct applicability of currently fashionable theories to the very different circumstances of such countries, are we really that sure that we know how those tax structures “should be changed?” There is, in the end, no short-cut that permits us to dodge the hard work of looking in detail at what is going on, and why, in the current context of the particular countries in which we are interested. I turn to some aspects of this question in the next three sections.

3. Broadening Tax Bases

No one likes taxes. People do not like to pay them. Governments do not like to impose them. Unfortunately, taxes are necessary both to finance desired public spending in a non-inflationary way and to ensure that the burden of paying for such spending is fairly distributed. Since even necessary taxes impose real costs on society, good tax policy seeks to minimize those costs. However, tax policy is not just about economics but also reflects such political factors as the degree of concern about fairness. Increased economic growth has in recent years increased disparities between the rich and the poor, and taxes may influence not only the after-tax distribution but also the before-tax distribution through their effects on economic incentives. How people see the distributive effects of tax systems matters. In the end, however, no matter what any country may want to do with its tax system, or what it should do from one perspective or another (ethical, political, or developmental), in reality what it does do is always constrained by what it can do. A country’s economic structure and its administrative capacity as well as its political institutions all tend to reduce the tax policy options available. Nonetheless, some options almost always exist. The holy trinity of options proposed by most fiscal experts are broader bases, lower rates, and better administration. I begin the review of these options in this section by considering what gets taxed – the tax base.

The Tax Mix

The manner in which countries raise taxes differs as widely as the amounts they raise. The pattern of taxes in any country depends upon economic structure, history, the tax structures

17

found in neighboring countries, administrative capacity, and political institutions. Tax design is, for instance, strongly influenced by economic structure. Many countries have a large traditional agricultural sector that is not easily taxed. Many also have a significant informal economy that is largely outside the formal tax structure. The potentially reachable tax base thus constitutes a smaller portion of total economic activity than in developed countries. To some extent, of course, the size of the untaxed economy is itself partly a function of tax policy (Alm, Martinez-Vazquez and Schneider 2004). For example, high social insurance tax rates in many countries in eastern Europe and Latin America foster the informal economy both by discouraging employers from hiring and by encouraging the under-reporting of wage levels.

Different countries may also attach different importance to such commonly accepted characteristics of a good tax system as fairness, economic effects and collection costs. Fox and Gurley (2005) found that consumption taxes accounted for almost 40 percent of total taxes around the world, with income taxes (including special taxes levied on extractive industries) being almost equally important. Within the income tax category, personal income taxes were a bit more important than corporate taxes (including extractive taxes). Within the consumption tax category, value-added taxes (VATs) accounted for about 40 percent of the total, but excises were almost equally important. Most remaining tax revenues come from taxes on imports and exports. A similar picture is shown for a longer period on a regional basis in Figure 2, which shows that income taxes play a much smaller role in the tax mix (24.3 percent) in developing countries than in developed countries (38.6 percent). Moreover, personal income tax revenues are often three to four times corporate tax revenues in developed countries, while in developing countries corporate tax revenues often exceed personal income tax revenues, sometimes by substantial amounts (Tanzi and Zee 2000). As a percentage of GDP, personal income tax revenues in developed countries average about 7.2 percent compared to only about 1.9 percent for developing countries. The differences in the relative use of income taxes and the relative proportions of revenues from personal income taxes and corporate taxes are even more pronounced when examined on a regional basis. In particular, revenues from personal income taxes accounted for only 5.5 percent of total tax revenues and 1.0 percent of GDP in Latin America compared to 10.6 percent of total tax revenues and 1.8 percent of GDP in Asia and a perhaps surprising 13.6 percent of total tax revenues and 2.8 percent of GDP in Africa.

A country’s revenue structure depends upon its economic structure. In small island countries such as Barbados, for instance, international trade taxes may play an unusually important role. Trade taxes tend on the whole to be more important in poor countries, where they account for 24 percent of tax revenues compared to only 1 percent in rich countries. Trade taxes (mainly customs duties) decline steadily as countries become more developed.18 I t is thus the poorest countries that face the greatest challenge in replacing such revenues when they liberalize trade. As Baunsgaard and Keen (2005) demonstrate, many of these countries have not been able to rise to this challenge in recent years.

18 The coefficient in a regression of per capita GDP on international trade taxes as a share of GDP is negative and

statistically significant. An interesting exception are the transitional countries of central and eastern Europe which -- although many of them fall within the low-income group as defined here -- have traditionally relied little on trade taxes (Martinez-Vazquez and McNab 2000)

18

Figure 2. Tax Structure by Region, Percentage of Total Tax Revenue, 1975-2002

Total Individual Corporate Total

General

Consumption Excises

International

Trade

North America

1975-1980 78.4% 56.9% 20.5% 15.0% 7.7% 6.5% 6.6%

1986-1992 78.8% 63.5% 14.4% 17.0% 9.8% 6.3% 4.3%

1996-2002 83.3% 66.3% 15.8% 14.8% 8.8% 5.1% 1.8%

Latin America

1975-1980 32.7% 11.1% 17.6% 40.4% 17.1% 19.3% 26.8%

1986-1992 31.1% 8.5% 17.6% 47.3% 20.9% 21.0% 21.5%

1996-2002 30.4% 6.2% 18.5% 56.3% 34.0% 16.1% 13.3%

Western Europe

1975-1980 42.7% 33.3% 8.5% 50.6% 28.6% 16.5% 6.7%

1986-1992 43.4% 32.9% 9.3% 53.4% 33.4% 14.9% 3.2%

1996-2002 47.2% 32.8% 13.0% 52.4% 31.8% 15.0% 0.3%

Asia

1975-1980 38.8% 22.9% 20.5% 37.2% 14.3% 18.3% 24.1%

1986-1992 39.3% 20.8% 19.2% 39.5% 17.4% 16.7% 21.2%

1996-2002 46.9% 24.2% 21.4% 40.2% 19.6% 15.3% 12.9%

Africa

1975-1980 32.1% 14.6% 16.1% 29.7% 18.4% 13.5% 38.2%

1986-1992 27.4% 14.6% 11.4% 31.9% 18.3% 11.9% 40.7%

1996-2002 30.7% 17.7% 11.6% 36.2% 21.8% 11.3% 33.2%

Income Tax Domestic Goods and Services

Sources: Bird and Zolt (2005)

On the other hand, the higher the level of per capita income the more countries rely on

direct taxes, especially those on personal income. Less strikingly, domestic consumption taxes also become relatively more important as countries become more developed.19 These differences reflect both differences in economic structure and also differences in collection capacity, with low-income countries raising more revenues at the border where relatively few collection points need to be controlled. For the same reason, such countries are more likely to rely on excise taxes such as those on tobacco and alcohol for significant shares of their revenue. VATs, like direct taxes, tend to require not only more effective tax administration but also taxpayers who are more sophisticated -- both conditions more likely to exist in more developed countries.

The most striking trend in tax mixes around the world in recent years has been the increase in the share of revenues generated by taxes on domestic consumption taxes. Particularly notable has been the widespread adoption of broad-based VATs (Ebrill et al. 2001), which rose

19 The income elasticity of direct taxes is 0.80, for consumption taxes 0.61, and for trade taxes 0.09 (Fox and Gurley

2005).

19

from 34 to 40 percent of all consumption taxes even in the short period considered by Fox and Gurley (2005). The 4.1 percent rise in consumption taxes as a share of total collections has been almost exactly offset by the 4.3 percent decline of trade taxes in the same period (Fox and Gurley 2005). Trade taxes have declined even more over the longer term: in 1981, for example, trade taxes accounted for 30.6 percent of developing country revenues (Tanzi 1987), compared to only 24.3 percent for the same countries in 1998.

How countries structure tax systems also depends upon such factors as the need and desire for increased public services and the capacity to levy taxes effectively as well as preferences for such public policy goals as attaining a desired distribution of income and wealth and increasing the rate of national (and perhaps regional) economic growth. In a recent study based on observations for 100 countries over the 1975-92 period Kenny and Winer (2006) show that, unsurprisingly, countries tend to utilize all tax bases (including some not normally included in ‘tax ratio’ studies such as seignorage) more as tax levels rise: countries that spend more, tax more.20 More interestingly, the degree of reliance on different tax bases over time increases more on bases that become relatively more important. For example, as oil production and prices increases, oil countries get more revenue from this source.21 In addition, Kenny and Winer (2006) demonstrate that –along the lines of the traditional ‘tax handles’ approach (Musgrave 1969) -- taxes on particular bases tend to increase when the administrative costs of imposing those taxes decline. For example, rising education levels lower the cost of imposing personal income taxes and are hence associated with more reliance on such taxes. Finally, and in some ways most interestingly, Kenny and Winer (2006) suggest that a critical factor influencing tax structure choices is the extent to which reliance on particular tax sources can be translated into effective political opposition.22 From this analysis, they draw the conclusion (p.209) that “…the onus should be on tax reformers … to justify why any particular country’s tax mix should be substantially altered in relation to its existing political equilibrium.” I return to this important point later. Developing countries face many difficult challenges in designing and implementing suitable tax systems. Many countries have large traditional agriculture sectors that everyone finds difficult to tax (Bird 1974). Other significant components of the potential tax base lurk in other equally “hard-to-tax” sectors ranging from small business and the informal economy to cross-border investments (Bird and Wallace 2004). Such countries in the past have often relied heavily on border taxes on international trade, but this tax base too is becoming increasingly hard to exploit in the face of pressures for trade liberalization. Economic growth generally expands tax bases and such growth is often encouraged by (and usually results in) closer involvement with the international economy. As countries develop, the mass modern production and consumption activities on which the tax systems of developed countries rest -- taxes on wages and personal income, on corporate profits, on value-added -- expand and need to be brought into

20 Similarly, Barreix and Roca (2006) notes that if one compares OECD countries (average tax ratio of 35.9 percent) to Latin American countries (average tax ratio 20.2 percent), the latter collect less as a share of GDP from every tax source. 21 On the other hand, as mentioned earlier, expanding trade has in recent years not been associated with increasing dependence on trade taxes (Baunsgaard and Keen 2005). 22 For a related analysis of tax structure developments in a single country (Canada) over a very long period, see Ferris and Winer (2003).

20

the tax base without overstraining administrative capacity or unduly discouraging the expansion of such activities. But globalization may also exacerbate fiscal problems. The leading edge of growth – outward-oriented development -- may become the bleeding edge of the fiscal system as it becomes more and more difficult to levy taxes effectively on capital income, thus potentially exacerbating internal inequalities and political pressures on the tax system. Life is not easy for tax people in developing countries, and is not becoming any easier. Tax Base Policy

As discussed further in Section 4, good tax policy requires minimizing unnecessary costs of taxation. To minimize costs, one lesson from experience is that tax bases should be as broad as possible. A broad-based consumption tax, for example, will still discourage work effort, but such a tax will minimize distortions in the consumption of goods if all or most goods and services are subject to tax.23 Conventionally, it has also been long argued that the tax base for income tax should also be as broad as possible, treating all incomes, no matter from what source, as uniformly as possible, although recent discussion (Boadway 2005; Barreix and Roca 2006) is beginning to cast increasing doubt on this once conventional piece of wisdom. Much discussion of taxation in developing countries seems to assume, as it were, that “unto each a base (or bases) is given.” If the tax base is indeed ‘given’ then the only policy issue to be addressed is how it can be best exploited -- for example, by reducing exemptions and efforts to bring non-payers into the tax net. While such measures are indeed important in most countries, this focus is perhaps too narrow. In reality, tax bases are not simply ‘given’: they can be ‘grown’ – or destroyed – through the manner in which a given tax burden is collected. For example, taxes may discourage, or encourage, the ‘formalization’ of the economy, they may foster or discourage the growth of such ‘tax handles’ as imports, or they may be used to shape and direct economic growth into particular channels in a variety of ways and for a variety of purposes. In the hurly-burly of politics and the technical wonderland of fiscal analysis it is all too easy to overlook such ‘developmental’ effect of tax policies. Yet, as Emran and Stiglitz (2005) have recently reminded us, in the long run the manner in which (and from whom) taxes are collected may affect not only growth and distribution but also the future level and mix of revenues itself.24 The long-run ‘fiscal impact’ (or ‘revenue productivity’) effects of tax policy and administration decisions needs more attention.

23 In theory, in order to minimize efficiency losses different tax rates should be imposed on each commodity, with higher rates imposed on those goods and services where the changes in behavior are the smallest. To do so, however, requires much more information about how taxes alter behavior than is available in most countries. Moreover, this approach does not take administrative and equity concerns into account. For these reasons, in practice it seems generally advisable to impose a uniform tax rate to the extent possible. A few items, such as gasoline, tobacco products and alcohol, may be taxed at a relatively higher rates, either because of regulatory reasons or because the demand for these products is relatively unresponsive to taxation Such taxes need not be highly regressive with respect to incomes: see e.g. the discussion of alcohol excises in sub-Saharan Africa in Bird and Wallace (2006).

24 The same inference emerges from other recent papers e.g. Auriol and Warlters (2005) and Gordon and Li (2005).

21

Consider, for example, four questions often raised with respect to the challenges facing tax policy in developing countries:

1. Should more reliance be put on consumption than on income taxes? 2. Are broader tax bases always better than narrower bases? 3. Should tax policy be designed to reduce the size of the ‘informal economy’? 4. What should be done with tax incentives?

The conventional answers of fiscal experts to these questions are straightforward: (1) Consumption taxes are better. (2) So are broader bases (3) Every effort should be made to tax the informal sector. (4) Tax incentives are almost always a bad idea. How do these answers hold up when considered from the perspective of the long-run development of ‘tax base policy’?

Taxation and growth Over the past 50 years, there have been many policy prescriptions for economic growth (Easterly 2002). Policy advisors have (in rough chronological order) urged increased capital investment, improvements in education, population control, reduction of government controls on market activities, and loan forgiveness programs as “silver bullets” that would result in improved economic performance in developing countries. Unfortunately, none of these cures has worked as advertised. Nor is there any magic tax strategy to encourage economic growth. Some countries with high tax burdens have high growth rates and some countries with low tax burdens have low growth rates. Looking at the relationship between growth rates and tax rates in the United States over the last 50 years, for instance, reveals that the U.S. has had its greatest periods of economic growth during those years where the tax rates were the highest (Slemrod and Bakija 1996). This does not mean that high tax rates are the key to economic growth since growth rates might have been even higher in those years with high tax rates if the rates had been lower. But it does suggest that there is still much that we do not understand about the relation between taxes and growth. Still, if one thinks about what a ‘pro-economic growth’ tax system might look like, several characteristics come to mind:

• First, a purely growth-oriented tax strategy would presumably tax consumption more than income. The difference between consumption and income is saving, and because income taxes tax saving they tend to discourage the growth of future consumption. One can argue about many aspects of this proposition: Are there offsetting bad effects on distribution? Is saving all that responsive to tax changes? On the whole, however, the proposition that countries that want to ‘grow’ their economies (and thus their tax bases) should place more emphasis on consumption than on income taxes seems robust. Developing countries for the most part already seem to do this, but if one looks beneath the revenue numbers, however, matters look a bit different in several respects. On one hand, most

22

‘income’ taxes in developing countries in practice amount to little more than taxes on labor income and such taxes, since most employment income is consumed, amount to little more than advance payment of consumption taxes. On the other hand, however, even if developing countries get little or no revenue from taxing capital income,25 their tax systems may nonetheless have adverse effects on investment and growth owing to the high marginal effective tax rates (METRs) often implicit in such systems. For example, Mintz (2006) reports METRs ranging between 55.7 percent (DRC) and 0.4 percent (Nigeria) for a broader sample of 45 developing countries.

• Second, it follows from the first point that developing countries aiming for growth should impose little or no taxation of profits in order to avoid discouraging entrepreneurship and risk-taking in addition to saving. Under a pro-growth tax system, no good case exists for taxing normal profits.26 Of course, if there is a personal income tax, some tax on profits may be needed to prevent people from hiding income through companies to avoid personal income taxes, and if there are cross-border capital flows some tax on profits may be required to obtain a share for the host country of profits earned by foreign investors (Bird 2002). But any such tax should be reasonably stable, broad-based, and impose a marginal effective rate that puts the country in a competitive position with respect to attracting capital from abroad (as well as retaining its own savings).

• Third, and often neglected, even in the most growth-oriented tax systems, taxes should kept be as low as possible on the poorest, who must consume to be productive. Just as some so-called “investment” (for example, in luxury homes or elaborate office buildings) may not be conducive to productivity, so some “consumption” is clearly productive. If people lack food to eat or lack basic clothing and shelter, or if they are not sufficiently healthy and educated to engage in productive work, they are unlikely to be economically productive. To some extent, equity (in the sense of not taxing the poor) and growth (in the sense of enhancing the productivity of the labor force) are thus fully compatible in poor countries. For instance, a good VAT in such a country might be one with a somewhat narrower base that excludes items that constitute a significant fraction of the consumption of poor people.27

• Finally, a growth-oriented tax system in developing countries would increase the cost of operating in the informal sector (including the non-monetized traditional sector) in order to encourage movement into the monetary (modern) sector. Imposing higher taxes on traditional agriculture may be difficult politically and administratively, and it may not

25 Neither do developed countries: see Gordon, Kalamokidis, and Slemrod (2004) on US and Becker and Fuest (2005) on Germany. 26 Of course, there is an excellent case for taxing so-called supra-normal profits (economic rent) as heavily as possible. Unfortunately, it is not easy in practice to tell “normal” from an “excess” profit.

27 In the case of Jamaica, for example, exempting five narrowly-defined items cut the VAT burden on the lowest 40 percent of the income distribution in half (Bird and Miller 1989). In principle, a more inclusive tax base combined with a targeted subsidy would of course be preferable but such refinement is not attainable in the circumstances of most developing countries.As Bird and Gendron (2007) suggest, it may be better for a variety of reasons to subject such items to a reduced rate rather than to exempt them completely, but such decisions can of course be made only after detailed consideration of the circumstances prevailing in the country in question.

23

always be equitable, but it is likely conducive to growth by shifting resources away from the traditional agriculture sector. Much the same can be said about presumptive taxes on informal sector activities, although such taxes are clearly generally horizontally inequitable, seldom yield much revenue, and usually suffer from many defects in both design and administration (Bird and Wallace 2004). Nonetheless, such taxes often the only levies effectively imposed on the large informal sectors common in such countries (Schneider and Klingmaier 2004), and as Auriol and Warlters (2005) say, even a bad tax on a ‘good’ base may be a good idea.28

In summary, a purely growth-oriented tax system would be one with a relatively low and stable tax on profits and some taxation of the traditional agricultural and informal sectors, but with major fiscal reliance being placed on a broad-based consumption tax that makes some allowance for necessary consumption. What is conspicuously missing in this picture, of course, is any explicit concern for fairness in taxation. I return to this point in Section 4.

Tax incentives Three of the questions posed above have now been answered from a growth perspective, in each case essentially by adding some qualifications to the conventional answer noted earlier:

• Should taxation focus more on consumption than income? Yes: but there is not all that much difference between the two in the context of most developing countries, some ‘productive’ consumption should not be taxed, and special attention should be paid to rates at the margin.

• Should bases be broader rather than narrower? Yes, but sometimes (as in the productive consumption case and perhaps also in the capital income case) broader may be worse.29

• Finally, should taxes be used to punish the informal economy? Yes, but again great care is needed in both design and administration lest the good become the enemy of the better (as Bird and Wallace 2004 discuss in detail).

The implications for equity and distribution of all this is considered further in Section 4. First, however, to conclude this very selective look at tax base policy, consider the fourth and final of the earlier questions: what should be done with tax incentives?

28 In addition, presumptive taxes may sometimes serve as a useful backstop for taxes in the formal sector. For example, Mexico imposes a minimum tax on the gross assets of a business: if the profits reported for tax purposes exceed a certain minimum rate of return on the assets, the profits tax is applied as usual, but if the reported rate of profits is below the minimal return, the business is instead subject to a tax based on assets. 29 There is also a strong case for heavy narrow excise taxes on a small range of products both in economic and revenue terms. Essentially, a good excise tax system is one that (1) taxes few products, (2) taxes those products correctly, and (3) is administered well. Most excise systems in developing countries fall well short of this standard. Excises are imposed on too many products; the rate structure is not logical; and administration leaves much to be desired. The result is all too often a complex structure that produces less revenue and more distortions than it should and is not well administered: for further discussion, see Cnossen (2006).

24

In contrast to the previous cases, when it comes to tax incentives there is no reason to add a ‘but’ to the conventional answer: eliminate them. Many countries have sought to improve their economy by introducing a variety of tax incentives for investment, for savings, for exports, for employment, for regional development, and so on (Shah 1995): Figure 3, for example, shows what was going on a few years ago in the Central America-Caribbean region. Similar pictures could be painted in other parts of the world for most years: incentives may come and they may go, but they seem always to be with us (Bird 2000).

Figure 3. Incentives in Central American and Caribbean Countries

Country CIT CCA Other Loss WHT

Belize 35.0 No Yes Indef 0.0

Antigua 40.0 Yes IBC 6

Barbados 40.0 Yes Export 9 0-15

Costa Rica 30.0 No Reinvest 3-5

Dominican Republic

26.0 No No 3 18

El Salvador 25.0 No Export 0

Guatemala 34.0 No Export 4

Guyana 35.0 Yes Export Indef

Honduras 35.0 No Export 3

Jamaica 33.3 Yes No Indef 15-22.5

Panama 34.0 No Export 5-15

St. Lucia 33.3 Yes 6

Trinidad and Tobago

45.0 Yes Indef 15-20

Notes: CIT - standard statutory corporate income tax rate (1995); CCA - some form of accelerated depreciation, initial allowance or investment allowance; Other - various special incentives for export industries; or other purposes (e.g. reinvested profits); some minor incentives for specific industries and locations are not indicated; Loss - Number of years loss-carryforward permitted; "indef" = indefinitely; WHT -Withholding tax rate on dividends paid to treaty country; often different rates for royalties and interest, and generally higher rates for all forms of payments to nonresidents in nontreaty countries. Source: Compiled from Boadway and Shah (1995).

Despite their continuing popularity, however, tax incentives are usually redundant and ineffective: they reduce revenue and complicate the fiscal system without achieving their stated objectives. Even to the limited extent that some incentives are effective in inducing investors to behave differently than they would have done in response to market signals, the result is often inefficient and diverts scarce resources into less than optimal uses (McLure 1999). Tax

25

incentives improve economic performance only if government officials are better able to decide the best types and means of production than are private investors. Economists tend to doubt this will often be true since observation suggests strongly that people are likely to spend ‘other people’s (taxpayers) money’ with considerably less care than they do their own. Moreover, experience suggests that, on the whole, such non-tax factors as sound macroeconomic policy, good infrastructure and stable governance are more important factors in locational decisions than tax benefits. A limited role for certain simple incentives may exist as part of a growth-oriented fiscal policy (as some East Asian experience suggests) but tax incentives cannot compensate for the absence of such critical factors.30 Excessive use of tax incentives complicates administration, facilitating evasion and corruption. Once created, concessions usually prove hard to remove and may easily be enlarged at the initiative of taxpayers who lobby for more concessions or simply redefine existing concessions in unforeseen and presumably undesired ways. To maximize the likelihood of beneficial results from tax concessions and to reduce the damage that may be caused by poorly-designed and implemented incentives, countries should at the very least stick to three simple rules (Bird 2000):

• Keep it Simple. The more concessions there are and the more complex they are, the less likely they are to produce desirable results at reasonable cost and the more likely they are to be conducive to evasion and corruption. Concessions should therefore be as few in number and as simple in structure as possible.

• Keep Records. Whatever concessions exist, consistent records should be kept as to who receives what concessions and at what cost in revenue forgone. If a concession is intended to achieve a particular result such as encouraging investment in a particular region, then results in terms of investment, employment, and so on should be systematically reported. In the absence of such information, government is simply throwing money away. Poor countries can not afford to do so.

• Evaluate the Results. It is not enough just to gather potentially useful numbers. Such numbers must be used to be of any value. Ideally, at regular intervals—say annually or at most every three or five years—data on each tax concession should be examined carefully to assess whether the concession is achieving results worth its estimated cost. If not, it should be eliminated.

Few countries at any level of development follow such prescriptions, perhaps because the political advantages of ambiguity outweigh the potential social gains from transparency. One result is that too often tax incentives are seen as costless because their costs are not recorded anywhere. Ideally, not only should the fiscal cost of tax expenditures be reported annually but it should done as part of the annual budgetary process—a so-called tax expenditure budget. Many developed countries now do this, and the result, though not dramatic, has arguably been at least to call attention to the existence and cost of tax concessions and perhaps to reduce their number and scope to at least some extent (Surrey and McDaniel 1985). Some have called for similar tax

30 For a review of tax policy and tax incentives in the Asian ‘growth miracle’ of the 1990s, see Bird and Chen (1998).

26

expenditure exercises in developing countries (Swift 2006).31 At the very least more should be done to assess and evaluate tax incentives on the basis of evidence rather than the largely unwarranted faith in their efficacy that seems the chief explanation of their existence and proliferation around the world.

4. Lowering Tax Rates

As noted earlier, the challenge for many developing countries is not so much whether to increase revenues but rather how to do so. Essentially, there are only three possibilities: raise rates, expand bases, and improve administration. Raising rates within the existing system is the most obvious approach, and it is often also the most politically acceptable approach. On the other hand, it is generally the least economically desirable solution. Raising rates when traditional tax bases are not expanding, new bases can shift abroad, and administration is weak is unlikely to increase revenue much. Even if revenues do increase, so may inequity and inefficiency. Distortions associated with taxation increase (broadly) with the square of the tax rate, so inefficiency increases with rate increases, especially those affecting economically mobile sectors such as foreign investors. Horizontal inequity may also increase because only those few unfortunates inextricably trapped within the tax system bear the burden. When those who comply are penalized and those who cheat escape, a country is not on the path to building a sustainable state revenue system. Moreover,. if as seems often to be the case, it turns out to be most expedient to increase taxes most on the politically weaker segments of society, vertical inequity may also be exacerbated. Few developing countries are likely to be imposing ‘revenue-maximizing’ (average) tax rates,32 but as mentioned earlier many do appear to impose ‘marginal effective tax rates’ (METR) that constitute a major barrier to new investment, particularly in certain sectors. Mintz (2006), for example, found that Argentina, where the statutory rate of corporate tax is 35 percent, imposed a METR of 47 percent on new investment in its manufacturing sector and 44 percent on similar investment in the services sector.33 A more detailed recent (unpublished) study of METRs in 10 African countries found that the marginal effective tax on new capital investment varied from a low of -140 percent under the incentive regime for investment in equipment in the services sector in Mauritius to a high of 71 percent on investment in the finance sector in Mozambique. Such studies demonstrate clearly that statutory rates alone do not tell the tale when it comes to the effects of taxes on economic decisions. Nonetheless, with some exceptions

31 Gomez Sabini (2005) reports on the relatively few estimates of tax expenditures in Latin America. 32 As tax rates rise, economic agents (individuals and companies) seek to lower their tax liability through legal means (tax avoidance) or illegal means (tax evasion). Because of these relationships there is, theoretically, a tax rate that maximizes government revenue called the “revenue maximizing tax rate” (RMTR). At rates lower than the RMTR, an increase in the rate produces a higher level of revenue; at rates higher than the RMTR, reducing the rate will yield more revenue. For discussion, see Bird and Wallace (2005). 33 While this is not the place to discuss the pros and cons of different ways of measuring ‘effective’ tax rates – see e.g. McKenzie, Mintz and Scharf (1997) and Devereux and Griffith (1998) - METR is a well-established way of comparing marginal tax impacts across assets, sectors, and countries.

27

(such as Mauritius), on the whole when statutory rates decline, effective rates are also likely to decline. It is thus striking how much statutory income tax rates have declined around the world in recent decades, as illustrated for Latin America in Figure 4. On the other hand, VAT rates have increased. Although in most cases, as discussed earlier, VAT revenues have replaced declining trade taxes and not income taxes, considerable concern has been expressed over these trends by many who view income taxes as the flag under which progressivity sails. In the remainder of this section, I first review the arguments for reducing the ‘costs’ of taxation that in part underlie recent trends in rates and then consider the evidence on the distributive effect of these tax changes.

Figure 4. Income Tax Rates in Latin America, 1985-2002

Country

Corporate

income tax

1985

Corporate

income tax

2003

Personal

income tax

1985

Personal

income tax

2003

Argentina 33 35 45 35

Bolivia 30 25 30 13

Brazil 45 25 60 28

Chile 10 17 50 40

Colombia 40 35 49 35

Costa Rica 50 30 50 25

Domincan Rep. 49 25 73 25

Ecuador 40 25 40 25

El Salvador 35 25 60 30

Guatemala 42 31 48 25

Honduras 46 25 57 25

Jamaica 45 33 57 25

Mexico 42 34 55 40

Panama 50 30 56 30

Paraguay 30 30 30 0

Peru 55 30 50 27

Trinidad 50 34 70 35

Venezuela 50 34 45 34

Average 41 29 51 28

Median 43 30 50 29

Source: Lora and Cardenas (2006) The Costs of Taxation

Some seem to think that economists overemphasize the costs of taxation and the importance of efficient resource allocation. But taxes do impose real costs, and developing countries -- where resources are by definition scarce – should strive to keep such costs as low as possible in order to free resources for socially desired objectives. Of course, taxes are not themselves a cost but rather just a means of transferring resources from private to public use. Economic costs arise only when the resources available for society’s use, whether for public or

28

private purposes, are reduced by taxes. There are several ways taxes can reduce the size of the economic pot from which all must draw. Administration Costs

To begin with, it obviously costs something to collect taxes. Depending on the types of tax, the actual cost of collecting taxes in developed countries is roughly 1 percent of tax revenues. In developing countries, the costs of tax collection may be substantially higher: Gallagher (2004) reports administrative costs ranging from 0.9 to 3.9 percent for six developing countries; Warlters and Auriol (2005) report results for an additional nine countries in the range of 1.1 to 3.6 percent. Compliance Costs

Less obviously, but more importantly, taxpayers incur “compliance costs” over and above the actual payment of tax. Third parties also incur compliance costs. For example, employers may withhold income taxes from employees, and banks may provide taxing authorities information or may collect and remit taxes to government. Compliance costs include the financial and time costs of complying with the tax law, such as acquiring the knowledge and information needed to do so, setting up required accounting systems, obtaining and transmitting the required data, and payments to professional advisors. Although the measurement of such costs is still in its infancy, studies in developed countries (Evans 2003) suggest that compliance costs are, as a rule, about four to five times larger than the direct administrative costs incurred by governments. One of the few reported studies of compliance costs in developing countries (by Chattopadhyay and Das Gupta 2002, for the Indian personal income tax) actually found compliance costs to be more than ten times higher than in developed countries. Similarly, Shekidele (1999) found compliance costs for excises in Tanzania to be more than 15 times higher than similar costs in more developed countries. As World Bank (2006) reports, compliance costs vary considerably from one country to another: for example, complying with the tax code in Brazil takes a representative firm 2600 hours a year compared to only 30 hours in Singapore and 50 hours in Namibia. With a very few exceptions, costs of paying taxes are generally considerably higher in poor than in rich countries for several reasons. One is the sheer complexity and cumbersome administrative methods employed in some countries. Another is because compliance costs are sensitive to the stability of the tax legislation as well as to such changes in the external environment as inflation, and such factors are more prominent in developing countries. Moreover, since compliance costs are generally quite regressively distributed, and are typically much higher with respect to taxes collected from smaller firms, in many developing countries they arguably constitute a significant barrier to the ‘formalization’ of small and medium enterprises. Indeed, such estimates actually understate the importance of the ‘tax

29

barrier’ in many countries since significant time and resources are also usually required before beginning operation, in order to register for tax purposes and to obtain necessary fiscal licenses.34

Efficiency Costs

The resources used in administering and complying with taxes (or, for that matter, evading them) are real economic costs: they reduce the ability of the economy to provide goods and services. Good tax policy requires keeping such costs as low as possible while also achieving revenue, growth, and distributional goals as effectively as possible. This is no small task. Taxes impose “deadweight” (distortionary) costs that alter decisions made by businesses and individuals as the relative prices they confront are changed.35 In most circumstances, the resulting changes in behavior reduce the efficiency with which resources are used and hence lower output and potential well being. For example, taxes on wages (personal income taxes, payroll taxes, social security contributions) reduce incentives to work. Moreover, at the margin such taxes tend to discourage work in the formal sector.36 Of course, taxes not only alter relative prices (in this case, the after-tax wage), but also income, so the net effect on work reflects both the income effect (do people work more to compensate for lost income?) and the substitution effect (do they work less because they are paid less per hour worked?) The impact of taxes on work depends on the structure of taxes, the nature of the workforce, and the nature of the economy and varies from country to country37 Nonetheless, on the whole it seems plausible that the imposition of significant taxes on wage income earned in the formal sector is likely to hamper the development of that sector. As a result, output (and the tax base) will be lower than would otherwise have been the case. No matter how well government uses the resources acquired through taxation, such negative consequences from tax-induced changes in behavior should be avoided if possible.

Virtually all taxes affect resource decisions at the margin. Consumption taxes, such as the value- added tax, may discourage the consumption of taxed as opposed to untaxed goods

34 Compliance and administrative costs may sometimes be substitutes to some extent – e.g. when taxpayers are required to provide more information, thus increasing their costs while presumably reducing the administrative costs that would otherwise be incurred to secure that information. On the other hand, administrative and compliance costs may also be complementary, as when a more sophisticated administration both requires more information and makes use of it to undertake more audits and other actions. 35 As Adam and O’Connell (1999) note, there are of course exceptions. First, when taxes are ‘lump sum’ – i.e. the tax burden is the same regardless of behavioral responses – there are no distortionary effects. But such taxes are of no importance in the real world. Second, to the extent that taxes fall on economic rents – payments to factors above those needed to induce them into the activity concerned -- they may not affect economic activity. Well-designed taxes on natural resources and land, for example, may thus to some extent produce revenue without economic distortion. Finally, some taxes may not only create no distortions in economic behaviour but may even induce desirable behaviour. Certain environmental levies, for example (even such crude proxies such as taxes on fuel), may to some extent have such effects. “Good” taxes – those with no bad economic effects – should of course be exploited as fully as possible, but most revenue needed to finance government inevitably comes from less ‘harm-free’ sources and hence gives rise to efficiency costs. 36 Note that consumption taxes also discourage work to some extent since they increase the amount of time one must work to pay for goods and services through the marketplace. 37 For a recent country study of the effect to taxing wages, see Alm and López-Castaño (2006).

30

(e.g. housing in some countries). Taxes on gasoline, alcohol, and cigarettes may reduce the consumption of these items.38 Income taxes, because they tax the return to savings, may alter the amount of savings or the form in which savings are held. Failure to tax capital gains until they are realized (when the asset is sold) encourages the holding of assets (lock-in effect). Taxes may also affect investment, and such effects may be especially important when economies are more open to trade and investment. Foreign investors may choose to locate their activities in a particular country for many reasons -- the relative costs of production, access to markets, and sound infrastructure -- but taxes may also influence their choice. When taxes lower the after-tax return on investments, the level of investment and hence growth is lower than it would otherwise be. Corporate income taxes may also influence the decision to incorporate, the composition of a firm’s capital structure (use of debt or equity financing) and dividend policy. The importance of such tax effects is a matter of considerable debate, but the current consensus is that they are much more important than was thought thirty or forty years ago. Efficiency costs of taxation in developed countries are usually estimated to be some multiple of the administrative and compliance costs mentioned above. The lowest estimates of the efficiency costs of taxes for developed countries are at least 20-30 percent of revenues collected, and much higher estimates (ranging well over 100 percent) are common in the literature (Auerbach and Hines 2002).. Such estimates are both hard to make and controversial, so it is not surprising that relatively few such empirical studies exist for developing countries. Broadly, unless public expenditures produce social benefits at least equal to the ‘marginal cost of public funds’ (MCF), they are not worthwhile. A recent study found the average MCF in 38 African countries to be close to those found in a number of developed countries, namely, around 1.2 (Warlters and Auriol 2005). The study suggests that this similarity perhaps results from two offsetting factors: first, developed countries tend to have higher taxes and heavier reliance on income taxes, both of which are associated with higher MCFs; second, developing countries tend to have higher administrative costs and larger informal sectors, both of which are again associated with higher MCFs.39 Whatever their size, efficiency losses from taxation are real. However, they are not directly visible: they arise essentially because something does not happen -- some activity did not occur or occurred in some other form than it would have in the absence of the distortionary tax. Output that is not produced is nonetheless output (and potential welfare) lost, so poor countries need to design taxes to minimize such possible adverse consequence. Unfortunately, the absence of visible concrete evidence means that there is seldom much political weight behind this concern.

38 As noted earlier, not all such effects need be bad: for instance, if tobacco consumption falls, people may live longer, healthier and more productive lives. For further discussion, see Cnossen (2005).

39 Warlters and Auriol (2005) do not take compliance costs explicitly into account. An additional factor, which also cannot easily be taken into account in the CGE framework used for MCF estimates, is that, as Shah and Whalley (1990) argue, since economic rents are more prevalent in the fragmented economies of developing countries, many taxes that might in more integrated market systems impact economic margins may fall on such rents and hence create less distortion.

31

The costs of taxation could be reduced by shifting to consumption taxes (VAT) instead of income taxes. They could also be reduced by downplaying taxes on production since such taxes affect the location of businesses, alter the ways in which production takes place, change the forms in which business is conducted, and so forth. Developing countries often impose such taxes for several reasons. First, countries with limited administrative capacity often find it easier and less expensive to collect sales taxes at the point of manufacture, although (as noted in Section 5), this does not mean that they cannot do so in the form of a limited but more ‘business-friendly’ VAT. Second, to the extent that taxes represent the costs of public services provided to businesses that are not recouped by user charges, businesses should bear the cost via taxation for those services. Again, however, there are more and less efficient ways to do this.40 Finally, even if it is somewhat inefficient, countries with income taxes need to tax corporate income to prevent tax avoidance by individuals as well as to collect taxes from foreign-owned firms (Bird 2002). Of course, as Kaplow (1996) and others have argued, to the extent that efficiency costs of taxation result from rational policy decisions (for example, to redistribute income through the fiscal system), they may be worth incurring. A key question emerging from this discussion is thus whether income taxes are in fact the best or even an effective redistributive factor in developing countries. If they are, the efficiency cost of taxation may be worth paying – though one should of course still strive to reduce administrative and compliance costs. But if progressive income taxes do not do much for distribution, presumably countries should at the very least reduce rates to reduce their marginal distortionary effect.41

Progressivity and Fairness

Fairness is a key issue in designing any tax regime. Indeed, from one perspective, taxes exist primarily to secure equity. National governments do not need taxes to secure funds: they can simply print the money required to fund operations. The tax system may thus be viewed as a mechanism designed to take money away from the private sector in as efficient, equitable, and administratively inexpensive way as possible. Of course, what is considered equitable or fair by one person may differ from the conceptions held by others. Some may stress horizontal over vertical equity, for example, as OECD (2006) argues is increasingly true in developed countries and as Boylan (1996) and Bergman (2003) have suggested is also the case in parts of Latin America. Others may tilt the balance the other way, as ‘progressive’ thinkers have long done. Broadly, horizontal equity requires those in similar circumstances to pay the same amount of taxes, while vertical equity requires appropriate differences among taxpayers in different economic circumstances. Those who have the same ability to pay should of course bear the same tax liability; equally, fairness would seem to require those taxpayers with greater ability to pay to pay relatively higher taxes. Both concepts have considerable appeal. Unfortunately, neither concept is very useful in actually setting up a tax system.

40 Bird and Mintz (2000) argue, for instance, that a low-rate origin-based VAT is the most appropriate form of taxation for this purpose. 41 If countries are below the RMTR mentioned earlier, reducing rates may increase revenues. Lower rates will also reduce the ‘barrier to formalization’ created by compliance and administrative costs (both of which are conventionally measured as a share of revenue collected).

32

I consider here only the question of whether ‘growing’ governments in developing countries on the basis of VATs rather than income taxes is a bad idea.42 First, I ask whether even progressive income taxes in developing countries are very redistributive: the answer is – not much. Second, I then ask whether VATs in such countries are regressive: the answer is – not necessarily. And finally, I ask whether in some circumstances a ‘good’ VAT may not be better than a ‘bad’ income tax; the answer is, of course it may, depending upon the details of both the country and the two taxes.

Progressive Income Taxes?

Taxes may tax the rich relatively more (progressivity) or less (regressivity) than the poor. Are income taxes in developing countries progressive? Even if they are, do they have much effect on the distribution of income and wealth? Chu, Davoodi, and Gupta (2000) surveyed 36 studies of tax incidence in 19 different developing countries: 13 found the tax system progressive, and 7 each found the tax system proportional and regressive; the others had mixed findings or insignificant effects. Most of the reported progressivity came from income taxes. So the answer to the first question is that income taxes in most developing countries are likely somewhat progressive. Indeed, to the extent income taxes do not impinge on the poorest people – those outside the market sector – they are bound to be progressive. Even within the taxed sector, progressive rates mean that the impact of the tax is progressive at least within the group of those who must pay tax on all (most) of their income-- for example because they receive it in the form of wages from a public sector employer. In most developing countries, however, little if any tax is collected from capital income or self-employment (mixed) income and that little is most unlikely to be distributed very progressively. Does tax progressivity matter for distributional policy? As Harberger (2006) argues, it does not matter much: insofar as government policy affects the distribution of income and wealth, providing education (even when financed from regressive taxes) ismuch more powerful than taxing the rich. As Engel et al. (1999) demonstrate, even in a relatively advanced country like Chile with an unusually well-developed and effective tax administration the progressivity of the income tax simply cannot have any significant influence on distributional outcomes. One may of course argue that such taxes share the burden of government more fairly, but one cannot argue that they are effective redistributive tools. Indeed, experience everywhere suggests that trying to impose heavy taxation on the very rich is not likely to be very successful and is all too likely to create problems. As Lindert (2003) shows, it was not by taxing the rich but by taxing the growing middle class that developed countries ‘grew’ large states. In notoriously unequal Latin America, for example, income taxes have only really begun to be relatively efficient and effective revenue raisers recently as (some) countries have begun to bite into the middle class, essentially by combining reduced top rates with lowering the level at which those reduced rates come into play (Lora and Cardenas 2006).

42 For further discussion of a few of the many complex points touched on (or skipped over) in this section, see Bird and Zolt (2005).

33

Regressive Consumption Taxes?

Three types of taxes are levied on consumption in most developing countries: import taxes, excise taxes, and VATs. The most striking recent change in tax structure in such countries has been the extent to which VAT has replaced the other two types of consumption tax. This change has almost certainly made the tax system more progressive (Gemmell and Morrissey 2003), for a number of reasons. First, the impact of trade taxes (especially including the ‘burden’ imposed by their protective effect) is likely regressive in the context of most developing countries. Second, excise taxes (even those justified by good economic arguments such as external effects) are also generally -- with the notable exception of those affecting motor transport -- regressive. Third, VAT in most developing countries is not particularly regressive and may indeed in some be slightly progressive, as Figure 5 shows.

Figure 5. The Distributive Effects of VAT

CoCountruntry Source Selective Summary of Findings

Colombia Steiner and Soto (1999)

VAT found to be slightly regressive

Colombia Rutherford, Light, and Barrera (2005)

An increase in VAT would be relatively progressive with respect to the lowest-income groups

Colombia Zapata and Ariza (2005)

VAT appears to be slightly progressive

Dominican

Republic

Jenkins, Jenkins and Kuo (2006)

VAT found to be progressive across all quintiles of household expenditure; includes estimate of different rates of compliance at different income levels

Ethiopia Muñoz and Cho (2003)

VAT found to be progressive owing to exemptions (especially of in-kind consumption)

Mexico Huesca and Serrano (2005)

A more differentiated rate would both raise revenues and improve equity

Pakistan Refaqat (2003) VAT slightly progressive because of exemptions

Peru Haughton (2005) VAT found to be somewhat regressive

Russia Decoster and Verbina (2003)

Indirect taxes were progressive, including VAT rates

South Africa Botes (2001) Zero-rating actually made VAT a little more regressive

South Africa Go et al. (2005) VAT is mildly regressive

Source: Bird and Gendron (2007), which contains full references to all studies cited. Can VAT be More Progressive than an Income Tax?

A good VAT may even be more progressive than a bad income tax. But a deeper issue is relevant here. Consider, for example, the case of a developing country with a large shadow economy. Income taxes do not reach this sector – and indeed appear to be associated with its expansion (Schneider and Klingmaier 2004). On the other hand, to some extent a VAT functions

34

like a presumptive tax on the informal sector since credits are available only to registered firms and those earning income in the shadow sector are taxed when they purchase formal-sector commodities (Glenday and Hollinrake 2005). Increasing income tax rates in such countries often impinges primarily on government employees (sometimes engendering counterbalancing wage increases that may eat up any revenue gained) and on employees of large, formal market firms (thus discouraging their expansion). Increasing VAT may, on the other hand, to some extent tend to make life in the formal sector relatively more attractive. Such a substitition is not only clearly preferable in terms of growth but also likely increases the horizontal equity of the tax system. Indeed, if combined with, say, increased excises on such important higher-income consumption goods as motor vehicles, the substitution of indirect for direct taxes may well prove more progressive than the usual developing country personal income tax that affects only a limited group of formal sector wage earners Tax Mix and Tax Structure

On the whole, therefore, there is much to be said both for broad-based consumption taxes

like VAT and for a limited set of excise taxes as important components of the revenue system – as indeed they already are in most developing countries. But there are also good reasons for keeping direct taxes on both income and property in the tax mix as well. Corporate income taxes are needed to buttress personal income taxes, to ensure an equitable share of the returns on cross-border investment, and to tap economic rents to some extent. Moreover, to maintain and grow a state, the tax system must tap into those sectors that grow and since growth results in the growth of the employed middle class, a mildly progressive personal income tax (like a VAT) is an important way to ensure state revenues share in the prosperity. Only by visibly taxing the middle class though both income and property taxes can countries use the tax system as it were as a ‘state-building’ tool.43 Similarly, since sustainable tax policy needs to be accepted as fair by those affected, and automobiles and big houses are much more visible than income, the more taxation can be levied on such items, the better.

Of course even the best taxes will not make the poor richer--though bad taxes can certainly

make them poorer. Fiscal attempts at poverty alleviation in developing countries must focus primarily on the expenditure side of the budget, but the poor should not be made even poorer through taxes. Both in terms of the ‘productive consumption’ growth argument made above and in pursuit of poverty relief per se, it may, for example, prove necessary to treat certain “basic needs” items favorably under even the broadest-based consumption tax. At the very least, heavy selective taxes on items that constitute major consumption expenditures for poor people should be avoided. 44

43 The case for property taxes is developed at length in Bird and Slack (2004). The general point made in the text emerges in different contexts in various recent studies: see, for example, Sokoloff and Zolt (2005) and Hoffman and Gibson (2004). 44As Hughes (1987) notes, for example, taxing fuel correctly can be difficult in countries like Indonesia in which petroleum products (in this case, kerosene) are an essential consumption item for the poorest people. As mentioned earlier, the option of using the tax system to deliver income support to low-income people as is done is some developed countries requires both that the tax administration is efficient and that most people file tax returns. Neither condition is satisfied in most developing countries.

35

5. Improving Tax Administration

Reaping revenues from tax rate changes (in any direction) requires that tax administration

is effective to at least some extent. Raising revenues through base expansion requires even better tax administration,. New taxpayers must be identified and brought into the tax net and new collection techniques developed. Such changes take time to implement. The best tax policy in the world is worth little if it cannot be implemented effectively. Tax policy design must take into account the administrative dimension of taxation (McLaren 2003). What can be done to a considerable extent inevitably determines what is done. The importance of good administration has long been as obvious to all concerned with tax policy in developing countries as its absence in practice. One cannot assume that whatever policy designers can think up can be done or that any administrative problems encountered can be easily and quickly remedied. How a tax system is administered affects its yield, its incidence, and its efficiency. Administration that is unfair and capricious may bring the tax system into disrepute and weaken the legitimacy of state actions. In many transitional countries, for instance, the failure to improve tax administration when introducing new tax structures resulted in very uneven tax imposition, widespread tax evasion, and lower than anticipated revenue. In developing countries, corporate tax liabilities are often negotiated rather than calculated as set out in the law. Bribery is sometimes so common that it is considered a regular part of the compensation of tax officials. Such corruption undermines confidence in the tax system, affects willingness to pay taxes, and reduces a country’s capacity to finance government expenditures (Fjeldstad 2005). Tax administration is a difficult task even at the best of time and in the best of places, and conditions in few developing countries match these specifications. Revenue outcomes are not always the most appropriate basis for assessing administrative performance.45 How revenue is raised - the effect of revenue-generation effort on equity, the political fortunes of the government, and the level of economic welfare - may be equally (or more) important as how much revenue is raised. Private as well as public costs of tax administration must be taken into account, and due attention must be paid to the extent to which revenue is attributable to enforcement (the active intervention of the administration) rather than compliance (the relatively passive role of the administration as the recipient of revenues generated by other features of the system).46 Assessing the relation between administrative effort and revenue outcome is by no means a simple task. Neither is improving administrative efforts and outcomes. Assessing Tax Administration

It is useful to think of the problem of tax administration at three levels – architecture, engineering, and management (Shoup 1991). The first level concerns the design of the general legal

45.For example, even if it costs only $1 to collect $1000, it does not follow that to get another $100 in revenue one simply needs to spend an additional dollar on tax administration. Not only are such figures sensitive to tax rates but the marginal revenue yield equals the average only under very special circumstances (Vazquez-Caro et al., 1992). More importantly as Slemrod and Yitzhaki (2002) show, the optimal size of a tax administration is likely to be where marginal revenue exceeds marginal cost, perhaps by a wide margin. 46.In one of the few books on how tax administrations actually function in developing countries, Radian (1980) stresses the extent to which officials tend to be passive recipients of funds rather than active collectors of them.

36

framework - not only the substance of the tax laws to be administered but also a wide range of important procedural features. Once the general architectural design has been determined, the engineer takes over and sets up the specific organizational structure and operating rules for the tax administration. Finally, once the critical institutional infrastructure has been erected, the tax managers charged with actually administering the tax system can do their jobs. One cannot assess how well a tax administration is functioning, let alone suggest how to improve it, without taking into account not only the environment in which it has to function but also the laws it is supposed to administer and the institutional infrastructure with which it has been equipped. To appraise the efficiency or effectiveness of tax administration one thus needs to take into account both the degree of complexity of the tax structure and the extent to which that structure remains stable over time. Complexity and its implications for tax administration has long been a concern even in the most developed countries. Even the most sophisticated tax administration can easily be overloaded with impossible tasks (Hood 1976). Such concerns are obviously critical in countries in which less well-equipped administrators are asked to tackle inherently complex tasks in a generally hostile and often information-poor environment. The life of the tax administrator is made even more complicated by the propensity of many governments, reflecting in part the unstable political and economic environment, to alter tax legislation annually, or even more frequently. Both the complexity of the tax structure and its stability are thus important factors to be weighed in assessing tax administration. Disaggregation of the ‘black box’ of tax administration along such lines is particularly important since the main ways in which to improve administrative outcomes is either to alter the tasks with which the administration is charged or to strengthening the tools with which it is equipped. Simple exhortations to "do better" are of little use to resource-strapped administrators faced with impossible tasks. Nor are the various gimmicks or quick-fixes that seem to come easily to the minds of clever policy designers of much use in resolving tax administration problems.47

Experience around the world demonstrates that the single most important ingredient for effective tax administration is clear recognition at high political levels of politics of the importance of the task and willingness to support good administrative practices -- even if political friends are hurt. Few developing countries have been able to leap this initial hurdle.48 Frequently, urged by international agencies or simply desperate to get more revenues, countries have launched frantic efforts to corral defaulters or to rope in new victims without hurting politically powerful interests and without providing the time, resources and consistent long-term political support needed to do a good job. The widespread reluctance to collect taxes efficiently and effectively without fear or favor may be understandable in countries which are fragile politically, but without such efforts, no viable long-term tax system can possibly be put into place. If the political will is there, the techniques needed for effective tax administration are not a secret.

47 For discussion of a number of such schemes, see Bird (2004a). 48 See, for instance, the telling comparison in Bergman (2002) of Argentina, which conspicuously has not leaped the hurdle, and Chile, which has. As IDB (2006) notes, there is still much we do not understand about why Chile has been able to do so much: however, as Bergman (2002) shows, the willingness of Chile’s leaders – of very different political persuasions – to support effective administration stands out.

37

How to Fix Tax Administration

Three ingredients are essential to effective tax administration: the political will to

administer the tax system effectively, a clear strategy for achieving this goal and adequate resources for the task. It helps, of course, if the tax system is well designed, appropriate for the country in question, and relatively simple, but even the best designed tax system will not be properly implemented unless these three conditions are fulfilled. Most attention is often paid to the resource problem -- the need to have sufficient trained officials, adequate information technology and so on. However, without a sound implementation strategy even adequate resources will not ensure success and without sufficient political support even the best strategy cannot be effectively implemented.

If the political will exists, the blueprint for effective tax administration is relatively straightforward. The tax administration should be given an appropriate institutional form, which may (or may not) mean a separate revenue authority: as discussed later, the jury is still out on this question. It should be adequately staffed with trained officials. It should be properly organized, which usually means an organizational structure based on function or client groups (e.g. large – and small – taxpayer offices) rather than on a tax-by-tax basis.49 Computerization and appropriate use of modern information technology are important, but as discussed later technology alone is not sufficient and these improvements must be carefully integrated into the tax administration. Putting all this into place takes time, resources, direction and effort. But it can be done, as countries from Singapore to Chile have shown.

The first task of any tax administration is to facilitate compliance: make sure that those

who should be in the system are in the system and that they comply with the rules. To do so:

1. Taxpayers must be found. If they are required to register, the registration process should be as easy as possible. Systems must be in place to identify those who do not register voluntarily. Tax authorities should adopt an appropriate unique taxpayer identification system to facilitate compliance and enforcement.

2. The administration needs a process to determine tax liabilities. This may be done

administratively (as with most property taxes) or by some self-assessment procedure (as with most income taxes and VATs).

3. Taxes must be collected. In many countries, this is best done through the banking

system. It is seldom appropriate for tax administration officials to handle money directly.

4. The authorities should provide adequate taxpayer service in the form of information,

pamphlets, forms, advice agencies, payment facilities, telephone and electronic filing, and so on, to make taxpayer compliance with the system as easy as possible.

49What is never sensible is to assign specific taxpayers to specific officials for prolonged periods of time, a practice still common in some countries.

38

This approach rests on treating the taxpayer as a client (albeit not a willing one) to be served and not a thief to be caught. Unfortunately, the latter attitude still prevails in many developing countries.50

Of course, some taxpayers are not honest, so a second important task of any tax administration is to enforce compliance and reduce tax evasion. To do so, the administration needs to understand the extent and nature of the potential tax base in order to estimate the “tax gap.” Without at least some idea of the number and type of individuals and firms not complying, no coherent enforcement strategy can be devised. Sometimes, the problem is that large groups of prospective taxpayers are simply ‘not known to the authorities.’ Or it may be that many taxpayers who are in the system are substantially under-reporting. Without some knowledge of the unreported base, and its determinants, no administration can properly allocate its resources to improve tax collection and ensure everyone bears at least a roughly fair share of the tax burden.

In addition to exploring the nature of the tax gap and extending the reach of the tax system into the informal economy to the extent feasible, close attention must also be paid to such simple tasks as ensuring that those who are in the system file on time and pay the amounts due. Immediate follow-up of non-filers and those whose payments do not match their liabilities is often unduly neglected. Adequate interest charges must be imposed on late payments to ensure that non-payment of taxes does not become a cheap source of finance. Similarly, an adequate penalty structure is needed to ensure that those who should register do so, that those who should file do so, and that those who under-report their tax bases are sufficiently penalized to increase the costs of evading tax. As Harberger (1989) once noted, good tax administration is not rocket science: it’s more like being a good accountant. But good accountants are scarce in most developing countries, especially in the tax department.51

A third major task is to keep not just taxpayers but tax collectors honest. No government

can expect taxpayers to comply willingly if taxpayers believe the tax structure is unfair or that the revenue collected is not effectively used. Even a sound tax structure and sound expenditure policy can be vitiated by capricious and corrupt tax administration. Developed countries took centuries to develop and implement systems to prevent dishonest tax officials from corrupt practices (Webber and Wildavsky 1986). To do so, tax officials must be adequately compensated so that they do not need to steal to live. Ideally, they should be professionally trained, promoted on the basis of merit, and judged by their adherence to the strictest standards of legality and morality. Temptation should be reduced by reducing direct contacts between officials and taxpayers and reducing discretion (and increasing supervision) when they do have such contact. Developing countries trying to sustain relatively large governments on precarious fiscal foundations find it hard to deal with such problems.

There is no single prescription - no secret recipe - that, once introduced, will ensure improved tax administration in any country. Countries exhibit a wide variety of tax compliance

50Some countries seem to think that taxpayers should be expected to respect tax laws even when the government does not – as when it fails to provide rebates in a timely fashion. 51 Perhaps it should be noted, however, that if a country does have only a few good accountants, it is highly unlikely that the socially best use of this scarce resource would be to put them to work in the tax department!

39

levels, reflecting not only the effectiveness of their tax administrations but also taxpayer attitudes toward taxation and toward government in general. Attitudes affect intentions and intentions affect behavior. Attitudes are formed in a social context by such factors as the perceived level of evasion, the perceived fairness of the tax structure, its complexity and stability, how it is administered, the value attached to government activities, and the legitimacy of government. Government policies affecting any of these factors may influence taxpayer attitudes and hence the observed level of taxpayer compliance. Measures sometimes recommended for countries with very low compliance levels - such as the massive application of administrative penalties may be inappropriate and even have perverse effects in countries with higher compliance levels (Frey 2002) But some basic points seem universal.

Keep It Simple One important lesson suggested by experience is that a precondition for the reform of tax administration is often to simplify the tax system so that it can be applied effectively in a low-compliance context.52 In Bolivia, for example, much of the initial success achieved in reforming the tax administration in the 1980s was clearly attributable to the extensive simplifications made in the tax system. Indeed, as Bahl and Martinez-Vazquez (1992) argued with respect to Jamaica, it seldom makes sense to reform tax administration without simultaneously reforming tax structure to be both sensible and administrable. Considerable improvements can often be made in administration by small simplifications in tax policy. Reducing the number of income tax deductions, for instance, may permit countries to eliminate filing requirements for most wage earners. In addition to giving the administration simpler and more enforceable laws to administer, it is equally important to simplify procedures for taxpayers, for example by eliminating demands for superfluous information in tax returns and consolidating return and payment forms. Once procedures are simplified, the tax administration can then concentrate on its main tasks: facilitating compliance, monitoring compliance, and dealing with non-compliance.

The Taxpayer as Client

The prevalent attitude in many tax administrations appears to be that all taxpayers are potential criminals. The key administrative problem is thus seen to be identify and control taxpayers and to catch those who cheat. These tasks are indeed important, and this emphasis is understandable in a country undergoing rapid transition, but no modern tax system can function on fear alone. Problems of tax enforcement cannot be solved simply by calling in the ‘tax police.’ Extensive research in a number of countries shows that there is much to be gained from viewing tax payers more as clients — perhaps not very willing clients, but still clients — than as would-be criminals.

52 For a review of various country experiences, see e.g. Bird and Casanegra de Jantscher (1992) and Das Gupta and Mookherjee (1998).

40

Facilitating compliance, for instance, requires improving services to taxpayers by providing them clear instructions, understandable forms, and assistance and information as necessary. Monitoring compliance requires the establishment and maintenance of taxpayer current accounts, management information systems that cover both ultimate taxpayers and third-party agents (such as banks) involved in the tax system, and appropriate and prompt procedures to detect and follow up on non-filers and delayed payments. Improving compliance requires a judicious mix of both these approaches as well as additional measures to deter non-compliance such as establishing a reasonable risk of detection and the effective application of penalties. Successful reform strategies require an appropriate mix of all these approaches.

Of course, what tax administration can do depends largely upon the environment in

which it operates. That environment is often adverse in developing countries. Rewriting reality so that citizens are induced to comply with tax laws voluntarily is difficult in countries that face severe institutional limitations arising from large informal sectors, poor salary structures for public servants, ineffective and uncertain legal systems, and an entrenched distrust of government. Nonetheless, despite the many other tasks that tax administrations in developing countries must cope with, they must also pay much more attention to taxpayer services than is common. Studies on taxpayer behavior around the world suggest that services to taxpayers that facilitate reporting, filing and paying taxes, or that impart education or information among citizens about their obligations under the tax laws, are often as or more cost-effective in securing compliance than measures (auditing, penalties) more directly designed to counter non-compliance. In some countries it is difficult for even the most educated and honest taxpayer to comply with tax obligations. The taxpayer service perspective emphasizes reducing taxpayer uncertainty by clarifying legal ambiguities (often about issues as simple as the tax rate applicable to certain transactions), communicating clearly what the law is, and ceasing to change it so often that no one quite knows what it is. Taxpayer compliance costs should also be taken into account in designing legal and administrative procedures: for example, some countries require excessive numbers of tax payments to be made in inconvenient places. Much evidence demonstrates that improving tax compliance is not the same as discouraging noncompliance (Slemrod 1992). Compliance in most countries most of the time is at best "quasi-voluntary" (Levi 1988) because taxpayers have little choice, as when taxes are withheld. However, there seem to be two distinct groups: those who comply and those who do not. Some comply because they do not have the opportunity to evade; others because they are exceedingly risk-averse; and still others because they think it is the right thing to do -- and, importantly, they think other right-thinking people are also complying. Taxpayers do not view the decision to pay or not as a simple gamble. Some always pay; some always cheat; and some cheat when they think they can get away with it. An important task of tax administration is to prevent the mix from tipping in the direction of pervasive non-compliance. Low compliance may to some extent be a function of high compliance costs, as well as of such more basic problems as lack of state legitimacy, inadequate connection between taxes and benefits, and perceptions of tax fairness. Exactly why taxpayers dodge taxes is the subject of a large theoretical literature on the economics of tax evasion (Slemrod and Yitzhaki 2002). Over the years

41

some progress has been made both in incorporating the strategic aspects of the evasion decision in a game-theoretic framework and in modelling it in principal-agent terms, but much remains to be done before such analysis have much to say about the real world tax game in developing countries. For example, most literature on tax evasion assumes that tax officials are honest. If not all officials are honest (and in the expected utility framework it is not clear why they should be expected to be), the game is very different than that usually modelled.53 Similarly, the literature has not as yet managed to model very well either the long-term, repetitive nature of the tax game or the role of norms in determining how people play the game. Consideration of the temporal dimension of tax administration emphasizes the importance both of the interaction of officials and taxpayers and of changes in tax technology and taxpayer attitudes to government.54 Consideration of the social aspect suggests that although non-compliers may be similar in some respects everywhere both the size and the nature of the factors inducing compliers to comply may be quite different in different countries (Alm and Martinez-Vazquez 2003). Optimal enforcement strategy anywhere is likely to include both rewards (support) for compliers and penalties for non-compliers, but the correct mix depends on many context-specific factors such as the value attached to ‘fairness’ (and its meaning), the degree of deference to authority (and the legitimacy attached to that authority), and the extent to which contributing to the finance of government activities is seen to be socially (as opposed to privately, as in the economic model of tax evasion, discussed below) desirable. Increased enforcement actions (or lessened enforcment such as an amnesty) may have quite different results on compliers than on non-compliers, as may increased efforts at public education about taxpayer rights and obligations or increased efforts by tax authorities to provide improved service to taxpayers.

Salvation through Reorganization? The scarcity of tax administration resources is a constant in developing countries. Despite the high potential pay-off in terms of increased revenue, it is usually difficult, and often impossible, for tax departments to obtain and retain qualified staff or even to meet such basic material needs as office space and computers. Tax administrators are civil servants and hence subject to all the constraints affecting civil services. Reform strategies that require substantial additional administrative resources - particularly staff - are hence usually doomed to failure, because the needed resources will not materialize fully or in a timely fashion.

One way around this problem is to set up independent revenue authorities that are to some extent freed from civil service restrictions on hiring and pay and may also be given access to some earmarked source of revenue. Experience with this approach has been mixed. In some instances (e.g. Peru) at first matters seemed to go well, but then they deteriorated quickly (Estela 2001). In others (e.g. Tanzania) it was not clear that much changed for the better (Fjeldstad 2002). In still other instances, however, considerable improvements do seem to have occurred (Taliercio 2004).

53 "Leakage costs," (Shaw 1981) -- tax revenues that flow into the pockets of officials rather than into the coffers of government -- may simply be transfers in economic terms, but they may nonetheless result in significant distortions as new taxes are invented and tax rates increased in an attempt to make up the revenue loss. 54See, for example, the historical discussion in Webber and Wildavsky (1986).

42

While the jury is still out, a tentative conclusion might perhaps be that any country that has the will, strategy, and resources to reform tax administration probably does not need an independent revenue authority -- and a country in which these critical ingredients are lacking is unlikely to be successful even if it creates such an authority.

Even when there is a independent authority, there is seldom much, if any, additional

funding. As a rule, successful administrative reform strategies – with or without revenue authorities -- have been based on better allocation of available resources rather than on accretions of major additional resources. One example is simply cutting down unproductive tasks, as was done in Chile and Colombia, which ceased to process most returns of wage earners and devoted the resources thus freed to more productive auditing. Another example of internal reorganization that has been considered successful in some countries (for instance, Uruguay) is the creation of special offices to deal with large taxpayers (Baer et al. 2002).

Broadly, there are three ways in which one can run a tax administration, whether it is

‘independent’, ‘semi-autonomous’ or a simple line agency. First, establish a set of rules and apply them in the same way to everybody. Second, establish special rules for some and apply other rules to others. Third, establish general rules that are applied initially only to some but with the clear idea and obligation of extending those rules subsequently to all. The real choice is usually between the second and third approaches. In this light - as a ‘pilot’ for the extension of similar procedures as and when it becomes feasible to do so - special tax offices for large taxpayers may make sense sometimes as a way to begin to reform tax administration. On the other hand, if the sole aim of the change is to maximize revenues, the result may be deleterious both because other essential administrative tasks may be unduly neglected and because in effect an extra tax ‘penalty’ (tighter control and enforcement) is put on more successful firms. Another approach in some countries has been to privatize certain tax administration activities traditionally performed by government. Countries such as Bolivia, Brazil, Chile, Colombia, and Ecuador, for example, have out-sourced much tax collection to banks, in part because of insufficient resources in the tax administration and in part because banks are already specialized in the handling and control of payments and it makes life easier to get the cash out of the hands of tax collectors. Of course, simply entrusting banks with the task of receiving payments or returns (and even, in some countries, processing returns) does not assure success. Proper systems must be designed, the tax department must exercise adequate supervision and the remuneration paid to the banks must be appropriate. Much time and effort has been spent on these matters in countries in which collection through the banking system operates successfully.55

55 Another important organizational decision relates to the degree of decentralization: for recent reviews, see Martinez-Vazquez and Timofeev (2005) and Mikesell (2007). Yet another is the ‘tax police’ approach popular in some eastern European countries, which does not seem likely to be either desirable or sustainable from the perspective of creating a sustainable and accountable state. Of course, as IDB (2006) stresses, there is always some trade-off between policy effectiveness and the degree of inclusiveness in the policy process, and countries obviously choose to settle at quite different points on this spectrum.

43

The IT Solution

Most recent attempts to reform tax administration center on information technology (IT), No modern tax administration can perform its tasks efficiently without using IT, but in many developing countries the expectation of greater effectiveness from computerization has not materialized. Those reform efforts that are successful did not simply computerize antiquated processes but re-engineered the whole system - for example, consolidating return and payment forms, eliminating unnecessary and unused information required from taxpayers, and so on. Successful computerization requires a fundamental reorganization in both systems and procedures; it cannot be used to side-step such reforms. Moreover, even the best computerized system will not produce useful results unless there are real incentives for tax administrators to utilize the system properly.

Radical improvement in tax administration does, however, require a transformation of its organization and methods, and modern IT greatly facilitates the needed transformation (Engelschalk 2000). A recent study on the enforcement efficiency of the income tax department in India, for example, identified the following problems: poor utilization of information collected by the central intelligence branch; ineffectiveness of surveys of business premises; absence of an adequate system of taxpayer identification numbers; absence of an adequate system of third party information collection; and the poor state of records and deficiencies in the record-keeping system. Much the same could be said in other countries, and in most cases IT is inevitably a critical element in resolving such problems.

The availability, cost, and accessibility of modern computers make them ideal for the large-scale information processing and coordination problems facing tax administrations in even the poorest countries. Among the areas that may be computerized are: (1) taxpayer records and tax collection (taxpayer compliance); (2) internal management and control over resources; (3) legal structure and procedures; and (4) systems to lower taxpayer compliance costs.56 The first of these areas lies at the centre of any computerization exercise and the single most important element within this area is basic information on taxpayers – for example, a taxpayer master file. Another reason why tax administrations need IT expertise is simply because some of their most important clients -- multinational companies and, increasingly, large domestic firms -- employ sophisticated computer systems which are beyond the investigative capacity of technologically backward tax administrations. IT is a ‘double-edged’ tool: in the hands of taxpayers, it may make tax administration more difficult (especially in an open economy); but in the hands of the administration it may enable it to respond more robustly to such challenges (Bird 2005).

Experience suggests a number of lessons with respect to the successful application of information technology in tax administration:

• First, and most important, an appropriate strategy must be developed that takes into account the obstacles and the constraints arising from such organizational rigidities as civil service salary structure and procedural hurdles in acquiring the necessary expertise,

56 A model of what can be done in most of these respects is Singapore, which has likely gone as far in using IT to modernize and improve its tax system as any country in the world (Jenkins 1996; Oldman and Bird 2000).

44

hardware and software. The susceptibilities of the existing staff and their resistance to change need to be taken into account. Experience in Kenya and elsewhere demonstrates that new technologies can only be introduced successfully if the key players in an administration are brought on side. Since a complex system is more likely to engender resistance and problems, the design, structure and operations of the system should be simple. In some situations (as in Mexico) it may be advantageous to entrust a part of the responsibility for setting up an information system to organizations outside the tax administration or even the government.

• Second, considerable organizational re-engineering is usually needed to gear the tax administration to a computerized environment. Sometimes (as for property taxes in Indonesia) it may be advantageous to reorganize tax administration by sector. As a rule, however, a functional approach is easier to operate when key information regarding a taxpayer’s obligations (like filing of returns and payments) is stored in the computer, with a tax ‘vector’ created for each taxpayer (as in Spain).

• Third, equipment and software should be standardized to facilitate operation, networking, and maintenance. Experience suggests that, whenever possible, software should be bought ‘off the shelf’ rather than developed internally, both for cost reasons and to accommodate subsequent technological developments.

• Fourth, the pace of change and the success of any modernization program depend ultimately on human resources - on the training and skills of the people who are expected to use and operate the technology. But technical expertise alone is not enough to assure success in application: appropriate incentives and accountability are also needed. Often neither the requisite expertise nor the desired incentives are easy to put into place in the prevailing government environment.

• Finally, IT can do little for tax administration unless a unique identification number is allocated to each taxpayer. In every country in which computerized tax administration has been successful, allotting a unique identification number has been one of the key steps. Without such a number, information can neither be stored properly nor used for any purpose.

An Example: VAT as a Self-Assessed Tax

To lend a little verisimilitude to what might otherwise seem the set of relatively bald and perhaps unconvincing assertions offered in this section about what must, should and can be done to improve tax administration in developing countries, consider the big item in most revenue structures these days: the VAT. The standard IMF treatise on the ‘modern VAT’(Ebrill et al. 2001) argues persuasively that the core feature of a modern VAT is that it is a ‘self-assessed’ tax. This argument is inarguably correct: a good broad-based tax really has to be one that essentially ‘runs itself’, that is, one that, once established, basically becomes part of standard business operations. The proper role of the tax administration with such taxes is not to assess who owes what but instead, so to speak, to guard the ‘borders’ of the system and to verify that those who should be ‘self-assessing’ are behaving as they are supposed to do.

45

Ebrill et al. (2001) set out seven ‘pre-conditions’ that must be in place for such a self-assessed tax to function properly, as set out in the left-hand column of Figure 6. In the right-hand column is a list what conditions might obtain in a country that completely failed the corresponding pre-condition test. Finally, and perhaps somewhat provocatively, I have filled in the cells in the table with a totally subjective appraisal (on a arbitrary scale ranging from 1 – very low -- to 5 – very high) assessing the extent to these pre-conditions seem to be satisfied in the countries listed, in all of which I have some recent personal experience.

Figure 6. Prospects for VAT Success: A Subjective Appraisal

Pre-

conditions

Canada South

Africa

Jamaica Colombia Egypt Ukraine Liberia The

opposite

Simple,

clear, stable

law

***** ***** *** ** * * * Bad,

changing,

law

Good

taxpayer

service

***** **** ** ** * * * No service

Simple

procedures

***** **** *** ** * * * Obscure

and

complex

Effective

enforcement

***** *** ** ** * * * No real

enforcement

Reasonable

audit

***** *** ** ** * * * No audit

Strict

penalties

***** *** ** ** * * * No effective

penalties

Good review ***** *** * * * * * No review

What Figure 6 suggests is that (1) a modern VAT should work well in Canada and that (2) a ‘standard’ VAT model should also work, although not always that well, in Jamaica and Colombia, but that (3) for very different reasons not spelled out in the table this model is much less likely to prove successful in Egypt, Ukraine and Liberia – the last of which does not have a VAT at present. Although I know less about South Africa than the other countries, it combines a well-developed modern sector – to which VAT mainly applies – and a large ‘small business’ sector where the impact of the country’s well-structured VAT is almost certainly quite different. To sum up, what Figure 6 suggests is that many countries that already have VATs fall far short of the ‘ideal’ conditions for a ‘modern’ ‘self-assessed’ tax.57

The conditions set out in Figure 6 relate only to the structure and functioning of the tax

administration and not to the underlying economic and political structure of the country within which that administration has to operate. But one cannot take a tax administration out of its

57 Of course the picture sketched in Figure 6 is based not only on my (inevitably only partially informed) assessment about the extent to which the various pre-conditions are met but also on my assessment (based on varying degrees of involvement and knowledge) of how well I think VAT actually does work in these countries. Informed observers may quibble and downright disagree with some of my entries, but I doubt if anyone familiar with how VAT really works in most countries would seriously dispute my main conclusion: the ‘best’ VAT for all countries is not necessarily identical.

46

environment and similar conclusions emerge from any systematic comparison of what might be called the ‘fiscal architecture’ (Wallace 2003) that underlies any tax system. . If one takes into consideration not only the ‘quality’of the tax administration (as in Figure 6) but also the size and location of the country, the level and distribution of income and wealth, the importance and pattern of international trade, the relative size and nature of the informal sector, the relative importance of ‘final’ and intermediate sales, as well as more nebulous but nonetheless important factors such as the prevailing level of trust in government and taxpayer morale two additional points become clear:

• First, the level, structure, and effectiveness of taxation – including VAT – varies enormously from country to country and over time within countries as these underlying factors change (Bird, Martinez-Vazquez and Torgler 2006).

• Second, the nature and effectiveness of any tax administration varies with both these factors and policy choices, as will both the extent to which conditions such as those set out in Figure 6 are satisfied and what satisfying such conditions means in terms of providing a foundation for a modern self-assessed tax like VAT (Vázquez-Caro 2005).

To design and implement successfully a modern tax like VAT, one must, as Ebrill et al (2001) say, satisfy conditions such as those set out in Figure 6 -- that is, essentially have both the right tax structure and the right tax administration. But one also needs the right clientele (taxpayers) as well as the right tax administration to make a self-assessed VAT work as it should. Before many developing countries can get to this happy starting position, however, they may need some very fundamental changes in underlying economic and political conditions.

None of this discussion means, of course, that only a developed country should ever have

a VAT. On the contrary, as Bird and Gendron (2006) argue, VAT is almost certainly the right way to impose general consumption taxes in almost all circumstances. The basic argument is simple: (1) almost every country needs a general consumption tax and (2) VAT is the best form of such a tax – the one with the fewest bad effects. The case for VAT remains strong everywhere. Even in the least developed of the countries mentioned in Figure 6 (Liberia), for example, the existing limited general sales tax imposed on manufacturing and imports should be moved much closer to a normal VAT by replacing the present ‘ring’ system of exempting imports by registered manufacturers by a VAT credit system under which VAT is applied to all taxable imports and this import VAT may then be credited (by registered firms) against output VAT due. Such a change makes both administrative and economic sense. Administratively, the (very weak) tax office would no longer have to prove that a duty-free import had been improperly claimed. Instead, taxpayers would have to demonstrate, by filing returns, that they were entitled to offset VAT on imports against VAT on sales. Evidence in many countries in Africa suggests that a high proportion of potentially creditable sales tax levied on imports is not in fact utilized as a deduction against tax due on sales in large part because the supplies are being diverted to the ‘informal sector’ (Glenday and Hollirake 2006).58 To the extent that the informal sector is being ‘fed’ by the diversion of tax-free imports, stopping this leakage would constitute a major step towards restoring the competitive balance between formal and informal sectors.

58 Keen (2006) sets out the analytical underpinnings of this argument in detail.

47

Even a very poor country like Liberia with a very weak administration may thus, by imposing at least a limited VAT, both have more cake (revenue) and at the same time stimulate a sounder pattern of development. What more can one ask? Tax administration can be and often is a serious constraint on tax reform, and countries should be both encouraged and helped to improve their administrations as much as reasonably possible, but significant tax improvements are often possible even with weak administrations.

6. The Politics of Taxation

Changes in such fundamental economic factors as a country’s openness to trade and the income elasticity of expenditures subject to VAT obviously affect tax outcomes. More basically, however, the level and the structure of taxation in any country reflect deep-seated institutional factors that, in the absence of severe shocks, do not change quickly. Tax policy decisions are not made in a vacuum. Nor are they made by a benevolent government. Instead, they reflect the outcome of a set of complex social and political interactions between different groups in society in an institutional context established by history and state administrative capacity. Taxation is not just a means of financing government; it is also a very visible component of the social contract underlying the state. Citizens are more likely to comply with tax laws if they accept the state as legitimate and credible and are thus to some extent both willing to support it and afraid of what will happen to them if they don’t. Improving tax outcomes thus depends in large part upon how different political groups perceive proposed changes and how they react to these perceptions. As Lledo, Schneider, and Moore (2003) put it, any major tax reform is thus always and everywhere“an exercise in political legitimation.” Those who will have to pay more must be convinced that they will get something worthwhile for their money. Those who will not pay more must not be able to block reform and, in the end, be willing to go along. Those who will have to implement the reform must also support it or at least not actively oppose it. And of course politicians have to see sufficient support to warrant putting reform not only on the agenda but on the ground. Accountability and Visibility

As an example of the importance of getting the politics of taxation right, consider two apparently minor questions about VAT design. Should a VAT be quoted separately from prices? And should it be called a VAT? Each of these questions delves deeply into the institutional cauldron. Neither has a simple or unique answer. Both need more consideration than they have received in most countries.

Tax Awareness: Good or Bad?

Retail sales taxes like those in most U.S. states are invariably stated as a separate explicit

charge imposed on the posted price when the consumer arrives at the cash register. While this process is cumbersome and unwelcome – no one ever has the correct change ready! –the very

48

fact that it is annoying may be considered good for democracy if one believes that citizens should be fully aware of the cost of government. However, such transparency also makes it more difficult to increase taxes because everyone is instantly aware of increases. In contrast, in most countries (with the notable exception of Canada) VAT is included in posted prices and hence is not immediately obvious to the consumer at the point of sale (although it may be stated explicitly on the cash register statement or invoice). Visible, or invisible: does it matter?

While there seem to have been no serious studies of this issue,59 it is intuitively plausible

that it should be easier to introduce (or increase) a tax if people are not painfully reminded that the tax exists by having to add it separately to the quoted price every time they buy something.60 Moreover, if VAT is relatively invisible it may be easier to have a better, broader tax base, for instance, by taxing a wider range of services. A broader base is undoubtedly more desirable on both administrative and economic grounds than one eroded by popular resistance to taxes on food, on medicines, on school books and so on. On the other hand, if people should know what they are paying for what they are getting from government, then even if the price of more transparency is a less perfect VAT, the price might be considered worth paying.61 Given these conflicting arguments, it is striking that almost every country with a VAT has chosen to hide the tax from the public. This outcome is of course understandable from the standpoint of those who make the decisions. But hiding the tax bill does not seem, on its face, a particularly attractive way to build a sustainable democratic consensus in support of fiscal equilibrium.

What’s in a Name?

A tax by any other name is still a tax, so does it matter what a tax is called? Many

governments around the world think that it does, as indicated by the many taxes with names like the Education Tax, the Employer Health Tax, the Hospital Tax, the Security Levy and so on. Sometimes the revenues from such taxes are in fact earmarked to the indicated objective. Often, however, this is not done: the name is the game. (Earmarking is discussed further later in this section.) Perception matters in politics, and labels affect perception. A good name may do as much to help a tax as a bad one can do to kill it.62

To illustrate, is a VAT that is not called VAT more or less likely to be accepted? Many

countries now call their VAT a GST – an acronym that has the useful dual meaning of Goods and Services Tax or General Sales Tax. Might a country wishing to introduce a VAT (or to raise

59 For two rather primitive looks at the question, see Tanzi (1970) and Bird (1982). 60 One reason some conservatives in the U.S. have resisted VAT and argued for replacing the income tax by a separately-quoted national RST is that they think a ‘hidden’ VAT would lead to a larger government – a belief the evidence for developed countries in Keen and Lockwood (2006) supports to some extent – while on the other hand they think that replacing the (largely withheld) income tax by a more visible RST would result in smaller government. 61 Even thorough-going democrats might prefer some taxes to be hidden for the same reason people sometimes support forced saving schemes of various types: namely, as a way of forcing themselves to behave more rationally in the long run (Elster 1984). Or they may prefer such taxes for paternalistic reasons similar to those of some supporters of social security systems: namely, to force others to behave more rationally (and thus ultimately in their own best interests) in the long run (Musgrave 1981). 62 Graetz and Shapiro (2005), for instance, stress that the label ‘death tax’ played a fatal role in the repeal of the U.S. estate tax.

49

the rate of an existing VAT) benefit by calling the new (or increased) levy something like, say, the Deficit Reduction Levy (though that may imply an undesired temporal limitation) or, say, the ‘No Child Left Behind Tax’ in support of increased education? Social security systems have done well by linking payroll ‘contributions’ – to a fiscal economist, a wage tax by another name – to expenditures that people desire. Can or should developing countries seek to follow this path, at least to some extent?

Such issues need to be considered more carefully in formulating tax policy than is usually

done. Marketing matters. Those who want serious reform in tax policy or administration in any country need to study the dark art of political salesmanship as well as analyzing the details or data of tax theory and practice. This warning needs to be taken particularly seriously by those are concerned not only with how tax revenues may build up a more sustainable state but also with how the way in which such revenues are collected may contribute to, or detract from, the long-term development of state legitmacy. Name changes may be an illusory game but when they produce real political effects, as fiscal illusions seem often to do, then the subject is worth closer attention than it seems to have received to date.

Earmarking and the Wicksellian Connection

More substantively, attention should also be paid to revenue sources such as earmarked taxes, benefit taxes, and user charges that explicitly emphasize the ‘wicksellian connection’ (Breton 1996) between the two sides of the budget. This linkage is critical to the whole issue of state legitimation: it is, for example, one of the key elements in explaining the popularity (and mixed results) of fiscal decentralization around the world.63 Earmarking – linking specific revenue sources to specific expenditures -- has existed since the earliest recorded fiscal practices (Webber and Wildavsky 1986). Both politicians and taxpayers often find earmarking an attractive and feasible way to finance social security, road works, education, environmental programs, and other good things. Politicians like earmarking as a means of reducing taxpayer resistance to higher taxes. Taxpayers like the greater accountability they perceive with respect to how their tax dollars are spent. Economists, however, have come only lately to the table when it comes to understanding and analyzing this common fiscal practice.64 Those concerned with budgeting almost unanimously conclude that earmarking is a bad thing, essentially because – as an unfriendly critic might say – they appear to believe that the alternative is for budgetary decisions to be made by a ‘benevolent dictator’ whose sole objective is to maximize social welfare. More charitably, budgetary experts are aware of the mess that rampant earmarking has created in some countries and argue that no rational budgetary process is conceivable unless the practice is essentially banned.65 This approach remained essentially

63 See later discussion: the theme that decentralization may be thought of a form of ‘earmarking’ is developed in, for example, Bird (2001). 64 For a more detailed discussion, see Bird and Jun (2005). 65 Earmarking (hypothecation) was widespread in Britain at the turn of the 19th century but was then rejected and replaced in mid-century by the ‘Gladstonian’ approach to public finance, an important feature of which was a consolidated budget with no earmarking (Daunton 2001). Interestingly, the stated reason for the turn away from earmarking was to restrict the growth the state in order to restore public trust in the neutrality of the public finances in the face of the then-common perception that hypothecated revenues were being (mis)used by the political elite to

50

unchallenged until Buchanan (1963) revived an important efficiency argument (made earlier by Wicksell) in favour of establishing as tight a linkage as possible between taxing and spending decisions. In this approach, earmarking is not simply a way to secure political consent for a tax increase but rather in principle the best way we have to deal with the fundamental normative problem of public economics -- how to provide people with the public services they really want - where ‘want’ is interpreted in the only economically relevant sense of what they (collectively, as determined through their political institutions) are willing to pay for. As with most fiscal institutions, however, in the real world earmarking comes in many variants of earmarking, as shown in Figure 7, and such arguments are much more persuasive with respect to some types of earmarking than for others.

Figure 7. Varieties of Earmarking

Variety Expenditure Linkage Rationale Example

A Specific Tight Benefit Public enterprise; user charges

B Specific Relatively loose

Benefit Gasoline tax and road finance

C Broad Tight in general

Benefit Social security

D Broad Loose ‘Benefit’ Tobacco tax and health finance

E Specific Relatively tight

None Environmental taxes and clean-up programs

F Specific Very loose None Payroll tax and health finance

G Broad Relatively tight

None Revenue sharing to localities

H Broad Very loose None Lottery revenues to health

Source: Bird and Jun (2005) The case for earmarking is strongest when there is a close benefit link between the payment of the earmarked tax and the use of the revenues to finance additional expenditures as in Type A and (depending on the details, perhaps Types B, C, and E) in Figure 7. Such benefit-related earmarking reveals taxpayer preferences for public services and sends a clear demand signal to the public sector about how much of a service should be supplied. Moreover, since the revenues received are spent on the service in question, supply automatically adjusts to demand and economic efficiency is achieved. Such earmarking may also be considered equitable in the sense that no one receives a service without paying for it or pays without receiving service. Providing the public service thus financed is like a privately supplied service in the sense that both an individual’s consumption of the service and the marginal cost of providing the service can be satisfactorily measured, most people would probably consider user charge or benefit-tax

expand the ‘fiscal-military’ state in their own interests. This case suggests that, as with many fiscal devices, whether earmarking increases public trust in government or reduces it depends very much upon the context.

51

payment fair in such cases, apart perhaps from general distributional concerns when it comes to people with low-incomes.66 Such earmarking establishes a different budgetary process from the alternative of general-fund financing prescribed by most budgetary experts, under which all government revenues flow into one big pot and all government expenditures are made out of this pot. In that system, taxation and expenditure decisions are made separately and no linkage exists between the revenues generated by any particular tax and the level of expenditure on any particular activity. With such general fund financing, if a government wishes to increase spending on any activity, to the extent that preferences are affected by how a particular program is financed citizens have no rational basis for knowing the true costs of any particular expenditure decision so that fully informed budgetary decisions cannot be made.67 In contrast, with strict earmarking the sequence of revenue and expenditure decisions is reversed: revenue collections drive expenditure levels. Since taxpayers are aware that when certain tax (or user charge) payments are extracted from them the funds will be used to pay for certain expenditures, they presumably support the charges if and only if they support the expansion in the supply of government services for which the earmarked revenues are targeted. Both tax and expenditure decisions may thus be made more rationally than under general fund financing. Of course, how well such systems work in the real world depends upon many things, and it is not surprising that in reality perfection seems seldom to have been achieved. The relatively disappointing experiences often observed in practice reflect such factors as the cost and difficulty of controlling many separate funds (Fullerton 1996), the inappropriateness of many of the linkages that have been established for political reasons between particular revenues and expenditures (McCleary 1991), and the understandable resistance of citizens to attempts to charge properly for services that initially were provided ‘free’ or at highly subsidized prices (Bird and Tsiopoulos 1997). Properly designed earmarking may also in certain circumstances be an effective way to enforce an inter-temporal agreement (Teja 1988). Suppose, for example, that an earmarked excise tax on chemical stocks is levied to finance the clean-up of toxic chemicals. Unless the proceeds of the tax are strictly earmarked for this purpose, taxpayers in ‘clean’ areas might resent having to pay for improving ‘dirty’ areas. Unless those living in the areas that benefit from clean-up are reciprocally bound in their turn to finance any future costs of cleaning up areas that may subsequently become contaminated, nothing may ever be cleaned up because no one is willing to pay for benefiting others without a firm commitment that they themselves will be looked after in the future, if necessary. If earmarking is seen as a way of insuring through such compelled reciprocity that funds will be available for clean-up no matter when such funds may be needed in the future, it may provide a acceptable political solution and overcome the incomplete contracting problem that would otherwise exist, in effect by increasing the political

66 Partial earmarking may be appropriate even if consumption of a particular public service generates external benefits for other households. In the limiting case in which the service is a pure public good and the marginal cost of extending service to another household is zero there is clearly no role for either user pricing or earmarking. 67 Bird (2005) examines in more detail the linkage between financing and expenditure evaluation in the context of benefit-cost analysis.

52

costs to future governments of breaching the contract (Cremer, Estache, and Seabright 1995).68 Such contractual (or quasi-contractual) arrangements may make most political and economic sense when there is some logical connection between the revenues and expenditures thus linked. Examples of earmarking that may to some extent be justified along these lines are not only some enviromental taxes and expenditure (Type E in Figure 7) but also various forms of road user charges (Type B), as in the case of ‘road funds’ (Gwilliam and Shalizi 1999). On the other hand, symbolic earmarking like Types F, G, and H (and, in reality, usually Type D also) in Figure 7 has no clear economic rationale. Nonetheless, it may sometimes make political sense, at least when it is first put in place. For example, if a government imposes a new payroll tax and calls it a Health Tax even though the funds flow into the general budget and have absolutely nothing to do with how much is spent on health, such labeling is clearly for political purposes only. A particularly egregious example is when, say, tobacco taxes or lottery revenues are earmarked to, say, education. Not only is there no logical link between smoking or gambling and education but there is also almost never any budgetary link either. How much is collected from the designated source and how much is spent on the designated activity are essentially separate decisions, decided independently. This may be just as well, of course, since the craving for nicotine or the lure of chance has absolutely no connection to the need for public education finance. Such linkages simply make no sense, and their only possible rationale seems to be for revenue enhancement e.g. by capitalizing on the presumed ‘halo effect’ of some popular expenditure to justify the imposition of additional taxes (Rajkumar 2004), as suggested earlier in the discussion of the potential importance of tax names. Apart from its role as misleading advertising, earmarking may perhaps also be motivated in some instances by rent-seeking behavior and be promoted by those who expect to benefit by securing more reliable funding for their favored expenditure. To illustrate, a decreasing number of North American adults smoke and most of them probably feel increasingly guilty about doing so. An increase in tobacco taxes with the proceeds earmarked for increased health spending would thus likely receive support on all sides: from non-smokers, who would not pay the tax; but want more spending on health; from smokers who feel guilty about smoking and are worried about the health consequences of smoking; and also, of course, from the very large number of people who are engaged in the health business and who (one assumes) not only really believe that higher taxes on tobacco and more spending on health are both positive things but also realize that they would be clear gainers because health spending would increase as a result – even though this need not necessarily happen. Even though some studies (Cnossen and Smart, 2005) suggest that tobacco taxes are both highly regressive and already, in many countries, higher than can be rationalized on any externality arguments, how can any politician be expected to resist such a win-win combination?69

68 For a related approach, stressing that in some instances earmarking may both provide voters a way to pin down politicians about whom they are uncertain and politicians a way to ‘signal’ their concerns to voters in a credible way, see Brett and Keen (2000). 69 Earmarking tobacco taxes to health programs can be rationalized to some extent on ‘benefit’ grounds but doing so seems unlikely to have any significant effect on health spending in most countries. Earmarking such taxes to e.g. ‘anti-tobacco’ advertising campaigns may significantly increase the flow of funds to such activities (Jha and Chaloupka 2000, chap. 10) but this does not mean it is the best use of such funds.

53

Moreover, what is earmarked for one use today may be diverted to an alternative use tomorrow as competing rent seekers continue to struggle for budgetary gain. If, for example, a tax on motor fuel is earmarked to pay for highways, groups interested in raising safety standards and reducing speeding may soon be successful in including highway patrol costs in the expenditures made for highways. Other groups may strive to include, say, rail passenger transport or urban mass transit within the favored expenditure category. As many (Bird 1976) have argued, this particular extension may make sense in many instances; but the point here is simply that what is ‘contracted’ with a government today and what is delivered by the (same or different) government tomorrow may differ substantially. In politics, where nothing is forever, no contract is ever ‘complete’. Nonetheless, even when people feel overtaxed and resist attempts to raise general taxation, new earmarking initiatives may be welcomed. Governments may both wish to impose their decisions on their successors through earmarking and also to enhance their revenues by creating the illusion that the funds collected will lead to greater expenditures on popular programs. A very few citizens may perhaps see earmarking as the rational choice mechanism it can be under certain conditions and welcome the opportunity to exert greater control over how their tax dollars are spent. Others may perceive an opportunity to engage the public sector in redistributive activities favorable to themselves. Still others may support earmarking as a way of protecting some activity that they support from future political whims (although the logical case for thus facilitating the imposition of the ‘dead hand’ of past political compromise on future political decisions is by no means obvious). And of course fiscal illusion may influence others who think the new taxes support something they like but do not realize that there is no guarantee that their favored activity will actually benefit. In the end, much as earmarking seems to appeal to politicians in countries from Korea to Brazil as an easy source of new revenues, if it is really nothing but fiscal sleight-of-hand. If one believes that in the long run one cannot fool all of the people all of the time, it seem unlikely that higher revenue levels than would otherwise be achieved can long be achieved through such means.70 Except when there is a clear benefit rationale, it is hard to defend any form of earmarking that actually affects expenditures (in the sense that the level of expenditure on any particular activity is determined at the margin by the amount of revenue collected from any particular tax or charge). But where sensible ‘user fees’ can be levied, they generally should: unfortunately, the fees, charges, and public prices found in most countries are seldom sensible.

Getting Prices Right

To some extent government may be viewed in effect as a firm delivering a package of services. As with any economic activity, those who get the benefits should want them enough to

70 A possible exception that deserves further exploration might be when the central tax administration is, for reasons that cannot be readily corrected, weak while separate ‘earmarked’ tax administrations, which in effect can keep what they collect, can do a better job – a situation that one might argue may exist in some countries with respect to social security taxes, for example. This may be one rationale for the separate administration of payroll taxes found in some countries: of course even if the result is to expand the size of the public sector it is not clear that such expansion is desirable.

54

be willing to pay for it. User charges are usually seen as simply an additional source of revenue, but their main economic value is to ensure that what the public sector supplies is valued at least at (marginal) cost by citizens and to promote economic efficiency by providing demand information to public sector suppliers. Whenever possible, public services should therefore be charged for properly rather than given away.71

User charges are more important an alternative to general taxes than is indicated by the relatively small amounts of money that most countries collect from this source. The appropriate policy in setting user charges is simply to charge the correct (roughly, marginal cost) price. Only thus will the correct amounts and types of service be provided to the right people, that is, those willing to pay for them, so only when such charges are levied will scarce public resources be as efficiently utilized as possible. Many argue that equity considerations mean user charges are a particularly bad idea in developing countries. In principle, the incidence of user charges – the price of services obtained by someone from the public sector -- is no more relevant than the ultimate incidence of the price of something obtained by someone from the private sector: education and health are no different than food or shelter in this regard. Moreover, the distributive consequences of charging for local public services may even be progressive in the circumstances of developing countries (Bird and Miller 1989a). For example, since the rich use much more water (for washing cars, watering lawns, and swimming pools) than the poor, under-pricing water is regressive, especially when, as is often the case, the result is to starve the water agency of funds and hinder expanding access to the poor. Trying to rectify fundamental distributional problems through inefficiently pricing scarce local resources is generally a bad idea, resulting in little, if any, improvement in equity and a high cost in terms of lost efficiency (that is, wasted resources).

Much as there is to be said in principle for the rational use of user charges in developing countries, most of the many prices fees, charges, and minor levies now existing could likely be abolished with little loss in revenue and some gain in efficiency. The few charges with any real economic rationale also generally need substantial revision if they are to be conducive to efficient resource use. User charges are seldom employed to the extent that is both possible and desirable, and that those charges that do exist are seldom well designed and consequently seldom produce any significant economic benefits. Unfortunately, even in the best circumstances, it is often surprisingly difficult to design and implement good user charges, and as a rule even good charges are not very popular with either administrators or citizens. Nonetheless, more can and should be done to improve revenue systems in this respect. There is too much to be gained in terms of efficiency, if not always in revenue, not to try.

Decentralization as a Partial Solution

Developing countries need to keep such ideas especially in mind when it comes to decentralization, a process that many are now undergoing to varying degrees. For example, in a number of countries (notably in Latin America) too large a share of national revenues has been earmarked to subnational governments, sometimes ruining both national budgeting and any

71 This efficiency objective is particularly important at the subnational level since the main economic rationale for subnational government in the first place is to improve efficiency (Bird 2001).

55

vestige of a ‘hard budget constraint’ at the local level (Rodden, Eskeland, and Litvack 2003). Or else the money flows down only with a heavily constrained set of ‘earmarking’ conditions attached as to how and on what it may be spent, thus ruining any chance at sensible local budgeting as well.

Reforming tax structures is generally difficult in any country. Reforming tax

administrations is often even more difficult, and certainly takes longer. But even countries that succeed in reforming tax structures and administrations may fail to reap the expected benefits unless they pay close attention to the often troubling problems arising from the finance of state and local governments. Such problems are by no means restricted to the relatively few (though important) developing federal countries such as Brazil, India, and Nigeria; they are also occurring around the world in countries as diverse as China and Colombia.72

Decentralization takes place for many different reasons and with many different results.

Some countries anticipate that better service delivery will result because the diverse demands and needs of the population can be served more effectively by local officials who have better information on what people want. There may be diseconomies of delivering some services at the national or even regional level, so local service delivery can be less expensive. Decentralization may be seen as a means of accommodating the varying interests of different ethnic groups by reducing the potential number of points of friction. Whatever the rationale, a common problem around the world is that decentralized governments seldom have the capacity to generate sufficient revenues to meet their expenditure needs. Many such governments do not have an adequate resource base. Most of those that do are prohibited from tapping that base by higher authorities. And even the few that could do so seldom do so in fact because they, like the rest of us, find it easier and more attractive to seek and spend ‘other people’s money’( through borrowing and transfers) than to face their own citizens with a bill for services rendered.

Discussion of such issues is often clouded by confusion as to how decentralized tax

revenues are, or should be. There are at least four dimensions of potential or actual local (or regional) control over revenues: ownership of tax revenue, choice of tax base, choice of tax rate, and tax administration. Revenue ‘ownership’ alone is not enough to consider taxes as decentralized. Even if local governments ‘own’ the revenue stream in some legal sense, it is still a transfer if, as is common, who gets how much is determined at higher levels. Transfers from higher-level governments may make sense to close the gap between what reasonable tax rates can raise from a region’s tax base and minimal acceptable expenditure levels as well as to further national interests in the provision of services with inter-jurisdictional spillovers (Bird and Smart 2002). But local governments can usually accountable for such revenues only ‘upwards’ to the donor level.73 They are not accountable politically to their own citizens in the same way they are for revenues that they raise directly from those who also receive local public services and can make the link between what they pay and what they get.

72 For discussions of the fiscal aspects of decentralization in the latter two countries see Wong and Bird (2005) and Acosta and Bird (2005). For a general overview, see Bird (2001). 73 The lack of effective public auditing in most developing countries makes even this accountability a charade: but see Ferraz and Finan (2007) for an interesting study of Brazil.

56

From the perspective of both economic and political accountability, the most essential element of fiscal decentralization is therefore to let local and regional governments determine the rates of taxes for which they are politically responsible – taxes that, ideally, should be sufficiently important in revenue terms to noticeably affect expenditures. Control over the tax rate both gives local officials the ability to influence the level of services provided. However, few developing countries have been willing to ‘trust’ local governments with either the authority or, more importantly, the responsibility to make their own tax decisions (and mistakes), despite the growing evidence of the importance of this link in developing both effective government and economic development (Sokoloff and Zolt 2005; Hoffman and Gibson 2004).74

7. What Should Be Done?

Reforming taxes in any country is always a one-off operation in the sense that it occurs in the unique circumstances of that place at that time.75 Nonetheless, some seem to believe that there must be some simple solution to be found somewhere else in the world that can replace the seemingly unending problems and process of negotiation found in their own country. Much can be learned from studying how different countries have coped with tax reform: the solutions reached may be different but the basic problems that must be faced are often rather similar. How one country dealt with a problem may provide useful hints on how another may do so.76 But comparative analysis can never provide a complete answer. Institutionalize Tax Reform

Most studies of tax reform focus on the substance of what should be considered. A more

fundamental question, however, is not what should be taken into account in developing a tax reform proposal but rather how the issues, whatever they are, should be approached (IDB 2006). McIntyre and Oldman (1975), for example, argue that more careful and comprehensive attention to institutional arrangements for tax reform will both improve the quality of the reforms proposed and increase the likelihood of their adoption and successful implementation. They give four reasons:

1. Better planning will lead to better “packaging” designed both to attract political support

and to undermine political opposition. 2. Changes in reform proposals for political (or other) reasons can be made more quickly,

while maintaining the basic intent of the reform.

74 Addison and Levin (2006) seem unduly worried about ‘bad’ local taxes, but local control over taxes does of course carry with it various possible costs and consequences although this issue cannot be further discussed here. 75 Much of what follows relates only to ‘major’ tax reforms (Bird 2004). Many countries constantly “reform” their tax systems by altering rates, redefining bases, and adding and clarifying interpretations to existing law, and it is not always a simple matter to tell when such ‘technical changes’ constitute a major reform but this issue is not discussed further here. 76 Bird and Slack (2006) apply this approach to the case of property tax reform, for example.

57

3. Incorrect but politically appealing arguments against reform can be refuted if proposals

are backed up by careful studies.77 4. Politicians introducing reforms will have more control of the process in terms of timing

and presentation Even the best provisions for studying and developing reform proposals will never be enough to bring about good policy changes in the absence of a coherent strategy, continuing support from above, and an acceptable level of administration.78 Good policy formulation is never enough; but presumably it is always better than its absence or its opposite.

The specific institutional arrangements appropriate for any country must of course be

developed specifically for the conditions of that country but a number of key common elements may be identified: the need to create capacity to draft tax legislation, to gather and analyzing data relevant to tax matters; to undertake the detailed planning of reforms; to develop the procedural systems and administrative capacity to implement them; and, above all, to ‘sell’ the reforms to those who must approve them (politicians), those who must administer them (officials) and, not least, those who must endure them (the public). Economists, lawyers, statisticians, and management and information specialists all have critical roles to play. Sometimes experts tussle over whether control should lie in the tax administration (usually a bad idea), the Ministry of Finance, the Ministry of Development (or Planning, or whatever it may be called), the President’s office, or a special body involving some or all of these groups. The best approach may be to have, as it were, a competition of reform ideas in play. Some might argue that skilled technical resources are too scarce in most developing countries to be thus ‘wasted’ in duplication but good, workable reform ideas are in too short supply to establish a monopoly and cut off potential alternative suppliers. I return to this point later.

A common approach in some countries has been to bypass the normal machinery by appointing some kind of special tax reform commission, whether foreign, domestic, or mixed. However, the track record of such efforts is not good, both because appointing an outside group has often been simply a way to avoid having to deal with a problem for a while and because whatever such a group comes up with is seldom ‘owned’ by those who must both sell it and make it work. In other cases, by the time the commission reports the initiating problem – usually a revenue crisis – has gone away, or the commissioning government has gone and the new brooms are not interested in old ideas. Ownership matters. So does leadership. So does a coherent strategy, and of course so do adequate resources. Good tax policy planning involves economists, lawyers, administrators, and – not to be forgotten – adequate discussion with taxpayers. Successful tax reform involves all this plus solid and continuing political support and adequate administrative follow-up. It is not easy anywhere. But it can be done.

77 For an example, see the discussion of tax incentives in Indonesia in Gillis (1985). 78 For example, as Yoingco (1976) shows, for many years the Philippines for many years had by far the best developed and institutionalized tax planning process in Asia, and indeed in some respects perhaps in the developing world. The results in terms of good policy, however, are not very evident, essentially for the reasons mentioned in the text.

58

The role of outsiders such as aid agencies is more to be supportive when countries want to reform their systems than to tell them when and how to do it. Specifically, such agencies might consider funding such outwardly ‘peripheral’ – though in the long run really central – determinants of good tax policy as informed non-government ‘think tanks’, university studies in relevant fields (including foreign training in some instances), as well as data gathering and analysis both within and outside government, and, where requested and required, specific foreign experts to supply specifically needed inputs. Such activities are not glamourous and seldom produce clear short-run pay-offs. Nonetheless, they are critical to establishing institutions that, over time, will be able to provide ideas and information that may inform (and perhaps even make more articulate) the interests that, in the end, invariably shape tax systems in every country (as discussed further below). Good planning and policy formulation focuses on what matters and what can be done and pays close attention to detail and implementation (Gillis 1989; Thirsk 1997). Building up adequate institutional capacity in the tax field, both inside and outside government, is critical to being able to adapt policies to changing circumstances and needs, thus ensuring both robustness and resiliency (IDB 2006). However, even the best planning and best implementation team in the world will not produce useful results unless there is also adequate political leadership, careful attention to building the necessary political coalitions, and close attention to the perceived needs of citizens as aggregated through parties, interest groups, and what is now often called civil society. Even the best product will not sell unless it is properly marketed – to the governing elites who will have to adopt it, to the officials who will have to administer it, and finally of course to the public who will have to accept it. In the end, therefore, tax reform is always and everywhere an exercise in practical politics. Play the Right Game

Major changes in taxation generally require major changes either in the political reality of a country or in its economic circumstances. In normal times a ‘good’ tax reform—one intended to raise more revenue in a more efficient and equitable fashion, for instance—may be a bit like a ‘good’ seat belt law: if everything else stays the same, lives will be saved (the tax ratio will increase). But things do not stay the same: some people drive faster when they are belted in (administrative efforts slack off, concessions to favoured groups multiply) with the result that death rates (tax ratios) show little change. Countries tend to achieve an equilibrium position with respect to the size and nature of their fiscal systems that reflects the balance of political forces and institutions and then to stay there until ‘shocked’ into a new equilibrium (Bird 2003). Two alternative explanations may lie behind this process. Either -- somewhat improbably -- ‘supply’ (‘capacity’) factors may alter over time in such a way as to offset all attempts to raise tax ratios. Or, more plausibly, ideas as to the ‘proper’ tax level – ‘demand’ factors -- may change over time. As discussed in Section 2 above, some evidence supports the latter (and in some respects more optimistic) position. I conclude with a few reflections on how the ‘demand’ for better tax systems may be shaped by policy.

To begin with, it is clear that the tax policy world is very different in many respects now

than it was 50 or even 20 years ago. Both the economic and the intellectual environment has

59

changed. Ideas matter: as Blyth (2002, 274) puts it, “…neither material resources nor the self-interest of agents can dictate…ends or tell agents what future to construct. Ideas do this.” Of course, institutions and interests also matter in shaping tax policy. An example may help make the point.

Some years ago an analysis of Central American tax policy in a ‘class’ framework argued

that in principle changes in tax level structure (e.g., the degree of emphasis on income taxation) essentially reflected the changing political balance of power between landlords, capitalists, workers, and peasants (Best 1976). Shortly after this article appeared, an explicitly ‘leftist’ regime (the Sandinistas) took over in Nicaragua. What happened to taxes? Three things:

1. As Best (1976) would have predicted, the tax ratio rose very quickly, from 18 to 32% of

GDP within the first five years of the Sandinista regime. 2. However, almost all the increase in tax revenue came from (probably) regressive indirect

taxes and not from the (at least nominally) progressive income taxes. 3. In many ways most interestingly, once Nicaragua’s tax ratio was increased, it stayed up

there even a decade (and three subsequent governments) after the defeat of the Sandinistas.79 As this example suggests, political ideas definitely matter in taxation; but they do not

necessarily dominate. Economic and administrative realities also matter. The fiscal reality found at any point of time in any country reflects a changing mixture of ideas, interests, and institutions. Few real-world tax structures have been designed with any particular objective in mind. Often, like Topsy, they ‘have just growed’ in ways shaped by both the changing local environment and the changing external context. In the case of the United States, for example, “economic crises and wars helped create a consensus for an income tax that falls most heavily on the wealthiest taxpayers. The consensus [was] forged in the period of 1860 to 1920….” (Weisman 2002, 366). The lengthy debate about taxes that took place over this period was not really about taxes at all but rather about what kind of society Americans wanted. Much the same can be said about other countries.

Ideas on the relevant balance between taxes and society forged over the first half of the

20th century have changed in many countries, as evidenced by the death of death taxes in developed countries80 and the limited success of developing countries in achieving the high levels of income taxation to which many of them aspired in the post-colonial period (Bird and Zolt 2005). Up to the present, however, reality in terms of both tax levels and the distribution of tax burdens has changed less in most countries than ideas about what it should be. The question of how to make the ‘wicksellian connection’ operational so that good decisions—that is,

79 Peacock and Wiseman (1961) had earlier explained a similar discrete jump in tax effort and public expenditure in post-war Britain as a '‘displacement effect’: general perceptions about what is a tolerable level of taxation tend to be stable until shocked by a social upheaval so that levels of taxation that would have been previously intolerable become acceptable and remain at the new higher level after the social perturbations have disappeared. 80 For a neat explanation of this trend, encompassing changes in both economic structure and income inequality, see Bertocchi (2007).

60

decisions that reflect people’s real preferences as closely as practically feasible—will be made on both sides of the budget cannot be separated from the question of the institutional structure within which political decisions are reached and expressed.

The only way we know to help relevant decision-makers make right decisions is to

ensure that they—and ideally all those affected—are as aware as possible of all the relevant consequences. One key to good fiscal outcomes is thus to have a public finance system that links specific expenditure and revenue decisions as transparently as possible. What all this means is that if any country is to have a better tax system—better in the sense of giving the people what they want—it must first have a better political system that transmutes citizen preferences into policy decisions as efficiently as possible. “Democracy,” as Churchill once said, “is the worst form of Government except all those other forms that have been tried from time to time.”81 Just as it was argued above that ‘representation without taxation’ is not the way to produce good outcomes in decentralizing countries so ‘taxation without representation’ is not the way to create a fiscal system that is likely to be either economically or politically sustainable in the long run. Of course, in any system, no matter how democratic or non-democratic it may be, taxation is of course always and everywhere a contested concept. Some pay and some do not. Some pay more than others. Some receive compensating services, some do not. Such matters are—and in democratic states, can be—resolved only through political channels. Indeed, history suggests that the need to secure an adequate degree of consensus from the taxed is one of the principal ways in which, over the centuries, democratic institutions have spread (Sokoloff and Zolt 2005). No non-dictatorial government in this age of information and mobility can long stay in power without securing a certain degree of consent from the populace, not least in the area of taxation. State legitimacy thus rests to a considerable extent on the ‘quasi-voluntary compliance’ of citizens with respect to taxation (Levi 1988). To secure such compliance in a sustainable way tax systems must, over time, represent in some real sense the basic values of at least a minimum supporting coalition of the population.82 The central problem in many Latin American countries, for instance, is clearly inequality (de Ferranti et al. 2004). The key, and related, governance problem in most of the same countries is lack of accountability. A better tax system is critical to the solution of both problems. Reforms that link taxes and benefits more tightly such as decentralization and more reliance on user charges may help accountability -- though not necessarily reduce inequality.83 On the other hand, reforms that replace highly regressive and inelastic excises by a less

81 This quotation actually had a somewhat different implication in its original context, but nonetheless seems largely right even if one’s main concern is growth: as Lindert (2004, 344), concludes, history tells us that “the average democracy has been better for economic growth than the average autocracy….” 82 Daunton (2001, 2002) shows that a great deal of attention was paid to precisely this task in Britain, with quite different tax levels and tax mixes being found most suitable to the ‘consensus-maintaining’ objective over the years. Gillespie (1991) tells a similar (more economics-focused) tale for Canada. Lieberman (2003) to some extent tells similar stories with respect to Brazil and South Africa. For a stimulating general model of the balancing of political and economic concerns in formulating and implementing tax policy in a democratic setting, see Hettich and Winer (1999). Tridimas and Winer (2004) in effect extend this framework to non-democratic settings also. 83 As discussed earlier, another such reform is earmarking (Bird and Jun 2005), but this too may instead of improving matters worsen them if it captured by a particular interest as may happen all too easily even in developed and democratic countries. There is no such thing as a free lunch when it comes to institutional design.

61

regressive and more elastic VAT may reduce inequality—especially of course if the increased revenues are invested in growth-facilitating activities such as education and infrastructure.84 The key function of taxes in developing countries is to provide (non-inflationary) funding for pro-poor and pro-growth spending programs particularly on developing human capital. The best way to achieve this goal in most countries is likely through a broad-based non-distortionary consumption tax like VAT, as has long been recognized (Heady 2004). What countries actually do, however, is always and inevitably determined in the first instance by political and not economic calculations.85 Countries vary enormously in the effectiveness and nature of their political systems. Some may be close to ‘failed states’ in which institutions are so ineffective that it does not matter much what they attempt to do: it will not work. Others may be ‘developmentalist’ and wish to use their fiscal systems as part of a relatively dirigiste interventionist policy. Still others may be of a more laissez-faire disposition. Some may be more populist, some more elitist, some more predatory.

The dominant policy ideas in any country (about equity and fairness, efficiency, and

growth), like the dominant economic and social interests (capital, labor, regional, ethnic, rich, poor), and the key institutions, both political (democracy, decentralization, budgetary) and economic (protectionism, macroeconomic policy, market structure), interact in the formulation and implementation of tax (and budgetary) policy. Uniform results are unlikely to emerge from this always boiling cauldron with its different mixes of ingredients in each country. The changing interplay of ideas, interests, and institutions affects both the level of taxation and its structure. Indeed, as Schumpeter (1954) emphasized long ago, taxation is one of the clearest arenas in which to witness the working out of these complex forces.

Viewed in long-term perspective, few developing countries have as yet completed even

the earlier parts of the long cycle that produced the (more or less) redistributive and (more or less) growth-facilitating fiscal states now found in most developed countries—the long preparatory period during which the idea of the desirability and even necessity of a larger state and a more or less progressive fiscal system became established to different degrees in different countries.86 Some countries in Latin America, for example, might be argued to have moved more from the colonial inequality of land-based maldistribution to the modern inequality of capital-based maldistribution.87

84 Note that one implication is that relevant analysis of the incidence of policy changes must consider both sides of the budget: as Break (1974) noted, the ‘differential’ tax analysis beloved of economists (and illustrated by all the incidence studies cited in Figure 5 above) is usually not the most relevant approach to incidence for policy purposes. 85 Of course, as Hettich and Winer (1999) develop in detail, political and economic factors are often interdependent. 86 Compare the different, but parallel, stories told by Lindert (2003) and Alesina and Angeletos (2003) about how different developed countries have reached quite different fiscal equilibria. Why should uniform outcomes be expected in the much more heterogeneous developing world? 87 Warriner (1969) once noted, despairingly, that many Latin Americans did not seem to know what a good land reform means—probably because they had never seen one. Equally, in most countries of the region (as Engerman and Sokoloff (2001) almost—but not quite—say) most people may not know what either moderate or justifiable inequality might mean since they have never seen it.

62

Many governments in developing and transitional countries -- not just those in Latin America -- are in dire straits. Even countries that have reached relatively safe harbors politically, achieving a certain degree of legitimacy and stability, almost always feel—often correctly—that they are in an economically precarious situation. The budget is politically and economically constrained. Life is difficult. Nothing can be done. All this may be true to some extent, but it is also both too much a counsel of despair and too easy a way out. Even in the most hopeless situations something usually can be done to improve matters. No doubt there will continue in most countries to be considerable dispute over what should be done to improve tax systems. Indeed, as I suggested earlier, in most countries it would be better if there were even more informed public dispute about such matters because unless and until an adequate degree of political consensus on what should be done is achieved, no significant tax changes are likely to be made. In short, to a considerable extent the main tax challenge facing many developing countries is simply that there is as yet no implicit “… social contract between governments and the general population of the kind that is embedded in taxation and fiscal principles and practices in politically more stable parts of the world” (Lledo, Schneider, and Moore 2004, 39).

History tells us that such principles do not become embedded either painlessly or quickly. The specific substantive suggestions that Lledo, Schneider, and Moore (2004) make to improve matters -- such as better VAT administration on a broader base --are already the stuff of countless existing reports. I agree. Most countries should do all (or most of) the good things that experts such as us tell them. But the real question is: why have so many done so little? Lledo, Schneider, and Moore (2004, 40) suggest, wistfully, that if Latin American countries wish to improve their tax systems they should “…improve political institutions in ways that broaden and deepen social contracts. For example, create more responsive and less clientelistic political parties, more cohesive and less polarised party systems, and improved capacity of civil society to monitor government and participate in tax debates.” In other words, what they say is, as I noted earlier, that there can be no good taxation without good representation. They are right. But how useful is it to advise countries they should be something other than what they are?

In the end, if a country needs or wants better tax policy or administration, it can have it: the answer largely lies in its own hands. Even those who want to do the right thing, however, can often use help in finding out just what is right and how it can best be done. It is always easy and seldom effective for those not in a game to give advice to those who are trying to play it. In general, however, outside agencies interested in fostering better sustainable tax systems in developing country will likely employ their efforts and resources most usefully if they play in the right game. Fifty years of experience tells us that the right game is not the short-term political game in which policy decisions are made. Instead, it is the long-term game of building up the institutional capacity both within and outside governments to articulate relevant ideas for change, to collect and analyze relevant data, and of course to assess and criticize the effects of such changes as are made. Such long-term ‘institution-building’ activities are seldom immediately rewarding. They are certainly out of fashion. It seems much more appealing and immediately productive to establish ‘benchmarks’ for success, to support this particular organizational change reform here (revenue authority) and that new technology (computerization) there, in the apparent belief that such simple ‘one-size-fits-all’ approaches can provide quick (but sustainable!)

63

answers to the many complex problems inherent in policy reform in difficult environments. It may be appealing. But it is wrong.

64

References

Acosta, O. L. and R.M. Bird (2005) “The Dilemma of Decentralization in Colombia,” in R.M. Bird, J. Poterba and J. Slemrod, eds., Fiscal Reform in Colombia (Cambridge MA: MIT Press). Adam, Christopher S. and Stephen A. O’Connell (1999) “Aid, Taxation and Development in Sub-Saharan Africa,” Economics and Politics, 11 (3): 225-53 Addison, Tony, and Jorgen Levin (2006) “Tax Policy Reform in Developing Countries,”Third draft 2006-06-17, WIDER Alesina, A.and G. Angeletos (2003) “Fairness and Redistribution: U.S. versus Europe,” Working Paper 9502, National Bureau of Economic Research, MA, February. Alm, J. and H. López-Castaño (2005) “Payroll Taxes in Colombia,” in R.M. Bird, J. Poterba and J. Slemrod, eds., Fiscal Reform in Colombia (Cambridge MA: MIT Press). Alm, J. and J. Martinez-Vazquez (2003) “Institutions, Paradigms, and Tax Evasion in Developing and Transition Economies,” in J. Martinez-Vazquez and J. Alm, eds., Public

Finance in Developing and Transitional Economies (Cheltenham UK: Edward Elgar). Alm, J., J. Martinez-Vazquez and F. Schneider (2004) “ ‘Sizing’ the Problem of the Hard-to-Tax,” in J. Alm, J. Martinez-Vazquez and S. Wallace, eds., Taxing the Hard-to-Tax (Amsterdam: Elsevier). Arbeláez, M. A., L.E. Burman, and S. Zuluaga (2005) “The Bank Debit Tax in Colombia,” in R.M. Bird, J. Poterba and J. Slemrod, eds., Fiscal Reform in Colombia (Cambridge MA: MIT Press). Auerbach, A. J. and J.R. Hines, Jr. (2002) “Taxation and Economic Efficiency,” in A.J. Auerbach and M.Feldstein, eds., Handbook of Public Economics, 3 ( Amsterdam: North-Holland). Auriol, E. and M.Warlters (2005) “Taxation Base in Developing Countries,” Journal of Public

Economics, 89: 625-46. Baer, K., O.P. Benon and J.A. Toro Rivera (2002) Improving Large Taxpayers’ Compliance: A

Review of Country Experience. Occasional Paper 215. (Washington: International Monetary Fund). Bahl, R.W. (1971) “A Regression Approach to Tax Effort and Tax Ratio Analysis,” IMF Staff

Papers, 18: 570-608.

65

Bahl, R.W. and J. Martinez-Vazquez (1992) “The Nexus of Tax Administration and Tax Policy in Jamaica and Guatemala,” in R.M. Bird and M. Casanegra de Jantscher, eds., Improving Tax

Administration in Developing Countries (Washington: International Monetary Fund) Barreix, A. and J. Roca (2006) “Arquitectura de una propuesta de reforma tributaria,” Universidad Católica de Uruguay, April. Baunsgaard, T. and M. Keen (2005) “Tax Revenue and (or?) Trade Liberalization,” Working Paper WP/05/112, International Monetary Fund, June. Becker, J. and C. Fuest (2005) “Does Germany Collect Revenue from Taxing Capital Income?” CESifo Working Paper 1489, June. Bergman, Marcelo (2002) “Who Pays for Social Policy? A Study on Taxes and Trust,” Journal

of Social Policy, 31 (2): 289-305. Bergman, Marcelo (2003) “Tax Reforms and Tax Compliance: The Divergent Paths of Chile and Argentina,” Journal of Latin American Studies, 35:593-624. Bertocchi, G. (2007) “The Vanishing Bequest Tax: The Comparative Evolution of Bequest Taxation in Historical Perspective,” IAZ DP No. 2578, Institute for the Study of Labor, Bonn, January. Best, M.H. (1976) “Political Power and Tax Revenues in Central America,” Journal of

Development Economics, 3: 49-82. Bird, R.M. (1974) Taxing Agricultural Land in Developing Countries (Cambridge MA: Harvard University Press) Bird, R.M. (1976) Charging for Public Services: A New Look at an Old Idea (Toronto: Canadian Tax Foundation) Bird, R.M. (1982) “Closing the Scissors, or The Real Public Sector has Two Sides,” National

Tax Journal, 35 (4): 477-81.

Bird, R. M. (2000) “Tax Incentives for Investment in Developing Countries,” in G. Perry, J. Whalley, and G. McMahon, eds., Fiscal Reform and Structural Change in Developing Countries (2 vols.; London: Macmillan for International Development Research Centre), 1, 201-21.

Bird, R.M. (2001) Intergovernmental Fiscal Relations in Latin America: Policy Designs and

Policy Outcomes (Washington: Sustainable Development Department, Inter-American Development Bank).

Bird, R.M. (2002) “Why Tax Corporations?” Bulletin for International Fiscal Documentation, 56 (5): 194-203.

66

Bird, R. M. (2003) “Taxation in Latin America: Reflections on Sustainability and the Balance between Efficiency and Equity,” ITP Paper 0306, International Tax Program, Rotman School of Management, University of Toronto, June.

Bird, R.M. (2004) “Managing Tax Reform,” Bulletin for International Fiscal Documentation, 58 (2): 42-55.

Bird, R.M. (2004a) “Administrative Dimensions of Tax Reform,” Asia-Pacific Tax Bulletin, 10 (3): 134-50. Bird, R.M. (2005) “Taxing Electronic Commerce: The End of the Beginning?” Bulletin for

International Fiscal Documentation, 59 (4, 2005): 130-40 Bird, R.M. (2005a) “Evaluating Public Expenditures: Does It Matter How They are Financed?” in Anwar Shah, ed., Fiscal Management. Public Sector Governance and Accountability Series (Washington: World Bank)

Bird, R.M. and S.Banta (2000) “Fiscal Sustainability and Fiscal Indicators in Transitional Countries,” in A. Shapleigh, F. Andic, and S. Banta, eds., Transition Economies and Fiscal

Reforms. Proceedings of the Conference on Central and Eastern Europe and the New Independent States, Istanbul, July 1999 (Washington: USAID, 2000), pp. 13-42.

Bird, R.M. and M. Casanegra de Jantscher, eds. (1992) Improving Tax Administration in

Developing Countries (Washington: International Monetary Fund). Bird, R.M. and D. Chen (1998) “The Fiscal Framework for Business in Asia,” in W. Dobson, ed., Fiscal Frameworks and Financial Systems in East Asia: How Much Do They Matter? (Toronto: University of Toronto Press)

Bird, R.M. and P-P. Gendron (2006) “Is VAT the Best Way to Impose a General Consumption Tax in Developing Countries? Bulletin for International Taxation, 60 (7): 287-96

Bird, R.M. and P-P. Gendron (2007) The VAT in Developing and Transitional Countries (Cambridge University Press). Bird, R.M. and J. Jun (2005) Earmarking in Theory and Korean Practice. Special Report, International Tax and Investment Center, Washington, 2005. Bird, R.M. and B.D. Miller (1989) “The Incidence of Indirect Taxation on Low-Income Households in Jamaica”, Economic Development and Cultural Change, 37 (2, 1989): 393-409 Bird, R.M. and B.D. Miller (1989a) “Taxation, Pricing and the Poor,” in R.M. Bird and S. Horton, eds., Government Policy and the Poor in Developing Countries (Toronto: University of Toronto Press) Bird, R.M. and J. Mintz (2000) “Tax Assignment in Canada: A Modest Proposal,” in H. Lazar, ed., The State of the Federation, 2000-01: Toward a New Mission Statement for Canadian Fiscal

Federalism (Montreal: McGill-Queen’s University Press for School of Policy Studies, Queen’s University)

67

R.M. Bird . and O. Oldman (2000) Improving Taxpayer Service and Facilitating Compliance in

Singapore, PREM Note Number 48, The World Bank, December Bird, R.M. and E. Slack (2004) International Handbook of Land and Property Taxation

(Cheltenham, UK and Northampton, MA: Edward Elgar). Bird, R.M. and E. Slack (2006) “Taxing Land and Property in Emerging Economies: Raising Revenue…and More?” ITP Paper 0605, International Tax Program, Rotman School of Management, University of Toronto, July Bird, R.M. and M. Smart (2002) “Intergovernmental Fiscal Transfers: Lessons from International Experience,” World Development, 30 (6, 2002): 899-912. Bird, R.M. and T. Tsiopoulos (1997) “User Charges for Public Services: Potentials and Problems,” Canadian Tax Journal, 45 (1, 1997), 25-86 Bird, R.M. and S.Wallace (2004) “Is It Really so Hard to tax the Hard-to-Tax? The Context and Role of Presumptive Taxes,” in J. Alm, J.Martinez-Vazquez and S. Wallace, eds., Taxing the

Hard-to-Tax: Lessons from Theory and Practice (Amsterdam: North-Holland)

Bird, R.M. and S. Wallace (2005) “Revenue-Maximizing Tax Rates,” in J. Cordes, R.Ebel and J. Gravelle, eds., The Encyclopedia of Taxation and Tax Policy (2nd ed.; Washington: The Urban Institute Press) Bird, R.M. and S.Wallace (2006) “Taxing Alcohol: Reflections from International Experience,” in S.Cnossen, ed., Excise Tax Policy and Administration (Pretoria: UNISA Press), 21-60. Bird, R.M. and E.M. Zolt (2005) “Redistribution via Taxation: The Limited Role of the Personal Income Tax in Developing Countries,” UCLA Law Review, 52 (6, 2005): 1627-95 Bird, R.M., J. Martinez-Vazquez and B. Torgler (2006) “Societal Institutions and Tax Effort in Developing Countries,” in J. Alm, J. Martinez-Vazquez and M.Rider, eds., The Challenges of

Tax Reform in a Global Economy (New York: Springer) Blyth, M. (2002) Great Transformations: Economic Ideas and Institutional Change in the

Twentieth Century (Cambridge: Cambridge University Press). Boadway, R.W. (2005) “Income Tax Reform for a Globalized World: The Case for a Dual Income Tax,” Journal of Asian Economics, 16: 910-27. Boadway, R.W. and A. Shah (1995) “Perspectives on the Role of Investment Incentives in Developing Countries,” in A. Shah, ed., Fiscal Incentives for Investment and Innovation (Oxford: Oxford University Press Published for the World Bank). Boylan, Delia M. (1996) “Taxation and Transition: The Politics of the 1990 Chilean Tax Reform,” Latin American Research Review, 31 (1): 7-31

68

Break, G.F. (1974) “The Incidence and Economic Effects of Taxation,” in A. Blinder et al., Economics of Public Finance (Washington DC: Brookings Institution). Breton, A. (1996) Competitive Governments (Cambridge: Cambridge University Press). Brett, C. and M. Keen. 2000. “Political Uncertainty and the Earmarking of Environmental Taxes,” Journal of Public Economics, 75: 315-40 Buchanan, J. (1963) "The Economics of Earmarked Taxes." Journal of Political Economy, 71: 457-69 Casanegra de Jantscher, M. (1990) “Administering the VAT,” in Gillis, M., C. Shoup, and G.P. Sicat, eds. Value Added Taxation in Developing Countries (Washington: World Bank). Chattopadhyay, S. and A. Das Gupta, “The Compliance Cost of the Personal Income Tax and its Determinants,” National Institute of Public Finance and Policy, New Delhi, 2002 Chu, K., H. Davoodi, and S. Gupta (2000) “Income Distribution and Tax and Government Social Spending Policies in Developing Countries,” Working Papers No. 214, UNU/WIDER, December. Cnossen, S., ed. (2005) Theory and Practice of Excise Taxation: Smoking, Drinking, Gambling,

Polluting and Driving (Oxford: Oxford University Press) Cnossen, S. ed., (2006) Excise Tax Policy and Administration (Pretoria: UNISA Press) Cnossen, S. and M. Smart (20050 “Tobacco Taxation,” in S. Cnossen, ed., Theory and Practice of

Excise Taxation (Oxford: Oxford University Press) Cremer, J., A. Estache, and P.Seabright (1995) "The Decentralization of Public Services: Lessons from the Theory of the Firm," in A. Estache, ed., Decentralizing Infrastructure: Advantages and

Limitations. World Bank Discussion Paper 290, Washington. Das-Gupta, A. and D. Mookerjhee (1998) Incentives and Institutional Reform in Tax

Enforcement (New Delhi: Oxford University Press). Daunton, M. (2001) Trusting Leviathan: The Politics of Taxation in Britain 1799-1914 (Cambridge: Cambridge University Press). Daunton, M. (2002) Just Taxes: The Politics of Taxation in Britain 1914-1979 (Cambridge: Cambridge University Press). De Ferranti, D. et al. (2004) Inequality in Latin America: Breaking with History? (Washington: The World Bank).

69

Devereux, M.P. and R. Griffith (1998) “Taxes and the Location of Production: Evidence from a Panel of U.S. Multinationals,” Journal of Public Economics, 68 (3): 335-67 Easterly, W. (2002) The Elusive Quest for Growth (Cambridge MA: MIT Press). Ebrill, L., M. Keen, J-P. Bodin, and V. Summers (2001) The Modern VAT (Washington: International Monetary Fund). Edling, Herbert and David Nguyen-Thanh (2006) “Taxes and Culture – Tax Reforms for Sustainable Development” (2006) Fiscal Studies No. 5, Public Finance and Administrative Reform Studies, GTZ, Eschborn. Elster, J. (1984) Ulysses and the Sirens: Studies in Rationality and Irrationality (Cambridge UK: Cambridge University Press). Emran, M. S. and J.E. Stiglitz (2005) “On Selective Indirect Tax Reform in Developing Countries,” Journal of Public Economics, 89: 599-623. Engel, E., A. Galetovic and C.E. Raddatz (1999) “Taxes and Income Distribution in Chile: Some Unpleasant Distributional Arithmetic,” Journal of Development Economics, 59 (1): 155-92. Engelschalk, M. (2000) Computerizing Tax and Customs Administration, PREM Note 44, World Bank, October 2000. Engerman, S.L.. and K.L. Sokoloff (2001) “Inequality, Institutions and Differential Patterns of Growth among New World Economies,” National Bureau of Economic Research. Estela, M.(2001) Manuel Estela, Strengthening Peru’s Tax Agency, PREM Note 60, World Bank, Washington, November Evans, C. (2003) “Studying the Studies: An Overview of Recent Research into Taxation Operating Costs,” eJournal of Tax Research, 1 (1): 64-92. Ferraz, C. and F. Finan (2007) “Exposing Corrupt Politicians: The Effects of Brazil’s Publicly Released Audits on Electoral Outcomes,” Paper presented to Conference on Fiscal Federalism in Latin America, UCLA Center for Latin American Studies, February. Ferris, J.S. and S.L. Winer (2003) “Searching for Keynes: With Application to Canada, 1870-2000,” CESifo Working Paper 1016. Fjeldstad, O-H. (2002) “Fighting Fiscal Corruption: The Case of the Tanzania Revenue Authority,” CMI Working Paper 2002, Bergen Fjeldstad, O-H. (2005) “Corruption in Tax Administration: Lessons from Institutional Reforms in Uganda,” WP 2005:10, CMI Working Paper, Bergen.

70

Fox, W.F. and T. Gurley (2005) “An Exploration of Tax Patterns around the World,” Tax Notes

International, February 28: 793-807. Frey, Bruno S. (2002) The Role of Deterrence and Tax Morale in Taxation in the European

Union. Jelle Zijlstra Lecture 1, Netherlands Institute for Advanced Study in the Humanities and Social Sciences. Fullerton, D. (1996) "Why Have Separate Environmental Taxes?" In J. Poterba, ed., Tax Policy and

the Economy 10 (Cambridge, MA: MIT Press) Gallagher, Mark (2004) “Benchmarking Tax Systems,” Public Administration and Development, 25 : 125-44. Gemmell, N. and O. Morrissey (2003) “Tax Structure and the Incidence on the Poor in Developing Countries,” Centre for Research on Economic Development and International Trade Research paper No. 03/18, University of Nottingham, October Gillespie, W.I. (1991) Tax, Borrow, and Spend: Financing Federal Spending in Canada 1867-

1990 (Ottawa: Carelton University Press). Gillis, M. (1985) “Micro and Macroeconomics of Tax Reform: Indonesia,” Journal of

Development Economics, 19:221-554. Gillis, M., ed. (1989) Tax Reform in Developing Countries (Durham NC: Duke University Press). Glenday, G. and D. Hollinrake (2005) “Assessment of the Current State of VAT Implementation in SADC Member States,” November. Gomez Sabaini, Juan C. (2005) “Evolución y situación tributaria actual en América Latina: Una serie de temas para discusión,” Septiembre. Gordon, R. and W. Li (2005) “Tax Structure in Developing Countries: Many Puzzles and A Possible Explanation,” Working Paper 11267, National Bureau of Economic Research, Cambridge MA, April. Gordon, R., L. Kalamokidis, and J. Slemrod (2004) “Do We Now Collect Any Revenue from Taxing Capital Income,” Journal of Public Economics, 88 (5): 981-1009. Graetz, M.J. and I. Shapiro (2005) Death by a Thousand Cuts: The Fight over Taxing Inherited

Wealth (Princeton: Princeton University Press). Gwilliam, K. and Z. Shalizi. 1999. “Road Funds, User Charges, and Taxes,” World Bank Research

Observer, 14 (no. 2): 159-85.

71

Harberger, A.C. (1989) “Lessons of Tax Reform from the Experiences of Uruguay, Indonesia, and Chile,” in M. Gillis, ed., Tax Reform in Developing Countries (Durham NC: Duke University Press). Harberger, A.C. (2006) “Taxation and Income Distribution: Myths and Realities,” in J. Alm, J. Martinez-Vazquez and M.Rider, eds., The Challenges of Tax Reform in a Global Economy (New York: Springer) Heady, C. (2004) “Taxation Policy in Low-income Countries,” in T. Addison and A. Roe, eds., Fiscal Policy for Development: Poverty, Reconstruction and Growth (London: Palgrave) Hettich, W. and S. Winer (1999) Democratic Choice and Taxation (Cambridge: Cambridge University Press). Hirschman, A.O. and R.M. Bird (1968) Foreign Aid - A Critique and A Proposal. Essays in International Finance, No. 69. (Princeton, N.J.: Princeton University)

Hoffman, Barak D. and Clark C. Gibson (2006) “Political Accountability and Fiscal Governance: Evidence from Tanzania and Zambia.” www.econ.ucsd.edu/seminars/0607seminars/gibson_paper.pdf Hood, C. (1976) The Limits of Administration (New York: Wiley). Hughes, G. (1987) “The Incidence of Fuel Taxes: A Comparative Study of Three Countries,” in D. Newbery and N. Stern, eds, The Theory of Taxation for Developing Countries (New York: Published for the World Bank by Oxford University Press). Inter-American Development Bank (IDB) (2006) The Politics of Policies (Washington) International Monetary Fund (IMF) (2005) “Monetary and Fiscal Policy Design Issues in Low-Income Countries,” August. International Monetary Fund (IMF) (2005a) Government Finance Statistics Yearbook 2005 (Washington). International Monetary Fund (IMF) (2006) “Fiscal Policy for Growth and Development: An Interim Report,” DC2006-0003, Development Committee, April. Jenkins, G., ed. (1996) Information Technology and Innovation in Tax Administration (The Hague: Kluwer Law International) Jha, P. and F. Chaloupka, eds.(2000) Tobacco Control in Developing Countries (Oxford University Press). Kaldor, N. (1956) Indian Tax Reform (New Delhi: Government of India Ministry of Finance).

72

Kaldor, N. (1963) “Will Underdeveloped Countries Learn to Tax?” Foreign Affairs, 41: 410-19. Kaplow, L. (1996) “The Optimal Size of Public Goods and the Distortionary Cost of Taxation,” National Tax Journal, 49: 513-33 Keen, M. (2006) “VAT Attacks! Second Best Perspectives on the Value Added Tax,” Paper presented to International Institute of Public Finance, Cyprus, August 2006. Keen, M. and B. Lockwood (2006)”Is the VAT a Money Machine?” National Tax Journal, 49 (4): 905-28 Kenny, L.W. and S.L. Winer (2006) “Tax Systems in the World: An Empirical Investigation into the Importance of Tax Bases, Administration Costs, Scale and Political Regime,” International

Tax and Public Finance, 13 (2/3): 181-216. Keynes, J.M. (1936) The General Theory of Employment, Interest and Money (New York: Harcourt, Brace). Levi, M. (1988) Of Rule and Revenue (Berkeley: University of California Press). Lieberman, Evan S. (2002) “Taxation Data as Indicators of State-Society Relations: Possibilities and Pitfalls in Cross-National Research,” Studies in Comparative International Development, 36 (4): 89-115. Lieberman, E.S. (2003) Race and Regionalism in the Politics of Taxation in Brazil and South

Africa (Cambridge: Cambridge University Press). Lindert, P.H. (2003) Growing Public: Social Spending and Economic Growth Since the

Eighteenth Century (Cambridge: Cambridge University Press). Lindert, P.H. (2004) “Voice and Growth; Was Churchill Right?,” Journal of Economic History,

63 (2): 315-50. Lledo, Victor, Aaron Schneider, and Mick Moore (2003) “Pro-poor Tax Reform in Latin America: A Critical Survey and Policy Recommendations,” IDS Report commissioned by DFID, March. Lledo, Victor, Aaron Schneider, and Mick Moore (2004) “Governance, Taxes, and Tax Reform in Latin America,” IDS Working Paper 221, Institute of Development Studies, Brighton. Lora, E. and M. Cardenas (2006) “La reforma de las instituciones fiscales en América Latina,” Inter-American Development Bank, Research Department, Documento de Trabajo #559, April Martin, A. and W.A. Lewis (1956) “Patterns of Public Revenue and Expenditure,” Manchester

School of Economic and Social Studies, 24: 203-44

73

Martinez-Vazquez, J. (2001) “Mexico: An Evaluation of the Main Features of the Tax System,” Working Paper 01-12, International Studies Program, Andrew Young School of Policy Studies, Georgia State University, November Martinez-Vazquez, J. and R.M. McNab (2000) “The Tax Reform Experiment in Transitional Countries,” National Tax Journal, 53 (2): 273-98. Martinez-Vazquez, J. and A. Timofeev (2005) “Choosing between Centralized and Decentralized Models of Tax Administration,” Working Paper 05-02, International Studies Program, Andrew Young School of Policy Studies, Georgia State University, January McCleary, W. (1991) "The Earmarking of Government Revenue: A Review of Some World Bank Experience," World Bank Research Observer, 6: 81-104. McIntyre, M. and O. Oldman (1975) Institutionalizing the Process of Tax Reform: A

Comparative Analysis (Amsterdam: International Bureau of Fiscal Documentation). McKenzie, K.J., J.M. Mintz, and K.A. Scharf (1997) “Measuring Effective Tax Reates in the Presence of Multiple Inputs: A Production Based Approach,” International Tax and Public

Finance, 4 (3): 337-59. McLaren, J., ed. (2003), Institutional Elements of Tax Design (Washington: World Bank). McLure, C.E. (1999) “Tax Holidays and Investment Incentives: A Comparative Analysis,” Bulletin for International Fiscal Documentation, 53 (8): 326-39 Messere, K., F. de Kam, and C. Heady (2003) Tax Policy: Theory and Practice in OECD

Countries (London: Oxford University Press) Mikesell, J. (2007) “Developing Options for the Administration of Local Taxes: An International Review,” Public Budgeting and Finance, 27 (1): 41-68. Mintz, Jack M. (2006) The 2006 Tax Competitiveness Report: Proposals for Pro-Growth Tax

Reform. (Toronto: C.D. Howe Institute). Moore, M. (2004) “Taxation and the Political Agenda, North and South,” Forum for

Development Studies, No.1-2004: 7-32. Morrisset, J. and A. Izquierdo (1993) “Effects of Tax Reform on Argentina’s Revenues,” World Bank Policy Research Working Paper WPS 1192, September. Musgrave, R.A. (1969) Fiscal Systems (New Haven: Yale University Press). Musgrave, R.A. (1981) “A Reappraisal of Financing Social Security,” in F. Skidmore, ed., Social

Security Financing (Cambridge MA: MIT Press)

74

OECD (Organisation for Economic Co-operation and Development) (2006) “Reforming Personal Income Tax,” Policy Brief, March. Peacock, A. and J. Wiseman (1961) The Growth of Government Expenditures in the United

Kingdom (Princeton: Princeton University Press). Radian, A. (1980) Resource Mobilization in Poor Countries: Implementing Tax Reform (New Brunswick NJ: Transaction Books) Rajkumar, A.S. (2004) “Tying the Hands of Government in Brazil Through Revenue Earmarking: How Can This Be Done Optimally?” World Bank, Washington. Rodden, J., G. Eskeland, and J. Litvack, eds. (2003) Fiscal Decentralization and the Challenge

of Hard Budget Constraints (Cambridge, MA: MIT Press) Schneider, F. and R. Klinglmair (2004) “Shadow Economies Around the World: What Do We Know?” Working Paper No. 0408, Department of Economics, Johannes Kepler University of Linz, April. Schumpeter, J.A. (1954) History of Economic Analysis (New York: Oxford University Press). Shah, A., ed. (1995) Fiscal Incentives for Investment and Innovation (Oxford: Oxford University Press Published for the World Bank). Shah, A. and J. Whalley (1990) “Tax Incidence Analysis of Developing Countries: An Alternative View,” World Bank Economic Review, 5: 535-52. Shaw, G.K.(1981) “Leading Issues of Tax Policy in Developing Countries: The Economic Problems,” in Alan Peacock and Francesco Forte, eds., The Political Economy of Taxation (New York: St. Martin’s Press) Shoup, C.S. (1991) “Melding Architecture and Engineering: A Personal Retrospective on Designing Tax Systems,” in Lorraine. Eden, ed., Retrospectives on Public Finance (Durham, NC: Duke University Press, 1991) Shekidele, C. (1999) “Measuring the Compliance Cost of Taxation Excise Duties 1995-96,” African Journal of Finance and Management, 7 (2): 72-84 Slemrod, J. (1992) Why People Pay Taxes (Ann Arbor: University of Michigan Press) Slemrod, J. and E. Bakija (1996) Taxing Ourselves: A Citizen’s Guide to the Great Debate on

Tax Reform (Cambridge MA: MIT Press) Slemrod, J. and S. Yitzhaki (2002) “ Tax Avoidance, Evasion, and Administration,” in Alan J. Auerbach and Martin Feldstein, eds., Handbook of Public Economics, vol. 3 (New York: Elsevier Science)

75

Sokoloff, K.L. and E.M. Zolt (2005) “Inequality and Taxation: Evidence from the Americas on How Inequality may Influence Tax Institutions,” Tax Law Review, 59 (2): 167-241. Steinmo, S. (1993) Taxation and Democracy: Swedish, British, and American

Approaches to Financing the Modern State (New Haven: Yale University Press). Surrey, S.S. and P.R. McDaniel (1985) Tax Expenditures (Cambridge, MA: Harvard University Press). Swift, Z.L. (2006) “Managing the Effects of Tax Expenditures on the National Budget,” World Bank Policy Research Working Paper No. 3927, May. Taliercio, R. (2004) “Designing Performance: The Semi-Autonomous Revenue Authority Model in Africa and Latin America,” World Bank Policy Research Working Paper No. 3423, October Tanzi, V. (1970) “International Tax Burdens: A Study of Tax Ratios in the OECD Countries,” in Taxation: A Radical Approach (London: Institute for Economic Affairs). Tanzi, V. (1973) “The Theory of Tax Structure Change During Economic Development: A Critical Survey,” Rivista de Diritto Fiananziario e Scienza delle Finanze, 32: 199-208. Tanzi, V. (1987) “Quantitative Characteristics of the Tax Systems of Developing Countries,” in D. Newbery and N. Stern, eds., The Theory of Taxation in Developing Countries (Published for the World Bank by Oxford University Press). Tanzi, V. and H. Zee (2000) “Tax Policy for Emerging Markets: Developing Countries,” National Tax Journal, 53 (June): 299-328. Teja, R.S. (1988) "The Case for Earmarked Taxes." International Monetary Fund Staff Papers. 35: 523-533. Thirsk, W.R. (1997) Tax Reform in Developing Countries (Washington: World Bank). Toye, J. (2000) “Fiscal Crisis and Fiscal Reform in Developing Countries,” Cambridge Journal

of Economics, 24 (1): 21-44. Tridimas, G. and S.L.Winer (2004) “The Political Economy of Government Size,” European

Journal of Political Economy, 21 (3): 643-66. UN Millennium Project (2005) Investing in Development (New York) Vazquez-Caro, J. (2005) “Benchmarking Complex Systems: The Search for a North Star,” Unpublished paper, Bogota, June.

76

Vazquez-Caro, J., G. Reid and R.M. Bird (1992) Tax Administration Assessment in Latin

America. Latin America and Caribbean Technical Department, Regional Studies Program, Report No. 13. (Washington: World Bank). Wallace, S. (2003) “Module 3: Fiscal Architecture”, World Bank Course on Practical Issues in Tax Policy, Washington, April. Warlters, Michael and Emmanuelle Auriol (2005) “The Marginal Cost of Public Funds in Africa,” World Bank Policy Research Working Paper No. 3679, August. Warriner, D. (1969) Land Reform in Theory and Practice (Oxford: Clarendon Press). Webber, C. and A. Wildavsky (1986) A History of Taxation and Expenditure in the Western World (New York: Simon and Shuster) Weisman, S.R. (2002) The Great Tax Wars (New York: Simon and Schuster). Wong, C.P.W. and R.M. Bird (2005) “China’s Fiscal System: A Work in Progress,” International Tax Program, Rotman School of Management, University of Toronto, October. World Bank (1991) Lessons of Tax Reform (Washington). World Bank (2003) Ukraine: Tax Policy and Tax Administration, Report No. 26221-UA, World Bank, Kyiv, March World Bank and International Finance Corporation (2006) Doing Business in 2006. Washington, DC. Yoingco, A.Q. (1976) “Tax Research: The Philippine Experience,” Graduate School, Centro Escolar University, Manila, March.