Journal of Development Economics 5(15’78)233-258. 0 North-Holland
TAXATION AND INCOME DISTRIBUTION: THE COLOMBIAN TAX REFORM OF 1974 Malcolm GILLIS* Harvard
Charles E. McLURE, Rice University, National
Btweatr of Economic
MA 02138, U.S.A.
Recerved April 1977, revised version received November 1977 In late 1974 the Government of Colombia implemented a major tax reform packag : embracing virtually all aspects of the revenue system. The reform was designed to serve allocative and stabilization as well as redistributive goals. This paper focuses upon the likely impact of the 1974 tax measures upon income distribution. Section 2 describes the prevailing distribution of mcome prior to the refarm. Section 3 presents a short outline of the principal features of the 1974 package, and sections 4 through 5 provide an evaluation of its redistributive effects. The authors conclude that the initial impact of the reform was undoubtedly progrersive, as it likely served to shift as muzh as 1.5 percent of GDP away from the top quintile of the income distribution.
1. Introduction During the last four months of 1974 the Government of Colombia introduced a major tax reform. Enacted under emergency powers prob.ided by the Constitution, this reform covered the sales tax, export taxes and incentives, adjustments in import surcharges, and the tax trea:ment of governmental agencies, as well as the personal and company income taxes.’ *The authors are, respectively, Institute Fellow of the Harvard Institute for International Development and Lecturer in Economics at Harvard University, and Professor of Economics at Rice University and Executive Director for Research of the NBER. They wish to-express their appreciation to Guillermo Perry Rubio for his patient and lucid explanations of the many complex features of the 1974 reform, and to thank an anonymous referee for providing several valuable suggestions adopted in the revision of this paper. But the opinions expressed here- and any errors of fact, interpretation, or analysis -a~ solelv the responsibility of the authors. ‘It might reasonably be asked how a government that came into office on August 9 could, by the end of the emergency period on October 31, produce a reform package of such magnitude and complexity. The answer is that a small group of Colombian experts appointed by then President-elect Lopez Michelsen had been working in secret since May, analyzing alternatives for tax reform and drafting legislation. Beyond that, many of the reforms were solidly based on the proposals that the Musgrave Commission had made during its extensive 1968-69 examination of the Colombian fiscal system. Most of the -sperms who served on the Michelsen task
M. Gillis and C.E. McLure,
Generally speaking, the goals of the tax reform were (a) to increase the progressivity of the tax system, (b) to reduce the distorting effect the tax system had upon the allocation of resources, (c) to promote economic stability, both by increasing revenue on a once-and-for-all basis and by enhancing the built-in response of the tax system to growth in national income, and (d) to simplify both administration and compliance, and in the process reduce evasion and increase yields. With a’ few exceptions, the new reforms go far in achieving all these objectives. In this pa?er we focus primarily upon distributional implications of various of the most important reforms.z In section 2 we set the stage for what follows by describing the prevailing distribution of income in Colombia and the likely incidence among income classes of taxes levied in 1970. Section 3 is devoted to a short outline of t? : tax reform. The evaluation of the reform is contained in Fections 4-6. That is. section 4 provides a qualitative evaluation of the distributional effects of the various components of the tax reform package; section 5 discusses the consequences of the judicial decisions that have nullified many of the gains that would have otherwise been made in tax administration; and section 6 presents results of crude attempts to quantify the distributional effects of the tax reform. Although some signs of erosion in the initial impact of the 1974 measures began to appear by 1976, the reforms are, for the most part, found to be progressive. In mos! cases the increase in progressivity comes primarily from higher taxes on upper income groups, rather than from lower’ burdens on low-income groups. This is inevitably the case when reform is so heavily focused on the personal income tax, which is paid in important amounts by only the top 10-15 percent of households. The final section offers a . few concluding remarks about the political feasibility of extensive tax reform iri developing countries. 2. Income distribution and tax incidence During the past decade nearly a half-dozen studies have been made of the distribution of income in Colombia. While these studies, summarized rn Berry and Urrutia (1976), have shown marked disparities in methodology, data sources, and the choice of whether to use individuals or households i, T the basis for the distribution estimates, they have produced a r’emarkably force had had experience on the staff of the Musgrave Commission or in follo&up analysis within the government or had been deeply involved in the public debate over the Musgrave proposals during the intervening half-dozen years. This experience and these debates enabled these experts to improve further upon the Musgrave proposals in many cases, and in others to go into questions not covered there-and to go from conceptual arguments to legislation in short order. Thus, the reforms were made quickly, but they were not ill-conceived. *We have elsewhere analyzed at some length the implications of the reform for resource aliocation and economic stability. See Gillis and McLure (1977).
M. Gillis and C.E. McLure,
Taxation and income distribution
uniform picture of income inequality. McLure (1975) concluded that the poorest two-thirds of families or individuals receive about one-quarter of personal income and that the most fortunate 10 percent receive some 40 to 50 percent of such income. Inequality is greater, it appears, in rural areas than in the urban sector.3 The national pattern of income distribution just described would strike anyone familiar with similar estimates for developed countries as exhibiting extreme inequality.4 But Berry and Urrulia found that even when compared with distributions in other developing countries, including those in Latin America, the Colombian distribution is heavily skewed. Prior to 1974, the Colombian tax system coniributed to the reduction of inequality in the distribut,ion of income, but only at the top of the distribution. Indirect taxation shows little systematic deviation from proportionality. (See column (d) of table 1.)5 Colombia has a relatively advanced system of taxation in the sense that it relies more heavily on income taxes than do many developing countries.’ But it is not at all clear to what extent the corporation income tax adds to the progressivity of the tgu system. Under standard assumptions the tax would be imputed entirely to Colombian shareholders, and therefore be borne entirely by Colombians in the very highest income groups. But under other assumptions we would expect the corporation income tax to add little to thz progressivity of Colombian taxation, since it would be borne primarily by consumers or workers, to the extent that it is not exported to foreigners.f Thus virtually -‘Berry and Urrutia
(1976, pp. 31 36, 41 -52). to bc 0.57 and tile ‘Berry and Urrutia (1976, p. 40) calculate the Colombian Gini -jefficicnt -. estimates of Gini coefficients for the U.S. of 0.33 to 0.35. In developing countries the top decile of lhe distribution ordmarily accounts ror about 30 perzcnt of income. ‘Table 1 is drawn from McLure (1975, pp. 170, 18:) where incidence estimates based on three different assumptions about the incidence of the corporation Income tax and two about the inclusion of coffee export duties are presented. Table ! is based upon the assumptions (a) that coffee export duties should be :reated like all other taxes, even though some revenues finance benefits for the coffee sector, and (b) that corporation income taxes are shilted to Colombian consumers. The first assumption, while it greatly affects the estimates of effective tax rates in the rural sector, has relatively little impact on the estimates for the entire nation. The assumption of forward shifting of the corporate tax, on the other hand, produces estimates or Incidence patterns which are quote different from those based on the assumption of incidence entirely on Colombian shareholders, but not so different (except in level) from those babed 01: assumed incidence of the tax on labor or foreigners. As ahvals, we would expect the progressivity (:f indirect taxes (and other taxes shifted to consumers) to be understated to 5omc unknown extcul because of the inability to base effective tax rates on permanent income. instead of observed income. This would, however, seem to be of little relevance so far as the perconal in ome tax (and unshifted corporate tax) is concerne,!. Moreover, this consideration would probdhly little affect the basic conclusion ‘SIX Chelliah
about the distributiona’
effcfeztsof the tax refor;n of 1974.
et al. (1975).
‘While we would not assume that Colombia could export ils corporate incomc tax to forrlgn constimers, partial exporting to foreigners occurs via the foreign tax credits available in many home countries of firms investing in Coiombia. The argument that noncredited taxes are shifted to consumers is essen:ially that made in the next footnote for the property tax: international
McLure (1975, pp. 160-165, 170, 181). This table incorporates
0.0 0.2 0.3 1.2 1.9 2.6 13.2
1.9 7.8 21.8 48.7 71.8 89.7 100.0
19.0 39.2 64.1 87.0 95.8 99.2 100.0
6,OOu- 12,000 12,000- 24,000 24,000- 60,009 60,000-120,000 120,000-24QOOO Over 240,000
Annual household income (pesos1
Personal income tax (c)
Cumulative percentage of: Households Income (a) (b)
9.2 9.1 9.4 8.9 !$I!
11.3 7.9 9.5 10.3 11.3 11.5 24.0
Subtotal, columns c and d (2)
5.8 3.7 2.7 2.7 2.4 3.2 6.0
Coffee export duties and propertg tax (fi
17.1 11.5 12.2 13.1 13.6 14.7 29.9
Total, all taxes k)
(b) of the article with coffee export duties included.
Indirect taxes and corporate income tax (d)
Effective tax rates
Effective rates of taxation, all households (prior to ?he !974 reform.)
M. Gillis and C.E. M-Lure,
and income distribution
the only tax that we can be confident contributes significantly to the progressivity of tht3 Colombian tax system is the personal income tax.8 If the corporate tax is shifted or borne by foreigners, the personal tax accounts for almost all of the difference between the effective rates of tax (tax as a percentage of income) paid by the top four percent of the population and those paid in lower income brackets. (Compare columns (c) and (e) of table l.)g Not only is the personal income tax the chief instrument of progressive taxation in Colombia; it is paid almost entirely by the upper income classes. Talus the top four percent of the distribution. pays two-thirds of all personal income tax and the top thirteen percent pays five-sixths. (See the information in table I.) This being the case, it is inevitable that reform of the income tax affects primarily those at the upper end of the distribution and has little effect at the bottom. Tax reform in other areas can, of course, reduce burdens at the bottom of the distribution. But this seems rarely to be the objertive of reform, perhaps because of recognition that, apart from limiting indirect taxes on certain fc odstuffs, it is difficult to provide more than minimal refief for poor persons by adjusting exemptions, rate structure, etc. of indirect taxes and that what gain is achieved may be bought at the price of considerable costs of administration and compliance. Rather, policy mzkers, including especially those in Colombia, seem to accept the dictum of Bird and DeWulf (1973, p. 673) that. ‘if the principal aim of redistributive policy is to level upto make the poor better off- the main role that the tax syst-m has to play is . . . the limited ar:d essentially negative one of not makinp them (the poor) poorer’. In so doing, they are perhaps recognizing that, in the words of the Musgrave report, ‘. . . it is difficult to bring about a major shift in lax burden or income distribution. The primary emphasis in any such efrort must lie on
mobili:y d capital precludes the burden of the tax falling on shareholders. While it is unfortunate that the uscertainty about the incidence of the corporate income tax clouds any discussion of ta’c policy-perhaps even more than it does in advanced countries-it seems inappropriate to devote more of the present paper to this issue, for lo do so would eniail giving short shrift to the primary topic of the paper. ‘Property taxes may also not have the markedly progressive effect assumed in McLure (1975, pp. 170-71). In a nutshell, the reasoning is that capital is suRicienlly mobile in and out of Colombia that investr.irnt would occur there only if a return eqlJal to that available elsewhere were available; that IS, if the tax could be shifted. For an extended discussion of this issue, see McLure (forthcoming). ‘Coffee export duties (2.6 percent of income) and property taxes (0.6 percent of income) are combined in column (f) after the sub-total of other taxes because of uncertainty as to whether the first should Je included on equal terms with other taxes and because property taxes probably should not be assumed to fall entirely on capitalists, as noted in the previous footnote. Omitting export duties and allocating property taxes in proportion to consumption or to land and labor wou!r! produce a pattern of effective rates similar to that shown in column le).
M. Gillis and C.8. McLure, Taxation and income distribution
the expenditure side of budget developtnent planning.“’
3. The reform outlined Space does not allow a full description of the many important provisions included in the recent reforms. But the most important c&n be summarized as follow0 : 4.1. Persona1 income and related taxes (1) Revision of marginal tax rates under both the income tax and net wealth tax; (2) Conversion of most income tax deductions and sper’sl exemptions to tax credits; (3) Elimination of almost all exemptions for particular kinds of income and net wealth; (4) Introduction of a presumption that income is at least eight percent of the ta.xpayer’s net wealth; (5) Taxation as ‘occasional income’ of such items as realized long-term capital ga.ins, inheritaalces, gifts, and gambiing receipts; and (6) Modifications in tax administration. 3.2. Company income taxes (1) Elimination of graduated rates on companies; (2) Consolidation of the three previous classes of companies (corporations, limited-liability companies, ard partnerships) into two classes, and abolition of a number of complementary income-based taxes, the combination of which often produced steep!y progressive rates: (3) Elimination of most exemptions and other tax incentives; (4) Taxation of all capital gains as ordinary income: (5) Replacement of many deductions by credits; (6) Adoption of the presumption that income is at least eight percent of net wealth of corporations; (7) Liberalization and clarification of existing depreciation allowances, while retaining historical cost as the basis for depreciation; (8) Restriction of the use of percentage depletion for mineral assets; (9) Introduction of loss carryforward, primarily for corporations; and (10) Inclusion of many st.ate-owned enterprises in the income tax base.
3.3. Internal indirect taxes (1) Sales tax : (a) Increased emphasis upon the value-added
“Musgrave and Gillis (1971, p. 33). Estimates of the distributional impact of public expenditures in Colombia are summarized in Berry and Urrutia (1976, chapter 7). Such estimates have, however, been criticized strongly in DeWulf (1975, pp. 61-131).
M. Gillis and C.E. McLurc, Taxation and
income di: tribuiion
Increased rates, especially on items judged to be consumed especially heavily by upper income groups, including personal and automotive services; (c) Expansion of the category of previously exempt goods, especially food; (d) Adoption of Andean Common Market variation of the Brussels Tariff Nomenclature (WN) to specify rates for narrowly defined categories of commodities; (2) Other internal indirect taxes: (a) Elimination of several important stamp taxes; and (b) Provision for increased Laring of revenues with subnational governments. 3.4. ForLIgn
(1) A slight incr>ase in import surcharges; (2) Subjection to import duties of the imports of government enterprises and decentralized agencies; (3) A reduction in the coffee export tax: and (4) Reduction in the rates at which certificate, of tax credit (CATS) are ?aid on nontraditional exports. Those provisions having an identifiable impact on income redistribution are explained more fully and evaluated in the following sections.
4. Evaluation For the most part it is convenient to discu;:s the various income tax reforms without regard to whether they are found in the personal or company taxes. Further. given the brief period that has elapsed since the reform was undertaken -and the subsequent paucity of hard datl for gauging changes in tax liability by income class- we can provide little more than qualitative Judgments about the likely distributive effects of the reform. However, where possible we cite results of simulations of distributional effects, done in 1975 by the Ministry of Finance, for several of the income tax components of the package. 4.1. The indiGduu1 4. I. I. Conwrsion
Tax deductions and credits have the well known characteristics that the value of a deduction, and therefore its stimulative effect, dcpcrds upon the marginal tax rate of the taxpayer, whereas the value of a credit is the same for everyone. Par.tly because of the distributional consequences of this difference, the Musgrave Commission had earlier proposed (1971), and Colombia adopted in part. a system of vanishing deductions-deductions that would diminish as income rose above selected levels. But the consequences of vanishing deductions for both equity and neutrality are rather capricious.
hf. Gillis and C.E. McLure, Taxation and
and experience in Colombia, albeit brief, has demonstrated that this device is complicated and particularly subject to erosion. For these-or other -reasons, many allowances that were deductions or vanishing deductions under the old personal income tax law were converted to credits in the 1974 reform. In particular, for individuals there is now (1) a personal credit of Col. $1,00011 for the taxpayer (and for the spouse, in the case of a joint return) and a credit of Col. $500 for each additional dependent, (2) a ‘special’ credit of the greater of $1,000 or the sum of ten percent of medical expenses and expenses for primary and secondary education plus twenty percent of the first Col. $10,000 of rent and five percent of the excess over that, (3) in the case of joint returns, an additional special cred.it of Col. $1,000 for the spouse, (4) a credit of twenty percent of contributioz (40 percent for corporations) to charity and governmental units (up to 2G percent of gross tax liability before credits), (5) a credit oi twe,. *I percent of the first Col. $40,000 of earnings received from savings accounts, dividends, and mutual funds by taxpayers with net wealth of no more than Cal. %2,@00,000,plus several other credits of less immediate interest. Because many, but not all, of the old deductions were included in the vanishing formula, assessments of the redistributive impact of the switch from deductions to credits cannot be made with any degree of precision. However, it seems that the change is likely to increase progressivity. First, the commonly applicable twenty percent credit is equivalent (in terms of tax savings) to a 100 percent deduction for someone in the *twenty percen: marginal tax bracket and worth more than a deduction to a taxpayer it 1 lower marginal tax bracket. Based on fragmentary information for 1968 and 1971, it appears that fifty percent or more of taxpayers may have paid marginal rates of less than twenty percent. At the other end of the scale, we might note that under the old law only one-ha!f of certain deductions vanished, and at a rate of twenty percent of the excess of income over Col. $60,000. Therefore, taxpayers in marginal rate brackets above forty percent could be no better off under the new regime than under the old. T$us, the shift from deductions to credits should increase progressivity, at least among taxpayers at the ends of the inctime distribution. 4.1.2. Changes in rate structure It is difficult to compare the old and new rate structures under the individual inrome tax, due to the shift from deductions to credits. But it does appear that at least in the upper income levels progressivity has been inNcreased. Similarly, the rate structure under the net wealth tax now adds somewhat more to progre;siviry, due primarily to the increase in the
exchange rate was Col. $33.04= U.S. $1.00 in January of 1976.
Gillis and C.E. McLure,
and income distribution
exemption from Cal. $20,000 to Col. $80,000 and the increase in the ma.ximum rate from 1.5 to 2.0 percent. Indicatii-e of the distributional effects of the shift to tax credits and of rate adjustments under the income tax is table 2, which presents the results of
Table 2 Imphcations of some key aspects of the 1974 tax reform (results of simulations comparing 1973 liabilities under pre-reform and post-reform Income tax law). _ -_Percentage change in income tax liabilities, by net income class, due to:
o- 6.000 6- 12,ooo 12-- 24.000 24- 50.000 50- 75,ooo 75.- 1oo.o!lo 100-150.000 1%200,ooo 20%300,000 3oGloo,ooO Over 400.000 ___Source:
Rate modilications and substitution of tax credits for deductions and exemptions (a)
Effects of eliminaG-q exemptions for capital income (b)
-100 -100 -96 -59 -37 -39 -36 -26 -16 -6 +8
-100 - 100 -96 -55 -21 -21 +lS +2 +25 +44 +186
.- 22 ___-__
Obregon and Ferry (1975, tables B. I -8.6).
simulations of the new income tax (a) shows that the modifications themselves, would have freed the all income tax liability and wouid highest income class (Col. $400 proportion to income.
law by the Ministry of Finance.‘” Column in tax rates and the shifts to credits, by three lowest income classes from virtually have reduced tax liabilities for all but the thousand and above), but in declining
“See Obregon and Perry (1975). The simulations arc based on data on the distribution of income and collections of income and complementary taxes for personal income taxpayers for 1973 (the relevant distributions for 1974 being unavailable). The simulations involved application of the post-reform income tax la~ to 1973 taxpayer liquidations. The revenue thus generated in each income class was then compared to actual tax collections, by income class, for 1973.
M. Gillis and C.E. McLure, Taxation and incnme distribution
income and wealth
The old law exempted from taxation such items as interest on mortgages and government securities, monetary corrections on bank deposits .of constant purchasing power (UPACs), and income placed in so-called extraordinary reserves, and provided incentives for private electric companies and manufacturers and assemblers of automobiles. These exemptions were probably of benefit. primarily to upper income groups and virtuaily none of them made sense as a matter of public policy; they had therefore generally been attacked by the Musgrave Commission. Thus, their elimination should be judged a step forward on grounds of resource allocation, revenue yield, and distribution. Similar statements could be made about exemptions for labor income, which probably also benefit primarily workers in the top part of the income distribution. But the Constitution prevents using emergency powers to reduce gains made previously by labor. These exemptions are, of course, a ,suitable target for further reforms. Running parallel to the exemptions under the income tax were a large number of exemptions under the net wealth tax. Virtually all of these-for such items as jewelry, works of art, government securities and mortgages, limited amounts of shares in Colombian corporations, and investments in non-profit enterprises, specified tax-exempt entities, assets yielding tax-exempt income, assets not capable of yielding income, and activities judged to be especially important for economic development -have been eliminated, leaving only a small number of relatively minor types of exempt wealth. Eliminating these exemptions should increase progressivity. Data concerning the effects of elimination of exemptions for non-capital income and exemptions under the net wealth tax were not available. However, the Ministry of Finance did undertake simulations of the effect of abolition of exemptions for capital income. As table 2 indicates, this measure was decidedly progressive in its distributional impact. Columb (b), which incorporates the effects of this change with those arising from rate modifications and the shift to tax credits, shows that the abolition of exemptions for capital income served to increase liabilities for all income classes except the three lowest (compare columns (a) and (b)), with the largest increase coming in the top two categories (Col. $300 thousand and above). 4.1.4. Presumptice
Due to the ease of hiding income, many wealthy Colombians, particularly those in agriculture, pay little or no income taxes. The presumptive income tax is: the!-t+re, potentially a quite important component of the new tax law, especially so far as yield and equity are concerned. But the law provides several escape clauses - including natural disaster - through which the taxpayer can circumvent the tax. This is especially noteworthy since a similar
M. Gillis and C.E. MrL we, Taxation and income distribution
law passed in 1973, but applicable only to agriculture, was rendered impotent by the declaration that the entire country was subject to natural disaster! The measure clearly had a progrziive impact, if presumptive tax liabilities are measured against net wealth. The aforementioned Ministry of Finance simt;lations of the 1974 reform on 1973 data show that the presumptive tax should have led to substantial iricreases in income tax liabilities for all taxpayers with net wealth in excess of Col. $240 thousand. For taxpayers in the bracket just above this amount, taxes under the new law were two percent higher than under the 1973 law. The differential rises steadily to 79 percent for taxpayers with net wealth of from Cal. $2 million to Co]. $3 million, and to 191 percent for taxpayers having between Col. $10 million and Cal. $100 million in net wealth [see Obregon and Perry (1975)]. 4.1 S. Irregl&r
Under the new law so-called irregular income of individuals is taxed in a novel and innovative way that provides for some averaging and some preferential treatment. Though several other items are included in irregular income, the most important components of the tax base, especially from the distributional viewpoint, are long-term capital gains and eighty percent of net inheritances and gifts. Twenty percent of net irregular income is added to ordinary income. The marginal rate that would be applied to the sum is reduced by ten percentage points and applied to all irregular income. Although the method adopted for achieving heavier taxation of capital gains is similar to one of the alternative approaches endtiised by the Musgrave Commission, l3 the new tax on irregular income goes well beyond the Musgrave Commission propo:ials in that it also applies to eighty percent of net inheritances and gifts. (The mandate of the Musgravp Commission did not include death and gift taxes.) Given the heavy concentration of share ownership and other types of assets (particularly real property) in Colombia [Quale (1971, p. 355) and Herschei (1971, p. 38811, t!le new tax on irregular income could add significantly to the progressiviiy of the tax system, not onl;l because it provides for heavier taxation of gains on property held for ,nore than tNo years, but because it may repres *c-qt a substantiaily simpler and more effective method for taxing enrichmer L rising from bequests and gifts than was formerly the case under the cony;.Icg maze of estate and gift taxes in force prior to 1974. Prior to the reform, the only capital gains th:lt were subject to any tax at illcolne not only all were those on real estate. The new lax on irregular brought other forms of capital gains into the L:w for the first time, but resulted in less favorable treatment for gains on all forms of real estate other ‘%e
and Gillis (1971. pp. 176 197)
M. Gillis and C.E. Mcture,
Taxation and income distribution
than owner-occupied housing. Given the nature of the inflationary process in Colombia over the past decade,14 together with the vocal resistance of the wealthy to capital gains taxation of any kind since the levy was first proposed there in 1923, it was perhaps inevitable that the new tax treatment of capital gains would incorporate some adjustment of gains arising from inflation. The method chosen was similar to alternatives proposed by the Musgrave Commission, in that taxpayers were allowed both to revalue assets to their values at the end of 1974, and to take an optional annual revaluation of eight percent.’ ’ However, because taxpayers must use the revalued asset figures for purposes of the net wealth tax and the presumptive income tax, revaluation becomes a mixed blessing for the taxpayer. Although revaluation reduces taxpayers’ future tax liabilities on irregular income realized in the future, it. tends to increase tax liabilities under other taxes that apply to asset value, whether directly, as in the case of the net wealth and presumptive inccme tax, or indirectly as under the gift and inheritance taxes. Because overall tax liabilities of upper income taxpayers opting for revaluation depend on the peculiarities of their particular circumstances (their marginal tax rates unde; the different taxes noted a,bove, the expected holding period, etc.) generalizations about the distributional consequences of allowing revaluation are not easy to make.16 However, one other feature of the new law does serve to moderate the redistributional impact of the tax on irregular income, to an extent that may not have been intended by the architects of reform. Contrary to the prt>posals of the Musgrave Commission [see Musgrave and Gillis, (1971, p. SO)] the new law does not provide for full taxation of unrealized capital gains (constructive reahzation) at the time of death. T’his important omission could not only undermine the Fotential progressivity of the new law, but could increase the ‘lock-in effect’ for investors who own appreciated assets.” 4.2. The company income tax 4.2.1. Rasic taxation of companies Unification of the taxation of company income should have positive effects on income distribution, as well as its more obvious salutory effects on neutrality. IUnder the old law business forms were ma::! dulated in order to “Annual rates of inflation averaged only seven percent per annum sharply accelerated to fourteen percent in 1971 and 1972, and averaged and 1974. “For details of these provisions, see Gillis and McLure (1977, chapter ‘“For an illustration of the effects of revaluation for taxpayers in stances, see Gillis and McLure (1977, chapter 2, table H-9). “For further dkcussion of this complex issue, see Gillis and McLure B 5).
from 1966 to 1970, but over 25 percent in 1973 2). widely different circum(i977, chapter 2, section
M. GiNis and C.E. MeLure, Taxation and income distribution
minimize the total amount of tax that would be paid on a given flow of income. The result was, of course, complicated business structures and little revenue. It seems likely that the new unified rate structure will be less subject to manipulation. Since the earlier advantages of manipulation accrued primarily to upper income groups, progressivity should be enhanced.” On the other hand, rates on corporations are somewhat lower in many cases than under the old regime, due to the elimination of many subsidiary taxes, and progressivity may have been reduced somewhat for that reason. But this change seems worthwhile, since these aggregate rates, which reached a maximum of seventy-two percent at the company level, may in many cases have done more to dull incentives, create inefficiency and induce evasion than to raise revenue. 4.2.2. Depreciation and depietion;Taxation
of state enterprises
The distributional ‘effects of loosening regulations for depreciation allowances and tightening those for depletion depend upon the extent to which the benefits and burdens are captured by (or shifted to) consumers, rather than accruing to owners of firms utili7+ tf,, dilowances. It seems reasonable to believe that competitive forces are sufficient that some, but not ail, of the benefits and burdens accrue to consullizrs. If this is rrjle, liberalizing depreciation allowances would benefit both industrialists - presumably upper i::conlt: individuals - and consumers of manufactured go&. Earlier work on t?x incidence in Colombia suggests that the consumer element of this tax reduction may be regressive at the bottom of the income scale (especially so ,;e concerned), but more or less proportionate far as poor rural farniL above that. The reduction ia the use of percentage depletion would have similar effects, but in the opposite direction. We have no estimate of the likely amounts of revenue involved in the changes in either depreciation or depletion allowances. But it seems that on balance these reforms should be given positive marks for their resource allocation effects and that taken together they are more or less neutral from a distributional point of view. Similar comments apply to the liberalization of provisions for loss carryforward, as well as to the inclusion of income of state-owned companies in the income tax base for the first time.” “Statements of this type are, to some extellt,!inconsistenlwith our previous suggestion thal the corporation income tax might be borne by consumers, rather than shareholders. Whcrclrs the business taxes might, on the average, be shifted to consumers. differentials above and below the average might affect profits. Space unfortul,ately does not allow the full discussion of this issue that would be desirable in another context. ‘“Decree 1979 of September 19 provided for the taxation of thrprofits of state industrial and commercial enterprises ot, the same basis as those of private corporations and for the abolition of virtually all customs exemptions enjoyed by most government agencies. Subsequent modific:dtions in the law served to dkipate most of these reforms. so far as customs duties are concerned. though the income t:\x provisions were allowed to stand essentially intact.
M. Cillis and C.E. Mchrc,
Taxation and income distribution
internal indirect taxes
The modifications in the sales tax were intended to increase the yield (and perhaps the elasticity), to increase the progressivity (or reduce the regressivity), and to improve the administration of the tax. The increased use of the value-added technique should improve both administration and revenue performance, because historically there has been a substantial amount of ‘slack’ in the administration of the tax, especially so far as firms allowed to purchase on a tax..exempt basis were concerned. To the extent that benefits of evasion have not been shared with consumers, improved administration should add somewhat to progressivity. ?rogressivity will also be enhanced somewhat by the increase in luxuy rates from twenty-five percent to thirty-five percent and the inclusion of a number of income-elastic services in the tax base for the first time. The reform measures involved substantial tax increases on a number of items important primarily in the budgets of high income families. Rates of tax on passenger automobiles, pleasure boats, jewelry, silks, and electric appliances such as hair dryers and razors were increased from twenty-five to thirty-five percent, while those applicable to most recreational equipment, stereos, movie equipment, cigarette lighters and pipes increased from lifteeL to thirty-five percent. Rates of tax on income-elastic services such as telephones, repairs, dues to social clubs, photo developing and parking lot services were increased from an effective rate of zero to fifteen percent in most cases. At the same time, items important in the budgets of low income families were classified in either the exempt or a new category of six percent. The exemption structure was expanded to include almost all foodstuffs, agricultural inputs and certain items of heavy machinery. Items taxable at six percent include such semi-necessities as clolhing, soap, shoes, and radios. According to estimates prepared within the Ministry of Finance, exempt items and those taxable at the lowest rate of six percent account for ninety percent of outlays of low and ‘medium-low’ income families. Goods that are not expressly exempt or subject to other sales tax rates are now taxed at the generally applicable rate of fifteen percent, a substantial increase over the generally applicable rate (four percent) in force prior to the reform. In general, the sales tax reform involved the placing of substantially higher rates of tax on items purchased primarily by upper income families. The net effect of expansion of the exemption structure and the increase of tax rates (from four to six percent) on so-called semi-necessities left the sales tax burden on lower income classes relatively unchanged. Previous studies of sales tax incidence in Colombia had concluded that even prior to the 1974 reform. the sales tax was basically progressive with respect to consumption expenditure, and somewhat progressive with respect to income [see Gillis and McLure (1977, chapter 411. If it can be assumed that the new sales tax
M. Gillis ad C.E. McLure, Taxation and inwrne distribution
will prove no more difficult to administerzO than was the previous system, the new measures should, on balance, strengthen the redistributive impact of the overall reform package. However, as in cases of sales tax reform undertaken elsewhere, expectations of massive income redistribution arising from heavier taxation of luxury consumption are not likely to be fulfilled, owing to the small share of such consumption in the total for the nation. But the 1974 sales tax measures did, in Colombia, result in very substantial revenues, and perhaps the real significance of this aspect of the reform lies in the greater scope it provided for meeting the governmeni’s expressed goals for redistribution on the expenditure side of the budget, in programs directed toward the poorest 50 percent of the population, such as rural development, nutrition and public health. Other aspects of the sales tax reform, including. the adoption of the BTN to classify goods for purposes of the sales tax, as well as other features of indirect tax reform (including the elimination of many nuisance stamp taxes) do not seem to have strong distributional implications, however advisable these neasures may have been on efficiency and administrative grounds. The same observation applies to the increased sharing of general sales tax revenues, except to the extent that it will increase the capacity for subnational governments to provide essential services to the poor within their jurisdictions. 4.4. Other measures The increased surcharge on imports should not seriously affect the aftertax distribution of income. On the other hand, the reduction in the export tax on coffee, from sixteen to twelve percent, will tend to improve the urban--rural distribution of income and alter the size distributions for both the entire country and the rural sector. Finally, it is difficult to know, without further study, the distributional eflects of the reduction in export‘ incentives provided through the CAT. But they are likely to be positive, since most workers (and, of course, most owners) in firms benefiting from CATS are probably relatively well-off. It should be remembered, however, that other policy adjustments [see Gillis and McLure (1977, pp. l&20, 1%157)] undertaken at the time of the tax reform may have left producers in export industries little affected by the reduction in CATS. These ndiustments involved exchange rate and financial policies, increased export promotion, and improved techniques of freeing exports from sales taxes. Orle measure conspicuously absent from the list of 1974 reforms was rationalization of the mix of tax and exchange rate policies affecting petroleum products, particularly gasoline. Because of exchange. rate subsidies, 20Experience in the first year following the enactment of the reform suggests that assumption is a reasonable one [see Gillis and McLure (1977. chapter 4)].
M. Gillis and C.E. McLure,
Taxation and income distribution
coupled with price controls on oil from old fields and low taxes pep gallon of fuel, Colombian gasoline sold for a pump price of about U.S. $0.11 in late 1974. Even if one were willing- and able- to ignore the rapidly worsening problems of pollution and congestion in the largest Colombian cities, one could hardly imagine a worse energy policy for the late 1970s. Though the adverse results of this policy are pervasive, for our purpose we need on’ly note that the subsidization of the consumption of gasoline benefited primarily upper-income groups, who own the vast majority of private automobiles, and that elimination of the subsidy would add to progressivity, especially if it could be combined with increased subsidization of urban bus transportation for the masses. However, the government finally took steps, in August of 1975, to end the exchange rate subsidy, on a gradual basis. By May 1976 the exchange rate for petroleum had been unified with the ‘certificate’ rate. Gasoline prices rose on this account, and because the taxes on fuel are ad valorzm taxes, collections from this source began to rise as well. Price contrcls on petroleum from older Co!ombian fields were retained, so that fuel prices in Colombia still did not approach opportunity costs, but even so, the price of gasoline at the pump had risen to U.S. $0.187 by May 1976. Elimillation of this last vestige of an outdated energy policy would serve to increase significantly revenues, neutrality, and the redistributive impact of the tax system. 5. Judicial and administrative aspects Economists who undertake to assess the impact-distributlve or otherwise- of a major policy change barely after the ink has dried $>n the enabling legislation or decrees do so at some peril. This is especially true in the case of tax reform, where successful implementation of administrative and procedural measures is crucial if legislative intent is to be realized. The effects of given structural changes in a tax system depend not only upon the efficacy of the tools and procedures available to tax administrators, but also upon morality in tax collection and compliance. Neither the commitment to improved administration nor the climate for taxpayer compliance is easy to gauge under any circumstances, least of all following a reform involving major departures from past policy; the reciprocal interactions between structural tax reform and tax administration and compliance become evident only with the passage of time. But the full effects of the Colombian reform began to be felt only in the second quarter of 1976, when taxpayers began filing income tax returns for 1975, the first completi: year in which the income tax measures were operative. The initial tax reforms announced in 1974 included a number of key changes in administrative procedures arid pract.ices for handling declaration
M. Gillis and C.E. McLure,
and income distribution
and assessment of income which, had they been allowed to stand, would have vastly strengthened the hand of the tax administration in dealing with various of the ingenious techniques of evasion developed by Colombian taxpayers over more than a century of experience with the income tax. However, many of these procedural changes were quickly ruled invalid by the Supreme Court and the Counci! cf State.21 Whether the climate for morality in tax administration and compliance has been altered since late 1974 is highly problematical. It is known that prior to the 1974 reform, administrator/ta:rpayer corruption was not uncommon.22 In addition, it might be claimed that !.he full implications of taxpayer resistance to the reform may not show up until 1977-1978 or beyond. Indeed, Vasquez and Palomeque (1976, p. 36) have argued that the 1974 reform, coming unexpectedly, caught taxpayers with their defenses down. But they also maintain that by 1976 the initiative had returnee! to the taxpayers, who by then had had time to establish new or modified ways to evade or avoid taxes, within the limits established by those administrative changes that did survive into that year. Lldeed, by 1976 several of the original architects of the reform had themselves expressed serious doubts as to whether much of the strong initial momentum of the reform so evident in 1975 could be sustained much beyond the short term. Their grounds for pessimism were two-fold, First, taxpayers had in fact apparently begun to deploy in 1976 new methods for exploiting loopholes and avenues of evasion. Second, there were signs that the govern:-.lent’s resolve for enforcing the measures as enacted in 1975 had begun to weaken: the government finally acceded to demands for income tax relief in 1976, both in the form of relaxed withholding requirements and tax reductions to compensate for inflation. While cognizant of the possible effects that adverse judicial rulings and deterioration in administration and compliance can have in thwarting the achievement of ambitious structural reform, the authors are not at all inclined to write off the ColombiaI, reform as merely an interesting b!.lt futile exercise in sensible fiscal engineering. To do so would be to ignore the feasibility. important implications of structural reform for administrative Unless the reform somehow served to cause a marked retrogresLIC ~hos:: that circumscribed the scope for legal artifices to delay or escape tax entirely. These arc outlined in Vasquez and Palomeque (1976, pp. 15 22). ‘“Vacquez and Palomcqc: (1976. p. 29) estimate that in recent ‘inducertlents’ pald to tax .w -ssors. the state lost 20 pesos in revenue.
years, for cvcry
Beyond the issues raised above, another important factor complicates efforts to assess the distributive impact of this or any tax reform. As noted, the 1974 Colombian reform was explicitly geared to achieve allocative, stabilization (in this case, revenue) and administrative, as well as distributive, goals. Fiscal reform that affects any of the first three goals will almost certainly have distributional consequences quite apart from those of measures adopted primarily for redistributive purposes. In some instances these consequences will support distributive aims; in others they will not. The task of sorting out the effects on income distribution of allocative (or revenue or administrative) measures is no less difficult than that of identifying the allocative impact of measures primarily intended as redistributive instruments. But to some extent it is possible to reconcile distribution, allocation, revejlue, and administration in the following ways. First, virtually any reform which increases the administrative feasibility and yield- of taxation has the potential of improving the lot of the poor, so long as a highly regressive tax is not involved. For one thing, if public funds are directed consciously toward providing services that benefit the poorest SO percent of households, as is claimed to be the official policy of the present Colombian Government, a larger budget may be better for the poor than a smaller budget. Beyond that, if inflation is both inevitab’le in the absence of higher taxes or lower expenditures and more burdensome to the poor than to other groups, as has been asserted to be the case in Colombia [see Gillis and McLure (1977, c. 1511, tighter tax administration will benefit especially the poor. Finally, it seems reasonable to believe, but hard to demonstrate, that in general a tax reform which improves the aliocation of resources, and therefore increases the real value of national output, is likely to improve the lot of the poor. This presumption is based upon tbe thought that tax policies that distort the allocation of resources did not just happen; more likely, they were enacted at the behest of taxpayers -often wealthy taxpayers. Since elimination of this kind of distortion can be expected to leave the preferentially treated taxpayer no better off than befoie, it probably benefits others, including the poor. But none of the points discussed in these two paragraphs should be pushed too far, nor are they the proper subject of this paper. Rather, suffice it to say that improved administration, increased yield (at least in the present context), and reduced distortion of resource alloca&ion are, like increased progressivity (again given existing incidence patterns), acceptable goals in their own right. 6. Indications of initial impact Whatever its ultimate distributive implications, the reform was clearly successful in augmenting government revenues in its first year of application
M. Gillis and C.E. McLure, Taxation and income distribution
(1973, even betore the full effects of income tax reform were felt. Gross tax collections rose by nearly 47 percent over 1974 levels, while net tax collectionsZ3 Increased by nearly 51 percent (see table 3). To be sure, a large part of the 1975 increase in revenues was not traceable to the reform as such, but to the built-in response of the tax system to changes in nominal income. 24 But collections in 1975 far outpaced the growth in nominal GDP, which is estimated to have risen by 27.3 percent over 1974 levels.‘5 Thus, the ex-post revenue elasticity of the tax system for 1975 was 1.71 on one view (gross collections) or 1.86 on another (net collections).26 The significance of these figures can best be appreciated by examining ex-post revenue elasticities over the period 1970 through 1974. Over that period tax revenues barely grew as fast as nominal GDP (even with frequent changes in tax legislation often designed to bolster revenues), as indicated by the overall ex-post elasticity, 1.01 [Gillis and McLure (1977, pp. 22-23)]. Further, revenue responsiveness to growth in GDP was highly variable over the period, with very low elasticity figures for 1972 and 1973 (0.80),2’ and moderately high ones for 1970 and 1971 (1.28 and 1.17 respectively). But in no year did ex-post elasticities approach those recorded for 1975. The surge in revenues in the first year following the 1974 reform had obvious consequences for the nation’s tax ratio, historically one of the lowest in Latin America.‘” For the period 1969-1974, the ratio of gross taxes to ZJThe distinction between gross and net tax collections is important in Colombia. Gross taxes refer to the collections imhsice of taxes paid with various types of tax credit instruments used in Colombia since 1967. The most important of these has been the CAT (Crrt$cado de Ahor Tributtirk~), an instrument issued to exporters of nontraditional products (typically equivalent to 15 percen! of export value, prior to Janu,.ry 1975). Nontaxable themselves, these instruments could thert be later used to meet a variety of tax obligations. L41t might be thought that the sharp rise in coffee prices and export values beginning in July 1975 (foRowing a disastrous freeze in Brazil) played a major role in the upsurge of tax collections for 1975. But closer inspection reveals that this was not the case. Revenues from colfee export taxes flow into a specud exchange account of the Central Bank. Computations based on pubhshed Central Bank information indicate that revenues from coffee export taxes in 1975 were about Col. $300 million, or 26.7 percent above 1974 levels. Thus, the rate of growth in coffee export tax revenues in the special exchange account was about half that for total revenues in 1975. “Based on Planeaci6n figures shl:wing 4.6 percent real growth and an increase of 21.7 percent in the implicit deflator for GDP. *‘F.x-post revenue elasticities reflect not only the built-in response of the system to income growth, but effects of changes in tax legislation (base and :ates) as well. An ex-ante elasticity coefficient would ircorporate only the effects of built-in revenue growth. Given the frequency with which Colombian tax legislation has been altered since 1969. computation of historic cxante elastcity coefftcients for this period would represent a nc:lrly impossible task. “The very low elasticity figures for 1972 and 19i.1 wcrc due to a variety of causes, includmg legislated tax reductions for business enterprises, turther proliferation of income tax exemptions. and the illcentives given in those yeat; by accelerating inflation to delay incclme lax r)aymcots until the last possible moment. *@SeeMusgracc and Gillis (1971, ch,tpter 3).
- .~.I. - _. _
_ _ _
22,172 5,410 3,561 1,594 7,758 1,726 77 1,658 390
18,315 18,063 34 - - 219 * -
advance payment of personal and business ta&es, and the net wealth tax. and for the benefit of the coffee fttzd. rates for Fondo Vial. development. CATS have been primarily used by exporters to pay income
J 6,035 4,849 3,054 805 3,945 1,536 59 1,791 -
11,260 10,725 30 380 12.5
“Includes personal income and business taxes, withholding and ‘Special surcharges on imports earmarked for export promotion ‘Includes 10 percent tax on gasoline in addition to ad valorem dTaxes paid with tax credit certificates for exports and tourist and complementary taxes, customs duties, and the sales tax. Note: Elements in columns may not add precisely to totals due Source: Direccicin General de Impuestos Nacionales.
21,675 (1,456) --
11,596 3,706 2,080 452 2,449 1,474 43 1,375 -
9,190 8,778 28 257 137
953 342 2,019 1,288 31 1,040 -
8,024 7,624 23 216 159
Total current revenues (gross basis) Total taxes paid by CATS
interest and penalties contractual rents
7,375 2,670 839 320 1,670 1,002 28 876 -
5,885 2,242 805 298 1,078 S30 25 607 _
ti7xe.s Customs duties Special exchange account Earmarked customs surcharges I” Sales tax Gasoline and diesel fuel taxesC Special 5 percent tourism tax Stamp taxesd National sales tax on departmental
1. 2. 3. 4. 5. 6. 7. 8.
6,956 6,492 24 254 176
5.959 5,624 13 174 149
current revenues, gross basis by type of revenue, inclusi~ve OF CAT, 1970-1975 (millions of pesos). _---. .._.-1970 1971 1972 1973 1974 1975
tu.ws Income and complemc;~tary taxes” Surcharge on property tax Death and gift taxes Cattle tax
1. 2. 3. 4.
Colombia - Central government .--___..-_^--Tax
M. Gillis and C.E. McLure,
Taxation and income distribution
GDP declined steadily 3om 9.8 percent in 1970 to 9.0 percent by 1974. But in the first year following the reform, the tax ratio rose to 10.5 percent, an increase of 1.5 percentage points, or by about 17 percent. Despite all the shortcomings involved in using the tax ratio as an indicator of fiscal performance,2g this increase is nevertheless of considerable significance: except in countries which, unlike Colombia, rely heavily on volatile taxes on natural resources (i.e. Zambia on copper and Indonesia on oil), such an increase in the tax ratio from one year to another is quite uncommon. Moreover, there is (admittedly crude) basis for stating that, at least for 1975, virtually the entire increase may have arisen from increased taxes on the richest 20 percent of the income distribution. In particular, table 4 suggests Table 4 Alternative
estimates of the share of 1975 GDP taken from the top quintile distribution by the 1974 tax reform.
(1) Overall increase in tax collections attributable to tax
of the income
Percent of tax increase paid by top quintile (Column 2 as percent of column l--a)
Percent of GDP shifted from top quintile (column 3 x column 1 b)
Alternative incidence assumptions
(a) Amount (COI. % millions)
(5) As percent of 1975 GDP
Revenue increase arising from top quinlile (Co). $ millions)
Maximum redistributive effecta
Intermediate redistributive effrct”
I .42 ‘,,
Low redistribulive effcvt
reform (4 T)
“Allocates 100 percent of /1I& and .rl7& and 80 percent of .47& to top quintile. 7;. 7;. and 7; represent. respectively, the personal income and wealth taxes. the corporate tax and the salts tax (see appendix). “Allocates 100 percent of AT,. 50 percent of ,~7;., and 80 percent of ,Ir, IO top quintile. ‘Allocates 100 percent of -17,. 25 percent of .i7;., and 60 pcrcwt of .17, to top quintilc. Snttrcr:
that under the most realistic assumptions (row II) as to the incidence of the 1974 measures, the reform may have shifted as much as 1.43 percent of 1975 GDP away from the top 20 percent of the income distribution. Under the 29See Musgrave and Gillis (1971. chapter 3. p. 23).
M. Gillis and C.E. McLure, Taxation and income distribution
most (least) progressive set of incidence assumptions, 1.74 percent (1.14 percent) was shifted from the top quintile. The extent to which this shift on the tax side of the budget was (or could be) translated into a net transfer in favor of less advantaged members of Colombian society depends upon one’s view of the incidence of public services in 1975, a question outside the scope of this paper.
7. Concluding remarks The design and initial execution of the Colombian tax reform of 1974 were impressive, in terms of equity, neutrality, and administrative efftciency. While the redistributive results of the reform may have fallen somewhat short of the expectations held by its architects, the tax reform did, at least in the first year, yield a significant increase in tax contributions from the topmost quintile of the income distribution. Notwithstanding the relatively minor structural faults cited in section 4.1 of this paper, the reforms of the income tax alone, if they prove able to withstand the pressures for retrenchment that inevitably follow a policy change of the magnitude implemented in 1974, would rank the recent Colombian reform among the most ambitious and comprehensive ever, in either developed or developing countries. In terms of scope, basic conception, and objectives of neutrality and equity, these reforms are perhaps most nearly analogous to those proposed by the Carter Commission in Canada in the mid-1960s. But whereas the Carter proposals; called by several authorities a ‘landmark’ in the history of fiscal reform, remain largely inoperative, the Colombian reforms, through the constitutional mechanism of emergency powers, were enacted into law. The inislal results of the reform, in terms of its implications for income distribution, improved resource allocation and revenues, were unquestionably encouraging for those who look to thoroughgoing tax reform as a powerful instrument for assisting in the restructuring of society and for allowing an economy more closely to fulfill its productive potential. Whether these initial results prove to be more than m +r transitory can be answered only from a much longer historical perspective than was possible in this paper. And skeptics on the efficacy of tax reform may be comforted to know that as early as 197-t, some of the earmarks of decay in effectiveness of the reform had indeed already begun t.o appear, as taxpapers began to deploy newly developed defense mechanisms against the new law, and to exploit more fully the procedural gaps in tax enforcement that the government triedunsucczssfully -to close at the outset of ths reform. Beyond these observations, the nature of the reform and the way it was achieved must leave the student of tax reform with mixed feelings. On the one hand, it is gratifying to learn that reform of this magnitude and
M. Gillis and C.E. McLure, Taxation and income distribution
excellr.nce is possible in a democracy. On the other han3, It is a bit depressing to contemplate that the reform could be achieved only through the exercise of the extraordinary emergency powers provided in the Colombian Constitution. One must still wonder whether there is hope for meaningful tax reform in other democracies withcjut resort to similar -. and of ien unavailable-emergency provisions. Appendix The share of’ 1975 GDP taken from hi+inc~tne reform
recipients by the I974
We know from section 2 that the personal income tax is paid almost entirely by the upper income classes, with the top 13 percent paying fivcsixths of the total. Given the configuration of the personal income tax element of the reiorm, it is fairly clear that rhe burden of any revenue increase arising from income tax reform wouid lie in the richest IO- 15 percent of the distribution. Furthermore, the pattern of the changes in the sales tax was such that it would be difficult to argue (assuming forward shifting) that any sigriificant proportion of the increased burden arose from the poorest 50- 70 percent of the income distribution. Further, we knvw from section 6 that the historical ex-post elasticity of revenue growth with respect to GDP over the period 1969-1974 was about unity. This information furnishes a crude basis for estimating the distributive impact of the 1974 reform. If tax revenues had grown in 1975 at the same relative rati: as in the five previous years (or in 1974, for that matter), they would have grown at about the same rate as &mi;ial GDP. That isi total 1975 tax revenues would have been as defined in (1) below: Tl =[l +c(I’~)]T~, where T = total tax revenues in year t, c = historical overall ex-post revenue elasticity I* rate of GDP growth, year r. ‘I =nominal
Given that 1974 revenues (r,) were Col. $28,480 million and J:= 1 and r, =27.3, then 1975 revenues (7’, ) would have been Cal. $36,255 million. However, overall actual revenue growth in 1975 exceeded that predicted in (1) by a substantial amount (AT). In fact, AT was Cal. $5,565 million, or about 20 percent of T, and abw: 1.5 percent of GDP. We could interpret
M. 6.. lis and C.E. McLure,
6 i7 as the unpredicted chiange in revenues due to a variety of possible attitudes, or factors, including the tax reform, changes in compliance improvements in collection efftciency. But changes in compliance attitudes and collection efficiency do not normally become manifest in a period of time as short as one year. It theref>re does not seem unreasonable, as a first approximation, to treat dT as the revenue increase resulting irom the 1974 reform. However, use of a global elasticity coefficient masks a number of important implications of the reform. Actual 1975 collections from three tax categories grew at a substantially faster rate than did overall tax revenues. The three categories were (a) the personal income and net wealth taxes, (b) the company income tax, and (c) the sales tax. If collections from these three tax sources had grown at rates commensurate with their histovical ex-post elasticities, we would have:
s,=Cl +M-l)1SO~ where I, = personal income and net wealth tax collection in year t, C, = company income tax collections in year t, S, = sales tax collections in year t. E,, Ed and sS are the respective historic ex-post revenue elasticities for I, C, and S.‘*’ Given r 1 = 27.3, and I, =Col. $5,363 million, C,, =Col. $5,362 million,
E, = 0.91, cc = 0.9 1,
So =Col. $3,945 million,
sS = 1.46,
then I, =Col. $6,695 mi!iion, C, =Col. $6,694 million, S, =Col. $5,517 million. ‘Colombian treasury data for income taxes do not distinguish between personal and company income tax collections. However, previous studies have found that the corporation income tax typically accounts for about one-half of total income tax revenues. [See McLure (1975) Musgra\*eand Gihii (1971, table 11-l)]. A recent unpublished analysis of the reform by a former Director of National Taxes confirms this ratio. ‘Values of E,. Ed and ss are as computed for 1964-74 in Gillis and McLure (1977, chapter I, table I-3).
M. Gillis and C.E. McLure,
and income distribution
However, revenues from all three- sources grew by considerably more than would have been pred ic:ed on the oasis of historical coefficients of revenue elasticity for each tax. If we identify AT,, AT, and AT, as the unpredicted change in personal and corporate taxes and the sales tax, respectively, we have : A T, = Cal. $2,337 million, d Tc = Col. $2,336 million, d & =Col. $2,241 million. The sum of AT,, AT, and A G was Cal. $6,914 million, or about 124 percent of 471 the overall increase in tax revenues over and above that which would likely have materialized in the absence of reform. That is to say, expanded collections from these three sources alone accounted for at least the entire increase of 1.5 percentage points in the overall ratio of taxes to GDP. On the basis of information presented in the text and at the beginning of the appendix, it would appear legitimate to assume that 100 percen: of the excess growth in personal income tax revenues (AT,) came from the top quintile, if not the top decide, of the income distribution. We have assumed that only one-half the growth in total income tax collections was due to personal income taxes. The procedure employed here may well understate the redistributive impact of the income tax reform, since, given the nature of the reform, it is not unlikely that more than one-half the extraordinary growth in income tax revenues occurred in the personal income tax. Other studies [McLure (19751-J suggest that even if the burden of corporation income tax is shifted to Colombian consumers in the long run, instead of being borne by capital owners, those in the top 13 percent pay roughly half of it. In the short run, the percentage paid by the top 20 percent is probably even higher. even though some allowance must be made for some short-run shifting to nonresident owners of capital. In any case, a reasonable assumption is that at least half of LIT,, or Col. $1,168 million, should be allocated to the top quintile. Given the pattern of changes in the sales tax described in the text, it appears legitimate to assume that at least 80 percent (Col. $1,793 million) of AT, was attributable to the top 20 percent of the income distribution. On this basis, we have a total of Col. $5,298 million arising from the tax reform and attributable to the top quint&:. Since the overall tax system generated (in 1975) only Col. $5,565 million over and above that expecled given historical elasticity of collections with respect to GDP, then higher taxes on the top 20 percent of the income distribution accounted for 9.5 percent of total revenue growth from the entire reform package. Thus, the reform may have served to shift a minimum of 1.43 percentage points of GDP (0.95 x 1.5) away from the top quintile of the income distribution. A
M. Gillis and C.E. McLure, Taxation and income distribution
brief sketch of plausible variations on the assumptions employed in arriving at this judgement has been provided in table 4. Even under the least yxogressive set of assu.mptions, the top quintile of the population would bear at least 76 percent of the burden of the reform (see table 4). References Berry, Albert and Miguel Urrutia, 1976, Income distribution in Colombia (Yale University Press, New Haven), 31-36,40-52, chapter 7. Bird, Richard M. and Luc DeWuif, 1973, Tax incidence and inco,ne distribution in Latin America: A critical review of empirical evidence, International Monetary Fund Staff Papers 20,672-673. Chelliah, Raja J., Hessel J. Bass and Margaret R. Kelly, 1975, Tax ratios and tax eiTort in developing countries, International Monetary Fund Staff Papers, 22, 196-197. DeWulf, Luc, 1975, Fiscal incidence studies in developing countries: Survey and critique, International Monetary Fund Staff Papers, 22, 61-131. Gillis, Malcolm and Charles E. McLure Jr., 1977, La reforma tributaria Colombiana de 1974 (Banco Popular, Bogota), chapters 1, 2, 3,4. i-lerschel, Fed&co J., 1971, Taxation of agriculture and hard to tax groups, in: Richard A. Musgrave and Malcolm Gillis, Fiscal reform for Colombia (International Tax Program, Cambridge). 388-390. McLure, Charles E., Jr., 1975, The incidence of Colombian taxes, Journal of Economic Development and Cultural Change 24, 160-165. McLure. Charles E., Jr., 1976, The rdevance of the new view of the incidence of the property tax in developing countries, presented at the 15th annual conference of the Committee on Taxation, Resources, and Economic Developme.rt (Cambridge, MA), 22-24, Oct., forthcoming. Musgrave, Richard A. and Malcolm Gillis, 1971, Fiscal reform for Colombia (International Tax Program, Czmbridge). Obregan Sanin, Ivan and Guiller no Perry Rubio, 1975, Efectos redistributivos de la reforma tributaria de 1975 en Colombia, paper presented at World Bank Seminar on distribution of income (Bogota), 1-15. Quale, Andrew C.. 14ii, The capital gains tax, in: Richard A. Musgrave and Malcolm Gillis, Fiscal reform for Colombia (International Tax Program, Cambridge), 351-364. Vasqu~z Caro, Jaime and Gabriel Palomeque, 1976, Reforma tributaria: Brechas entre propositos, normas y reahdades, paper presented at Third Seminar on the Colombian Economy (Bogota), 15-22.