a. Sources of the fiscal imbalance in Latin America

(3) Domestic developments

In addition to these external developments, some domestic developments that to be mentioned. Faced with cheap foreign credit, rising prices for their basic exports, and pressing needs that could be met through higher public spending, governments found it difficult to worry about the time when (i) real interest rates might rise sharply, (ii) the terms of trade improvements might reverse themselves, and (iii) creditors would be unwilling to extend additional loans.

Cheap foreign credit also probably produced a lowering of the standards for public project selection. Although it is difficult to prove it, much of the public sector investment made in the 1975-82 period was probably of highly questionable quality. As a consequence, the capital-output ratio increased considerably during this period. This increase in the capital-output ratio reinforced a low investment rate to reduce the growth of aggregate output in the 1980s (see Table 2 and Appendix Table 10).

Table 2.

Structure of Domestic Demand in Latin America, 1965 and 1986 ^1/

Source: The World Bank, World Development Report, Washington, D.C., 1988.

^1/See Appendix Table 10 for country grouping.

Table 2.

Structure of Domestic Demand in Latin America, 1965 and 1986 ^1/

	Government Consumption		Private Consumption		Domestic Investment		Domestic Savings
	1965	1986	1965	1986	1965	1986	1965	1986
Average	10	14	69	73	18	16	17	13
Lower income countries	10	14	75	77	18	17	15	13
Upper income countries	11	14	67	69	19	15	22	19

Source: The World Bank, World Development Report, Washington, D.C., 1988.

^1/See Appendix Table 10 for country grouping.

View Table

Table 2.

Structure of Domestic Demand in Latin America, 1965 and 1986 ^1/

	Government Consumption		Private Consumption		Domestic Investment		Domestic Savings
	1965	1986	1965	1986	1965	1986	1965	1986
Average	10	14	69	73	18	16	17	13
Lower income countries	10	14	75	77	18	17	15	13
Upper income countries	11	14	67	69	19	15	22	19

Source: The World Bank, World Development Report, Washington, D.C., 1988.

^1/See Appendix Table 10 for country grouping.

The debt crisis brought with it, in several countries, an acceleration of inflation and, as a consequence, a reduction in real tax revenue, thus increasing the fiscal deficit. Stabilization policy became more difficult because of the debt overhang. The need to generate a trade surplus to meet the foreign obligations often required sharp devaluations. The policymakers were often unable to accompany these devaluations with sufficiently restrictive monetary and fiscal policy to prevent inflation.

b. Policy measures to reduce the fiscal imbalance

While the link between the debt crisis and the fiscal problems in many Latin American countries has been discussed in a variety of studies, some salient features might be mentioned. The external debt was accumulated in the process of financing (i) increases in domestic consumption, (ii) some good investments, (iii) some bad investments, and (iv) capital outflows. The capital outflow was stimulated by the belief that, given current and future domestic policies, the expected rate of return on domestically invested saving would be very low or negative. ^10/

Of all these uses of the resources obtained through foreign borrowing, only the investment in efficient projects could be expected to generate a future stream of income, which would contribute to servicing the debt. The flight capital is a potentially important asset, which in time and under particular circumstances could make a substantive contribution to stabilization and growth for the Latin American countries. However, as long as the economic and political situation remains uncertain, and as long as inefficiencies remain, this capital is unlikely to return.

The part of the foreign debt which went into either straight consumption or bad investment projects will not generate the income for servicing the foreign debt. Thus, with the current low economic growth, particularly when accrued interest on foreign debt is capitalized, the book value of the foreign debt, unless it is reduced in some way, will tend to grow at a faster pace than the economy.

The resolution of the debt crisis has many facets. In this section, however, we focus on the debt servicing aspect. The servicing of the public debt, if it were to occur in a stable macroeconomic environment, would require two types of transfer: (i) the releasing of resources by the private sector to the public sector, and (ii) a transfer of the resources by the public sector to creditors. In other words, there must be a net positive change in the fiscal balance (i.e., a net transfer from the private to the public sector) which must accompany the external adjustment required to make the foreign payment (i.e., the trade surplus). ^11/ Putting it differently the dollars associated with the trade surplus must be bought by the public sector so that payments to foreigners can be made. The transfer from the private to the public sector can occur through (i) higher taxation, (ii) a reduction in public expenditure, (iii) voluntary lending by individuals to the state, and (iv) inflationary finance. All of these alternatives have costs, and the minimization of these costs (especially the inflationary cost) must be one of the guiding principles of fiscal policy.

An increase in taxation would force a reduction in private spending, hopefully more in consumption than in investment. ^12/ Thus, it is important that the choice of tax measures should aim at promoting a reduction in the least useful kind of spending, that is, unproductive consumption and unproductive investment. Particular care must be taken to avoid taxes with strong disincentive or distortionary effects, such as export taxes.

A reduction in public spending must be aimed at unproductive public expenditures and must try to safeguard the standard of living of the poorest sectors. Unfortunately, cuts in public spending have often been made in categories that may be highly productive, such as operations and maintenance outlays and efficient investments, but not in politically popular but unproductive activities. ^13/

Voluntary lending would enable the government to sell bonds to the private sector, while the private sector would reduce its own spending sufficiently to generate a surplus that can be transferred to the government, which in turn would make the foreign payments. An increased sale of bonds by the government requires an increase in the private sector’s saving rate, or a reduction in private investment or in net capital flight. It thus requires a rate of interest high enough in real terms to induce an increase in financial saving and/or a reduction in private investment. The studies available on the interest elasticity of private savings are inconclusive. This road may require very large percentage increases in real interest rate to induce modest increases in the saving rate. These studies imply that one should not pursue a high interest rate policy under the assumption that it can stimulate a higher rate of savings. However, to the extent that high interest rates encourage financial savings to remain within the country, rather than to be channeled abroad, and to the extent that high interest rates discourage unproductive private investment, such as luxury housing, this policy can still be justified for other reasons. It is important, however, to keep in mind that there are long-term costs to domestic debt accumulation. To the extent that the servicing of foreign debt results in domestic debt accumulation and that domestic debt accumulation requires (and results in) high real rates of interest, this policy will imply an increase in future government spending to service domestic debt. It will, at the same time, result in a reduction in private investment which may contribute to a slowdown of economic growth. When a government relies on domestic borrowing to finance its expenditure, particular attention must be paid to the taxation of financial transactions. If the taxes on financial transactions are increased at the same time the government is trying to absorb resources from the private sector through voluntary means, the cost of this policy will be much higher.

Money creation, in a growing economy, would enable the government to finance some public spending without aggravating inflation. However, in stagnant economies with acute inflationary expectations, inflationary finance is unlikely to give the government significant net control over resources. By stimulating inflation, it will induce (a) more capital flight, (b) dollarization of the economy, (c) losses in tax revenue because of collection lags, and (d) efficiency costs to the economy in general.

The importance of simple and efficient tax systems and of well-developed budgetary controls cannot be exaggerated in these situations. It is a mistake to look for perfection in the tax system or to ask the tax system to do much more than raise revenue in a relatively neutral or broadly equitable way. Highly progressive income taxes can virtually never be implemented in high inflation. Highly differentiated value-added taxes are often accompanied by major administrative difficulties. In these circumstances, the golden rule is that of relatively low and undifferentiated rates on as broad tax bases as possible. The more governments insist on promoting high progressivity and on maintaining tax incentives to particular activities, the lower is likely to be the tax level and the more distorted is likely to be the effective tax system. It is also important that the government be in a position to monitor, precisely and in a timely fashion, the process of expenditure, not just in the central government, but also in the rest of the public sector.

Unfortunately, when inflation accelerates and macroeconomic difficulties become prevalent, the monitoring capacity of the government deteriorates, and it begins to lose sight of the true dimension of the fiscal problem. The measurement of the fiscal deficit itself gets distorted, either because the deficit is pushed away from where it can be easily measured to where it becomes difficult to measure, ^14/ or because inflation itself distorts certain expenditure, such as domestic interest payments, and ends up giving a picture that is no longer a reliable reflection of reality.

In the final analysis, there are no magic formulas for achieving stabilization through fiscal policy. Given the level of expenditure, the total revenue needed to meet that expenditure must be mobilized without causing inflation or distortions. If this is not feasible, there is no alternative but to reduce public spending.

a. Evolution in the 1960s and the 1970s

Table 3 is based on some of the data used in Tanzi (1974) and more recent data from other sources (Part A). ^19/ Data are available for the early 1960s, the early 1970s, and the early 1980s for only seven countries. The sample countries for these three time periods are not identical, thus making it difficult to discuss the evolution of the pattern of income distribution (Appendix Table 11). However, for any sample period, the income shares of the bottom 20 percent of the population were notably small for all the sample countries. As in the early 1980s, the share of income received by the bottom 20 percent of the population ranged between 2 percent and 6 percent for the seven sample countries; the average was 5 percent. Although the average of 5 percent was higher than 3 percent in the early 1970s and 4 percent in the 1960s, it is unclear that given the changes in the composition of the sample over the years, the small sample size, and the diversity of data sources, the higher income share in the early 1980s for the poorest 20 percent should be interpreted as a genuine improvement of their relative income share. Nor do the data tell about the most recent situation of the poor, although their situation appears to have deteriorated since the early 1980s (see below). The situation in lower-income countries does not appear to differ from that in upper-income countries. ^20/

Table 3.

Income Distribution and Poverty in Latin America, 1960s Through 1980s

Sources: Part A: Vito Tanzi, “Redistributing Income Through the Budget in Latin America,” Banca Nazionale del Lavoro Quarterly Review, Vol. XXVII, No. 108, 1974, pp. 5-27; United Nations, National Accounts Statistics: Compendium of Income Distribution Statistics, Statistical Papers, Series M., No. 79; and The World Bank, World Development Report, Washington, D.C., 1988; Part B: International Labor Organization, Meeting the Social Debt, Geneva, Switzerland, 1988.

^1/See Appendix Table 9 for country grouping.

Table 3.

Income Distribution and Poverty in Latin America, 1960s Through 1980s

Part A. Evolution of the Size Distribution of Income1/
	Early 1960s		Early 1970s		Early 1980s
	Lowest 20 percent	Highest 20 percent	Lowest 20 percent	Highest 20 percent	Lowest 20 percent	Highest 20 percent
Average	4	59	3	57	5	48
Lower income countries	4	62	4	57	5	47
Upper income countries	4	56	3	57	4	49

Part B. Evolution of Poverty
		1960	1970	1980			1985
				Total	Urban	Rural	Total	Urban	Rural
Poor population
	In million	110	113	119	47	72	158	77	81
	Share of total	51	40	33	21	55	39	29	59
Total population		216	283	361	229	132	405	267	138

Sources: Part A: Vito Tanzi, “Redistributing Income Through the Budget in Latin America,” Banca Nazionale del Lavoro Quarterly Review, Vol. XXVII, No. 108, 1974, pp. 5-27; United Nations, National Accounts Statistics: Compendium of Income Distribution Statistics, Statistical Papers, Series M., No. 79; and The World Bank, World Development Report, Washington, D.C., 1988; Part B: International Labor Organization, Meeting the Social Debt, Geneva, Switzerland, 1988.

^1/See Appendix Table 9 for country grouping.

View Table

Table 3.

Income Distribution and Poverty in Latin America, 1960s Through 1980s

Part A. Evolution of the Size Distribution of Income1/
	Early 1960s		Early 1970s		Early 1980s
	Lowest 20 percent	Highest 20 percent	Lowest 20 percent	Highest 20 percent	Lowest 20 percent	Highest 20 percent
Average	4	59	3	57	5	48
Lower income countries	4	62	4	57	5	47
Upper income countries	4	56	3	57	4	49

Part B. Evolution of Poverty
		1960	1970	1980			1985
				Total	Urban	Rural	Total	Urban	Rural
Poor population
	In million	110	113	119	47	72	158	77	81
	Share of total	51	40	33	21	55	39	29	59
Total population		216	283	361	229	132	405	267	138

Sources: Part A: Vito Tanzi, “Redistributing Income Through the Budget in Latin America,” Banca Nazionale del Lavoro Quarterly Review, Vol. XXVII, No. 108, 1974, pp. 5-27; United Nations, National Accounts Statistics: Compendium of Income Distribution Statistics, Statistical Papers, Series M., No. 79; and The World Bank, World Development Report, Washington, D.C., 1988; Part B: International Labor Organization, Meeting the Social Debt, Geneva, Switzerland, 1988.

^1/See Appendix Table 9 for country grouping.

The data indicate that the income share of the top 20 percent was almost 50 percent in the early 1980s, lower than in the early 1960s and the early 1970s. However, it is not obvious that the lower share indicates a genuine increase in the share of the middle-income group. In the two countries for which data are available for both the early 1970s and the 1980s, these shares changed in the opposite directions.

The severely uneven income distribution, particularly the small income share of the bottom 20 percent of the population, stresses the need to examine the extent of poverty closely. To discuss the extent of poverty, one needs to establish a poverty line and to estimate the number of people whose income falls below the poverty line. Part B of Table 3, based on a study conducted at the International Labor Organization (ILO), indicates how the extent of poverty has changed since 1960. ^21/

Through 1980, in spite of a steady increase in the total population, the poor population had remained stationary at around 110-120 million, and the share of the poor population in the total population had declined from about one half in 1960 to one third in 1980.

What were the major factors underlying the evolution of income distribution and poverty? Given the size of aggregate income, its size distribution will depend on how the wealth (both human and physical) is distributed and how the rates of return to various types of wealth are structured. The stability of the relative income shares of the main income groups underscores the rigidity of the fundamental social and economic institutions and perhaps policies that prolong the status quo in the size distribution of income. ^22/

The evolution of basic needs indicators confirm an improvement of the social conditions in Latin America between 1960 and 1980. Per capita calorie supply increased. The death rate declined. The primary school enrollment rate increased. The improvement was fairly widespread in both lower-income and upper-income countries (Table 4 and Appendix Table 12).

Table 4.

Basic Needs Indicators, 1960-85 ^1/

Sources: The World Bank, World Development Report, Washington, D.C., 1988.

^1/See Appendix Table 12 for country grouping.

Table 4.

Basic Needs Indicators, 1960-85 ^1/

	Indicators			Percentage Changes
	1960	1980	1985	1961-80	1981-85
Calorie supply	2,310	2,445	2,532	0.28	0.70
Lower income countries	2,172	2,318	2,416	0.33	0.83
Upper income countries	2,607	2,719	2,783	0.21	0.47
Death rate (percent)	12	9	8	-1.40	-2.30
Lower income countries	13	9	8	-1.40	-2.30
Upper income countries	10	8	7	-1.11	-2.60
School enrollment (percent)	93	102	104	0.50	0.40
Lower income countries	89	100	103	0.58	0.59
Upper income countries	101	109	108	0.38	-0.18

Sources: The World Bank, World Development Report, Washington, D.C., 1988.

^1/See Appendix Table 12 for country grouping.

View Table

Table 4.

Basic Needs Indicators, 1960-85 ^1/

	Indicators			Percentage Changes
	1960	1980	1985	1961-80	1981-85
Calorie supply	2,310	2,445	2,532	0.28	0.70
Lower income countries	2,172	2,318	2,416	0.33	0.83
Upper income countries	2,607	2,719	2,783	0.21	0.47
Death rate (percent)	12	9	8	-1.40	-2.30
Lower income countries	13	9	8	-1.40	-2.30
Upper income countries	10	8	7	-1.11	-2.60
School enrollment (percent)	93	102	104	0.50	0.40
Lower income countries	89	100	103	0.58	0.59
Upper income countries	101	109	108	0.38	-0.18

Sources: The World Bank, World Development Report, Washington, D.C., 1988.

^1/See Appendix Table 12 for country grouping.

Given the relatively stable size distribution of income, the evolution of poverty apparently reflected the fluctuation in the aggregate income. Between 1965 and 1980, the increase (2.5 percent per annum) in per capita income contributed to the reduction in the share of the poor population in the total population, although the size distribution of income did not change substantially. The continued growth of the population, however, did not allow the countries in the region to reduce the poor population. ^23/ While declining from over one half in 1960 to about one third in 1980 of the total population, the poor population had remained stationary at around 110-120 million. Through the 1980s, the expanding world economy, the relatively favorable external terms of trade, the availability of foreign financing had helped both lower-income and upper-income Latin American countries achieve the relatively high economic growth, although the external terms of trade had been volatile (Table 5 and Appendix Table 9).

Table 5.

Growth of GDP, Population, and Per Capita GDP in Latin America, 1965-85 ^1/

Source: The World Bank, World Development Report, Washington, D.C., 1988.

^1/See Appendix Table 9 for country grouping.

Table 5.

Growth of GDP, Population, and Per Capita GDP in Latin America, 1965-85 ^1/

	Annual Rate of Change
	GDP		Population		Per Capita GDP
	1965-80	1981-85	1965-80	1981-85	1965-80	1981-85
Average	5.0	0.2	2.5	2.3	2.5	-2.1
Lower income countries	4.8	0.2	2.6	2.5	2.2	-2.2
Upper income countries	5.3	0.2	2.3	1.9	3.1	-1.7

Source: The World Bank, World Development Report, Washington, D.C., 1988.

^1/See Appendix Table 9 for country grouping.

View Table

Table 5.

Growth of GDP, Population, and Per Capita GDP in Latin America, 1965-85 ^1/

	Annual Rate of Change
	GDP		Population		Per Capita GDP
	1965-80	1981-85	1965-80	1981-85	1965-80	1981-85
Average	5.0	0.2	2.5	2.3	2.5	-2.1
Lower income countries	4.8	0.2	2.6	2.5	2.2	-2.2
Upper income countries	5.3	0.2	2.3	1.9	3.1	-1.7

Source: The World Bank, World Development Report, Washington, D.C., 1988.

^1/See Appendix Table 9 for country grouping.

The picture that emerges from this analysis coincides with the one described in a recent World Bank study:

“Growth went hand in hand with considerable social improvements. Between 1960 and the late 1970s, life expectancy at birth, perhaps the most important indicator of welfare, increased from 56 to 64 years, which is not far below the average for European middle-income countries. Clean water had been made accessible to more people—from less than 40 percent to more than two thirds of the growing population—and the share of the labor force in agriculture had fallen from one half to one third. And while income distribution remained extremely skewed in all but a very few countries, the income share of the lowest four deciles did not shrink; thus most of the poor saw their standard of living improve at about the same pace as the entire population.” ^24/

b. Development in the 1980s

Between 1965-80 and 1981-85, with the population growth remaining at roughly the same rate, the reduction in economic growth resulted in a reduction in the growth of per capita GDP from 2.5 percent per annum to negative 2 percent (Table 5). Both lower-income and upper-income countries suffered similar declines in their per capita income.

The events that had led to the region’s economic decline in the first half of the 1980s have already been discussed (section III). What is unclear is how these external forces and the policy responses of the region’s governments have affected the pattern of income distribution and poverty. In the absence of an adequate data base, it is not obvious how the relative income share of the population groups changed during this period. The ILO study suggests a significant increase in the capital share of the declining aggregate income, at the cost of a declining labor share. ^25/ However, with the rural poor, including smallholders, remaining the largest poor group, it is not obvious whether the declining labor share suggests a declining share of income received by the poor group (Table 3, Part B).

The deterioration in economic growth reversed the earlier trend of declining poverty. The ILO study suggests that the poor population increased from 120 million in 1980 to almost 160 in 1985; its share in the total population increased from one third to almost 40 percent.

In spite of this evidence, it is difficult to assess the impact of the deteriorating external forces and the governments’ policy responses on the social conditions. The methodological, theoretical, and empirical difficulties of assessing this impact have been pointed out in many studies. ^26/ While the deteriorating external developments led to a decline in per capita income, an increase in unemployment in the formal sector, and possibly a reduction in government social expenditure, some governments responded to these developments by raising the relative price of tradables, implementing structural measures to improve economic efficiency, and improving the efficiency of existing social programs. The limited data on basic needs indicators in selected Latin American countries do not necessarily show a deterioration of these indicators. Actually, the data indicate a continued improvement in per capita calorie supply, in the crude death rate, and in the primary school enrollment rate (Table 4 and Appendix Table 12). This may be explained by the fact that per capita consumption declined far less than per capita income.

c. The role of budgetary policy

Tanzi (1974) points out that the social groups who tend to get the greatest benefit from public sector budgetary policy appear to be the urban middle-income groups. Many more recent studies have obtained a similar conclusion not only in the Latin American context, but also in the context of the countries in other regions. ^27/ In a sense, it is not surprising that government programs tend to benefit urban middle-income groups, since these are the most vocal and politically influential groups.

Tanzi (1974) also notes that on the revenue side, the relatively light tax burden, the predominance of indirect taxation, and the limited use of capital and income taxes made it obvious that taxation had no significant redistributive impact. The paper also noted the substantial difference between the de facto and the de jure tax systems; the former are much more progressive than the latter. On the expenditure side, it notes that even the supposedly pro-poor social expenditure had little effect on income distribution and that the group that seemed to get the greatest benefit was urban middle-income group.

Since the early 1970s, the tax burden in the region increased from 17 percent of GNP in 1972 to 21 percent of GDP in 1985 (Table 6, Part A). The increase was substantially larger in upper-income countries than in lower-income countries. However, the share of income and profit’ taxes in total current revenue increased by only 2 percentage points. The combined share of domestic taxes on goods and services and taxes on international trade remained at about one half. The changes in tax burden were diverse in the sample countries. Some countries (e.g., Argentina, Brazil, Chile, Costa Rica, Nicaragua, Panama, and Venezuela) increased the tax to GNP ratio substantially, but others (e.g., Bolivia, Dominican Republic, Mexico, and Paraguay) saw a decline in the ratio (Appendix Table 13).

Table 6.

Central Government Revenue and Expenditure in Latin America, 1972 and 1986

Source: The World Bank, World Development Report, Washington, D.C., 1988.

^1/See Appendix Tables 13 and 14 for country grouping.

Table 6.

Central Government Revenue and Expenditure in Latin America, 1972 and 1986

Part A. Revenue1/
			Share in Total Current Revenue
	Current Revenue in percent of GNP		Taxes on income and profit		Domestic indirect taxes		Taxes on international trade
	1972	1986	1972	1986	1972	1986	1972	1986
Average	17	21	18	21	25	31	24	17
Lower income group	16	19	15	20	29	32	30	21
Upper income group	19	25	23	23	19	31	14	10

Part B. Expenditure1/
			Selected Social Expenditure
	Expenditure in percent of GDP		Total		Education expenditure		Health expenditure
	1972	1986	1972	1986	1972	1986	1972	1986
Average	20	24	26	21	19	13	7	7
Lower income group	20	20	29	25	22	16	8	9
Upper income group	21	27	23	17	16	11	7	6

Source: The World Bank, World Development Report, Washington, D.C., 1988.

^1/See Appendix Tables 13 and 14 for country grouping.

View Table

Table 6.

Central Government Revenue and Expenditure in Latin America, 1972 and 1986

Part A. Revenue1/
			Share in Total Current Revenue
	Current Revenue in percent of GNP		Taxes on income and profit		Domestic indirect taxes		Taxes on international trade
	1972	1986	1972	1986	1972	1986	1972	1986
Average	17	21	18	21	25	31	24	17
Lower income group	16	19	15	20	29	32	30	21
Upper income group	19	25	23	23	19	31	14	10

Part B. Expenditure1/
			Selected Social Expenditure
	Expenditure in percent of GDP		Total		Education expenditure		Health expenditure
	1972	1986	1972	1986	1972	1986	1972	1986
Average	20	24	26	21	19	13	7	7
Lower income group	20	20	29	25	22	16	8	9
Upper income group	21	27	23	17	16	11	7	6

Source: The World Bank, World Development Report, Washington, D.C., 1988.

^1/See Appendix Tables 13 and 14 for country grouping.

The expenditure to GNP ratio increased from 20 percent in 1972 to 24 percent in 1986, entirely as a result of an increase in upper-income countries (Table 5, Part B). The share of two main categories of social expenditure (on education and health) in total expenditure declined from 26 percent to 21 percent. It is unclear how the pattern of expenditure incidence changed between 1972 and 1986. Assuming that it did not change significantly, one may conclude that both middle-income and lower-income groups were adversely affected by the decline in the share of these social expenditures.

In the 1980s, many Latin American countries, facing severe debt service burden, pursued both macroeconomic and structural adjustment policies. A dominant question in this regard has been how the poor could be protected from the possible adverse impact of some adjustment measures. Episodes of various targeted programs have been discussed, but the role of the government in these programs was diverse. For example, in the Latin American region, the much-publicized Emergency Social Fund in Bolivia bypassed much of the normal bureaucratic channels in government. ^28/ Among the governmental programs implemented in the early 1980s, two have been often mentioned.

First, Chile’s infant nutrition program (National Program for Complementary Food) aimed at using food supplements to induce mothers to bring their children to health centers for monthly checks on their nutritional status. The school lunch program aimed at maintaining adequate nutrition for school children. The implementation of these programs benefited from Chile’s long history of intervention in the social sector dating back to the 1920s. The targeting benefited both from Chile’s long history of intervention in the social sector and from fairly effective institutions that had existed prior to the adoption of the programs.

Second, Mexico’s targeted food stamp program aimed at allowing the poor to have access to low-priced tortillas, the maize flatcakes which are the staple of the poor, thus protecting them from the increase in food prices resulting from a reduction of food subsidies. In addition, the Mexican government also expanded two existing targeted programs to distribute subsidized milk and to supply rural food stores. These programs required the establishment of a wide-scale infrastructure. ^29/

a. A stylized model of growth-poverty alleviation dynamics

These questions have both normative and positive aspects. In the following analysis, this paper focuses on two positive aspects. First, it will assess the duration required for the poor group to attain a certain level of income on the basis of alternative growth rates and redistributive policies. Second, it will discuss the relationship between redistribution and economic growth.

One way to assess the first aspect is to start from a situation resembling the present situation of the poor in the region and to ask how long it will take for a certain rate of sustained economic growth to increase their incomes to the poverty line. For this purpose, assume that the income share of the poor, say, the bottom 20 percent of the population, is 5 percent. This means that the average income of the poor is 25 percent of the average income of the entire population. Assume further that that income is 50 percent of the poverty line. The poverty line is 50 percent of the average income of the entire population. The population is growing at 2.5 percent per annum. Assume further that the government does not attempt any redistributive policy, that is, the income share of the poor groups remains at 5 percent. How many years would it take for the average poor to attain an income level equivalent to the poverty line? Table 7 suggests that it would take almost 30 years if annual economic growth were to remain 5 percent. The duration would be shortened by two thirds, to 10 years, if the pace of economic growth doubled to 10 percent (Table 7, column (1)).

Table 7.

Growth, Redistribution, and Poverty Alleviation: Years Needed to Raise Income, of the Average Poor to the Poverty Line ^1/

^1/The annual rate of increase in the population is assumed to be 2.5 percent. It, is also assumed that the growth rate and the distribution of the additional income of the population are independent; the average, income of the poor is assumed to be 50 percent of the average-income of the population; and the average income of the poor is assumed to be 25-percent of the average income of the population or 50 percent of the poverty line.

Table 7.

Growth, Redistribution, and Poverty Alleviation: Years Needed to Raise Income, of the Average Poor to the Poverty Line ^1/

	Percentage Share of Annual Increase in Income to be Distributed to the Poor
Annual Growth Rate of Aggregate Income (percent)	(1)	(2)	(3)	(4)	(5)	(6)
	5	6	7	8	9	10
	Years
10	10	9	8	7	6	6
9	12	10	9	8	7	6
8	14	12	10	9	8	7
7	17	14	12	11	9	8
6	21	18	15	13	12	10
5	29	24	20	17	15	13

^1/The annual rate of increase in the population is assumed to be 2.5 percent. It, is also assumed that the growth rate and the distribution of the additional income of the population are independent; the average, income of the poor is assumed to be 50 percent of the average-income of the population; and the average income of the poor is assumed to be 25-percent of the average income of the population or 50 percent of the poverty line.

View Table

Table 7.

Growth, Redistribution, and Poverty Alleviation: Years Needed to Raise Income, of the Average Poor to the Poverty Line ^1/

	Percentage Share of Annual Increase in Income to be Distributed to the Poor
Annual Growth Rate of Aggregate Income (percent)	(1)	(2)	(3)	(4)	(5)	(6)
	5	6	7	8	9	10
	Years
10	10	9	8	7	6	6
9	12	10	9	8	7	6
8	14	12	10	9	8	7
7	17	14	12	11	9	8
6	21	18	15	13	12	10
5	29	24	20	17	15	13

^1/The annual rate of increase in the population is assumed to be 2.5 percent. It, is also assumed that the growth rate and the distribution of the additional income of the population are independent; the average, income of the poor is assumed to be 50 percent of the average-income of the population; and the average income of the poor is assumed to be 25-percent of the average income of the population or 50 percent of the poverty line.

This is a rather bleak picture. An annual growth rate of 10 percent appears to be virtually unattainable in the region at this time or at any time. Moreover, the goals in the simulations were rather modest in two senses. First, the assumption that the bottom 20 percent constitutes the poor population is a rather narrow definition of poverty; poverty is obviously more widespread than that. Second, even after the average poor, have attained an income level equivalent to the poverty line, a substantial number of the population would still receive an income below the poverty line.

The government could attempt to pursue policies to distribute a larger share of the increase in aggregate income to the poor. For the sake of simplicity, assume that the government could carry out such policies without any political resistance or administrative costs. Table 7 indicates that on the basis of a sustained annual economic growth of 5 percent, distributing 10 percent of annual increase in aggregate income would bring the income of the average poor to a level equivalent to the poverty line in 13 years; on the basis of a sustained annual growth of 10 percent, similar policies would bring about the same result in 6 years.

b. Equity, poverty alleviation, and economic growth

These results underscore the need to reinforce growth policies with redistributive policies in favor of the poor. Would such policies slow the pace of economic growth? The conventional analysis shows a number of channels through which more equity could reduce the pace of economic growth. First, more equal distribution of income might reduce work efforts. For example, a targeting scheme that guarantees a minimum level of income for the poor could take away any incentive to work from those whose original income was below the poverty line. Second, more equal distribution of income might reduce domestic aggregate savings and thus resources available for productive investment. Third, the administrative cost of implementing a redistributive scheme, such as programs targeted at the poor, reduces the amount of resources to be allocated for other productive investment.

The implications of income redistribution for economic growth, however, are more complicated than suggested by this conventional reasoning. The crucial reason is that the level of nutrition, health, and education may be a binding constraint of the work efforts of the poor. This is particularly true in the long run. Better nutrition, health, and education would strengthen the poor as productive units and would enhance their productivity. In some cases, better nutrition and health would enhance labor productivity even in the short run. Similarly, targeting could aim at relaxing the binding constraints of the work efforts of the poor. For example, increasing the availability of credit for the poor engaged in informal sector activities or providing employment opportunities to them would most likely enhance, rather than reduce, their productivity. The argument that equity would reduce the resources available for investment does not appear to be convincing if the increase in savings may not be invested domestically, but join the flight capital abroad (see section III). The last argument, that is, targeting requires resources that could be invested for growth, has merit, but the choice between more poverty alleviation and higher economic growth is not entirely a positive one, it involves normative judgment. Furthermore, it is not obvious that these resources would actually be invested for growth.

The conventional analysis might be correct with regard to redistributing income from the top to the middle-income groups; it may not be correct, however, with regard to redistributing income to the poor or the ultra-poor.

c. The role of fiscal policy for equity and poverty alleviation

How could the role of fiscal policy be improved for equity and poverty alleviation? There are two important ways.

First, fiscal policy should be more focused on efficiency gains for maximum growth. Fiscal policy should provide a less unstable macroeconomic environment. As discussed in section III, in Latin America, as well as in other regions, fiscal policy has too often been a major source of shocks, instability, and uncertainties, all of which are obstacles to mobilizing domestic and foreign savings and utilizing resources for productive investment. Fiscal policy has often been a root cause of capital flight. ^30/ Fiscal policy should also support more rational use of resources through appropriate structural measures. Macroeconomic stability, while promoting growth, is often crucial in protecting the poor. Both in Latin America and elsewhere, the poor are the most vulnerable to high inflation, since they are the least prepared to protect their assets and real earnings from the adverse impact of inflation. Indirectly, macroeconomic stability and the right macroeconomic policy would help the government’s revenue efforts (see below). ^31/ The structural role of fiscal policy should aim at providing an efficient tax system to mobilize sufficient domestic resources without causing distortions. While the effort is being made to enforce the tax law, efforts should also be made to make the tax base broader and the tax rate structure simpler and perhaps lower. ^32/

Second, the government social programs should be improved to protect the poor and should be made more effectively targeted to the poor. A broadening of the tax base would be helpful in achieving a more equitable tax system. The present tax structure in Latin America relies excessively on indirect taxation, which is often more regressive than direct taxation. The reliance on a few commodities, such as alcohol and tobacco, has also been found to be adverse on some vulnerable groups. ^33/ Expenditure programs should also be made more effectively focused on the poor. As indicated earlier in the section, many social programs in Latin America tend to benefit middle-income groups more than the poor. Investment in education needs to be focused more on primary than higher education. Studies suggest that both private and social rates of return to education are higher for the investment in primary education than in higher education (Table 8). ^34/ Health and nutrition programs could also be made to focus more on the poor. For example, not only in Latin America, but also in many other developing countries, the investment in small-scale health clinics in both urban and rural areas has been found to be more effective in protecting the poor than large-scale medical centers. As mentioned earlier, targeting could be designed to relax the binding constraints of the work efforts of the poor, thus to enhance, rather than reduce, the productivity of the poor. Effective targeting requires a careful consideration of the cost and benefit. It also requires the establishment of effective institutions. The Chilean experience with the infant and child nutrition program has demonstrated the usefulness of an effective and efficient institution for successful implementation of targeted programs.

Table 8.

The Return to Investment in Education in Latin America

Source: George Psacharopoulos, “Education and Development: A Review,” The World Bank Research Observer, Vol. 3, No. 1, January 1988.

Table 8.

The Return to Investment in Education in Latin America

	Primary Education	Secondary Education	Higher Education
	(in percent)
Social rate of return	26	18	16
Private rate of return	32	23	23

Source: George Psacharopoulos, “Education and Development: A Review,” The World Bank Research Observer, Vol. 3, No. 1, January 1988.

View Table

Table 8.

The Return to Investment in Education in Latin America

	Primary Education	Secondary Education	Higher Education
	(in percent)
Social rate of return	26	18	16
Private rate of return	32	23	23

Source: George Psacharopoulos, “Education and Development: A Review,” The World Bank Research Observer, Vol. 3, No. 1, January 1988.

d. Targeting as a means of poverty alleviation

Over the long run, sustained economic growth is the only viable means of poverty alleviation. Only those countries with the record of sustained economic growth have been able to reduce absolute poverty. Across the countries that have achieved rapid economic growth, however, the record of poverty reduction differs substantially in the industrial and developing worlds. In the developing world, the inadequacy of the poverty alleviation strategy relying only on economic growth is apparent in view of the many years it takes to reduce poverty.

Currently, targeting is being intensely discussed as a means to reinforce the growth strategy to alleviate poverty. Making the present social programs more effectively targeted to the poor is an important task facing Latin American governments. However, targeting is not without problems. This paper has already mentioned the possible impact of targeting on economic efficiency, possible costs, and political constraints. Successful targeting requires the fulfillment of strict political, institutional, and economic conditions. Various studies have discussed some of these conditions. ^35/

Latin American countries, in comparison with most other developing countries, have both advantages and disadvantages in fulfilling these requirements. On the one hand, as middle-income countries, many Latin American countries have relatively more domestic resources to deal with poverty than many other developing countries. On the other hand, the high debt burden and the recent economic stagnation in many countries impose serious constraints on poverty alleviation efforts. Political conflicts that would arise from major government efforts to alleviate poverty could be more serious than in other developing countries where the majority of the population is poor.

Another constraint is high inflation. The average annual inflation in the region increased from 24 percent during 1966-80 to 86 percent during 1981-86. It is generally recognized that high inflation hurts the poor more than the nonpoor at least in the short run. Any attempt to expand government poverty alleviation programs, if accompanied by a higher inflation, could actually have a net negative effect on the poor. This macroeconomic constraint to targeting or poverty alleviation is serious, particularly when the targeting programs are not effectively focused on the poor.

To illustrate this point, a simple example may be used. Suppose a country’s aggregate output (or income) is 100. The government taxes and spends 20. Suppose initially that the government does not run a deficit. Suppose further that the public holds 20 of income in the form of money. The share of income for the poor is 5 percent. The price level remains constant. In this economy, suppose the government, in an effort to expand its social programs, attempted to raise expenditure by 2 by running a deficit of 2, which is monetized. With the initial money stock of 20, in the absence of a change in real output or the velocity of money, an increase of 2 in the money supply would increase prices by 10 percent. What would be the net effect of the increase of 2 in government social expenditure and the increase of 10 percent in prices? To have a meaningful discussion, two crucial parameters should be known. First, the incidence of the increase in government social expenditure should be known. Suppose that the share of 2 that goes to the poor were 25 percent, substantially higher than the overall income share of the poor. Second, the speed with which the poor’s nominal income responds to inflation should be known. In this example, on the basis of the assumptions made so far, if the nominal income of the poor were to remain fixed in the short run, the increase in social expenditure would result in an increase of 0.5 in the benefit for the poor, accompanied by a reduction of a same amount in their real income. The net benefit for the poor would be higher if the speed of the increase in nominal income for the poor were faster and if the share of the poor in the increase in government social expenditure were higher.

This example illustrates both the promises and the limitations of targeting. In Latin America, for targeting to be more successful, it is crucial that macroeconomic stability be enhanced and inflation be reduced. Moreover, it is crucial that revenue efforts be improved, the success of which, in turn, depends not only on efforts to reform tax systems, but also on sound macroeconomic policy measures.