Despite rapid economic growth and continued efforts by the government to improve social equity, a common perception in Chile is that some groups have lagged behind and income distribution has worsened. Using the best available analytical and statistical techniques to examine the developments since 1987, this report has sought to examine the validity of these concerns. It has also tried to identify plausible policy options that the government could follow towards reducing both poverty and income inequality, while not compromising the growth potential of the economy.
What We Have Confirmed
First, the report has confirmed that even within a short period of time (1987–1994), there has been a significant decline in poverty. The poverty headcount and poverty deficit were halved during this period. While 41 percent of the population lived in poverty in 1987, by 1994 this figure fell to 23 percent. The incidence fell from 13 percent in 1987 to 5 percent in 1994. The report confirms that high economic growth is strongly and positively correlated with declining poverty; during the sub-period 1992–1994 when economic growth fell, there was also a deceleration in the rate at which poverty declined.
Second, for the poorest decile group, the slowdown in growth from 11 percent in 1992 to 4 percent in 1994, and the accompanying rise in unemployment from 4.5 percent to 6 percent actually meant a small increase in this group's poverty. The reduction in unemployment to rates around 4.5 percent in 1995, by the same token, is likely to have reduced poverty for this group more recently.
Third, the report confirms that income inequality in Chile is high by international standards. However, sustained high growth has resulted in a significant reduction in poverty, despite this high level of inequality. Chile's success in reducing headcount poverty during 1987–1994 rivals the performance of countries such as Korea, Indonesia, and China; it should be noted, however, that these other countries are believed to have experienced increases in inequality in the process of high economic growth.
Fourth, the reduction in poverty during 1987-1994 has benefited almost all groups classified as vulnerable at the beginning of the period. While growth obviously helped those among the poor who could work, poverty reduction policies have benefited even non-workers: for example, older, poorly educated, male and female household heads in both rural and urban areas experienced significant declines in the probability of being poor.
Fifth, education is an important determinant of labor earnings and hence of household income. Differences in educational attainment account for almost one-third of overall income inequality, and are by far the largest single explanatory factor.
The report confirms that the income tax is progressive and the VAT is regressive -- the combined effect of which is a tax structure that is largely inequality-neutral. On the other hand, if public spending on pensions is excluded, the analysis confirms that total social expenditure has substantial redistributive impact, most of which is accounted for by basic education and health care.
Finally, overall, Chile's policy in recent years of growth with equity has been demonstrated as effective in reducing poverty without exacerbating income inequality. As with virtually any program, there is scope for improvement at the margin; this report attempts to draw out options for consideration by Chilean policymakers.
What We Have Refuted
First, contrary to a popular perception, this report documents that income inequality has not increased during 1987-1994; in fact, there is a slight improvement in the income distribution evidenced by a small decline in the Gini coefficient from 0.55 to 0.53.
Second, the report finds that inequality in labor earnings has declined considerably in the last decade. While the Gini coefficient for labor earnings in Gran Santiago between 1960 and 1987 rose from 0.43 to 0.58, it has fallen steadily since then to 0.46 in 1996.
Third, refuting growing concerns about precariousness of employment and earnings, the report finds that job and income security has increased substantially since 1987. Average expected tenure on a job has risen from 47 to 55 months between 1987 and 1995, and the average duration of unemployment has been halved from 5.5 to 2.8 months.
Fourth, allaying concerns that rapid growth has exacerbated regional income disparities, the analysis shows a convergent pattern. All indicators of inter-regional dispersion of income show a reduction between 1987 and 1994: real per capita income of the poorest region (region IX) grew by 10 percent, which is twice the national average for the period. Reassuringly, reductions in poverty have been shared by all regions (except region XI, which has only 0.6 percent of the country's population) and these gains have not been restricted to Gran Santiago. While Santiago did better in reducing headcount poverty ratios, other regions matched its performance if indicators that measure the depth of poverty are used instead.
Finally, less encouragingly, while educational attainment of all income groups increased between 1987 and 1994, the difference in years of schooling between the richest and poorest groups has increased. While the years of schooling for the poorest 20 percent increased by 0.8 years, the schooling of the richest 20 percent increased by 1.3 years.
Limitations of Our Statistical Analysis
Besides labor earnings and other income (including imputed rents of owner-occupied dwellings), measures of household income in this report include monetary Teamsters such as means-tested old age, disability, and family allowances, and other cash subsidies. Omitted from our income measures are welfare-enhancing improvements such as better access, lower prices, and higher quality of public utilities, which are believed to have improved more for poor than for non-poor households.
More importantly, this income measure does not include the value of government transfers in kind, though public housing, health, and education programs, which together were 6.5 percent of GDP and 30 percent of total government expenditures in 1994. Non-availability of household level data for all four survey years for these transfers made this adjustment impossible. Since the incidence of public spending on these programs is regarded as progressive, this results in a greater underestimation of the real income of poorer households than of the non-poor. Annex 1 in Volume II of the report does, nevertheless, report some of the amputations for public education and health services by household quintiles made by MIDEPLAN.
The findings discussed above, and their policy implications, must be considered keeping this caveat in mind: our estimates probably overstate the level of poverty in Chile and understate its reduction over the period examined. This shortcoming in data collection will need to be corrected for policymakers to be able to attain a finer degree of targeting of social programs than already achieved, viz., if the goal of "supertargeting" of public spending is to be achieved.
Implications for Labor Policies
The labor market has served Chile's poor well in recent years. Labor earnings have contributed more to poverty reduction than non-labor income (mainly transfers). While household inequality measures have not shown an improvement, inequality in labor earnings has fallen significantly since 1987. While these developments may not justify inaction, they do call for restraint in changing labor market policies.
Labor legislation that would make dismissals more costly for employers is likely to be counterproductive and could, as in other countries, actually increase unemployment.
An unemployment insurance system is largely unnecessary given existing severance benefit legislation, the low duration of unemployment, and the existence of means-tested transfers. Steps to introduce an unemployment insurance system should be carefully evaluated keeping in mind the distortions such a system can create.
While the minimum wage could be increased, on the basis of the analysis in Annex 3 it appears that further increases in the ratio of the minimum wage to per capita income is likely to result in greater unemployment and could actually worsen poverty.
Implications for Education and Training Policies
Compared with measurement of the effects of improved access to schooling (reflected in years of schooling), the benefits of improved quality of schooling are difficult to measure. Our estimates, however, indicate that increasing access to primary and secondary education yield smaller increases in earnings relative to both improvements in education quality at these levels, and increased access to higher education. But even access to higher education by the poor is best facilitated by raising the quality of their primary and secondary education. These findings suggest that efforts to improve education quality in municipal and private subsidized primary and secondary schools would help reduce poverty and inequality. Combined with institutional measures to improve the quality of instruction, reducing public spending on tertiary education (62 percent of which goes to the richest one-third of individuals) and allocating it to quality-enhancing initiatives at lower levels could be an effective medium-term strategy to reduce inequality of education, and hence of earnings.
The socioeconomic status of vocational-technical students is considerably lower than those enrolling in humanistic-scientific education programs, and secondary vocational education has lower economic returns than general education because it is more expensive. Until improvements in primary education quality for the poor have been achieved, however, measures to increase the relevance and quality of secondary vocational schools are likely to be equity enhancing.
Based on international experience, if designed well, such training programs may help carefully selected groups find jobs but are not likely to reduce earnings inequality of employed workers. Because these programs are always expensive, Chile's training programs should be scientifically evaluated (with a control group).
Implications for Tax and Social Expenditure Policies
Because introducing dual VAT rates or exemptions for goods comprising relatively large shares of the budget of poorer households will raise administrative costs, further study of the tax system is warranted despite the regressivity of the current VAT system, if looked at in isolation.
Despite the payoff in terms of inequality reduction that a greater reliance on income taxes would yield, the current strategy of concentrating on closing loopholes has the advantage of being equity-enhancing and increasing tax collection efficiency.
To better monitor the incidence of social spending as a whole, the classification of what constitutes social spending should be re-examined. Spending on items such as pensions of public servants, military and police should not be lumped with those with clearly identified social policy objectives.
Chile's past successes in improving the efficiency of targeting notwithstanding, there is considerable scope for further improvements. This can be done both by allocating a greater fraction of social spending to programs that have had better targeting records (e.g., primary education and health, and cash benefits, where the poorest 40 percent receive 60 percent of expenditures) and improving targeting efficiency of programs such as housing subsidies, which are essentially untargeted.
The report finds that despite extensive individual and household-level controls, simply living in poorer comunas (districts) remained an important factor in accounting for the likelihood of being poor. Strong location or "neighborhood effects" on poverty deserve further study, because they expand the set of potential measures to deal with residual or hard core poverty.
These findings and policy implications are based on five studies that were commissioned for this report. These studies, in turn, analyzed recent changes in income distribution and poverty, provided a profile of poverty, isolated critical policies that influence poverty reduction, examined selected labor market issues and policies, and evaluated social policies in Chile. A brief synopsis of the results of these studies, together with a more detailed discussion of these topics are available in Volume I of the main report. The findings and the analytical and statistical techniques employed in all five studies are contained in Volume II.
Back to Poverty Assessment Summaries — Latin America & Caribbean