SciELO - Scientific Electronic Library Online

 
vol.38 issue1The impact of income adjustments in the Casen Survey on the measurement of inequality in ChileAlleviating extreme poverty in Chile: the short term effects of Chile Solidario author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand

Journal

Article

Indicators

Related links

Share


Estudios de economía

On-line version ISSN 0718-5286

Estudios de Economía vol.38 no.1 Santiago June 2011

http://dx.doi.org/10.4067/S0718-52862011000100004 

Estudios de Economía. Vol. 38 - No 1, Junio 2011. Págs.67-100

ARTÍCULOS

 

The longer-term effects of human capital enrichment programs on poverty and inequality: Oportunidades in Mexico*

Los efectos a largo plazo de programas de incremento en el capital humano sobre la pobreza y la desigualdad: Oportunidades en México

 

Douglas McKee** Petra E. Todd***

** Post-doctoral scholar at the Economic Grow Center at Yale University.
*** Professor of Economics at the University of Pennsylvania.


Abstract

Previous empirical research has shown at Mexico’s Oportunidades program has succeeded in increasing schooling and improving health of disadvantaged children. is paper studies the program’s potential longer-term consequences for the poverty and inequality of esthe children. It adapts methods developed in DiNardo, Fortin and Lemieux (1996) and incorporates existing experimental estimates of the program’s effects on human capital to analyze how Oportunidades will affect futurthe earnings of program participants. We non parametrically simulate the earnings distributions, with and without the program, and predict at Oportunidades will increase future mean earnings but have only modest effects on poverty rates and earnings inequality.

Key words: Oportunidades, Human capital, Schooling, health, Poverty, Inequality.


Resumen

La investigación empírica previa ha mostrado que el programa mexicano Oportunidades ha tenido éxito en aumentar la escolaridad y mejorar la salud de niños desfavorecidos. Este artículo estudia las potenciales consecuencias de largo plazo en la pobreza y desigualdad que afectan a estos niños. Adapta modelos desarrollados por DiNardo, Fortin y Lemieux (1996) the incorpora estimadores experimentales existentes de los efectos del programa en el capital humano para analizar cómo Oportunidades afectará los ingresos futuros de los participantes del programa. Simulamos de manera no-paramétrica las distribuciones de ingreso, con y sin el programa, y predecimos que Oportunidades aumentará los ingresos medios futuros pero sólo tienthe efectos modestos en la tasa de pobreza y la desigualdad de ingresos.

Palabras clave: Oportunidades, Capital humano, Escolaridad, Salud, Pobreza, Desigualdad.

JEL Classification: H50, I00, J24, O12, O54, O15.


1.      INTRODUCTION

In recent years, governments in many Latin American countries have adopted conditional cash transfer (CCE) programs as a primary strategy for alleviating poverty and stimulating investment in human capital. Thesthe programs typically provide cash grants to poor families if ey send their age-eligiblthe children to school and subsidies for regularly visiting health clinics. Mexico and Brazil first adopted CCthe programs in e 1990's. Since en, programs wi similar incentives have been introduced in Argentina, Chile, Colombia, Costa Rica, El Salvador, Ecuador, Honduras, Nicaragua, Peru and Uruguay1.

The Mexican Oportunidades program (formerly called PROGRESA) was rigorously evaluated using bo experimental and non-experimental evaluation designs. In e first two years (1998-1999) of its implementation in rural areas, the program was evaluated using a place-based social experiment at randomized 506 villages in or out of the program. The experimental results demonstrated statistically significant program impacts on increasing schooling enrollment and attainment, reducing child labor, improving health and nutrition outcomes and reducing poverty2. Partly on e basis of esthe observed positive program impacts, e Mexican government expanded the program into urban areas in 2002. By 2005, the program covered five million families and had an annual budget of U.S. $2.1 billion. A non-experimental evaluation carried out in urban areas found statistically significant program impacts similar in magnitude to ose found in rural areas.

As noted, previous evaluation studies of the Oportunidades program documented the program's short-term impacts. This paper takes as a point of departure the observed impacts on education and nutrition and estimates the effects of ese changes on e futurthe earnings distributions of the children currently participating in the program. The question we consider is how the program's impacts on human capital, as measured by years of schooling attained and increases in height (interpreted as an indicator of long-term nutritional status), will affect futurthe earnings inequality and poverty of the younger generation. In e last decade, Mexico has ranked among e countries in Latin America with the highest income inequality. A study by Lopez-Acevedo (2004) finds at theducational inequality accounts for e largest share of Mexico's earnings inequality, suggesting at human capital enrichment programs could be an effective instrument for reducing inequality. Freije, Bando and Arce (2006) show at Oportunidades has significantly decreased the poverty rate among e current generation of recipients, but little is known about e longer-term effects of the program on poverty.

Our approach to simulating program impacts on earnings distributions adapts for use in program evaluation a nonparametric decomposition method originally developed in DiNardo, Fortin and Lemieux (1996) and extends is method to allow for probability mass at zero in earnings distributions. Existing micro-simulation approaches for predicting effects of conditional cash transfer programs have focused on e short-term and are mainly based on parametric modeling frameworks3. The parametric models can be quite rich, but ey typically impose strong functional form assumptions. The goal in microsimulation studies is usually to forecast the effects of programs prior to their implementation, whereas our aim is to understand how program impacts at have already been estimated will affect futurthe earnings and poverty. Our approach is fully nonparametric and does not impose any functional form assumptions on the earnings-height-education-work experiencthe relationship, other an continuity and differentiability. We find evidence of nonlineari-ties in e estimated relationship at shows e benefits of flexibility wi regard to model specification. We use the nonparametric simulation method to compare the earnings and employment distributions wi and wiout the program and to compare our inferences to ose at would be obtained using more standard parametric approaches.

The key implications from e analysis are at the program's impacts on education and height will increase mean futurthe earnings of beneficiaries but will likely have little impact on earnings inequality. The modest overall observed impacts on inequality are attributable to two main factors. First, the program targets children from poor family backgrounds, and family background is an imperfect predictor of futurthe earnings. Children from poor backgrounds ultimately get distributed roughout e adult earnings distribution due to substantial intergenerational mobility. Second, we find important nonlinearities in the relationship between earnings, education and height, the most notable being at the returns to education are greater for post-primary years of education. Such nonlinearities imply at people who would obtain higher levels of schooling in e absence of the program tend to benefit more from the intervention, which contributes to widening raer an lessening inequality.

Our empirical analysis is based on e first wave of e Mexican Family Lifthe survey (MxFLS-1) which was collected in 2002. the survey collected data for all members of 8,440 households and includes information about labor force participation, income for bo primary and secondary jobs (including self-employment), education, and health. It also contains measures of family background, at we use to simulatthe program targeting. Our final analyses use a subsample of 5,171 individuals age 25 to 40.

This paper proceeds as follows. Section two describes the nonparametric simulation method and how we adapt and use it to study how Oportunidades affects employment and e overall thearnings distribution. Section rethe describes e Mexican Family Lifthe survey and our analysis samples. the empirical results arthe presented in section four. Section five concludes.

2.      METHODOLOGY FOR SIMULATING PROGRAM EFFECTS ON POPULATION EARNINGS DISTRIBUTIONS

The simulation method at we use to study program effects on earnings and employment outcomes is adapted from a wagthe decomposition method originally proposed in DiNardo et al. (1996). Ththeir study uses e method to investigate the effects of institutional and labor market factors on changes in e U.S. wagthe distribution over time. Their approach writes e overall wagthe density at time t, ƒw = (w | 1) in terms of e conditional wagthe densities, where conditioning is on a set of labor market or institutional factors, z, whosthe effects on earnings ey analyze:

In their study, z includes variables indicating union status, industrial sector, and wheer e wage falls above or below e minimum wage. Counterfactual wagthe densities are constructed by replacing ƒz (z | 1) by a different hypoetical conditional density, gz (z | 1).

We apply e DiNardo et al. (1996) method to simulatthe earnings densities wi and wiout a program intervention, where the program intervention changes the distribution of z. We extend e method to account for simultaneous analysis of bo employment and earnings by permitting the earnings distribution to have a mass point at zero due to nonparticipation. In is section, we first describe e simulation approach in general terms, and en how it applies to evaluating the effects of the Oportunidades program.

2.1.      Basic method

Denote some outcome of interest (earnings) by y and define its density in terms of its conditional density (conditional on somthe observed characteristics x):

Suppose at the program intervention changes the distribution of x from but at the distribution of y conditional on x stays the same . e new unconditional distribution of y would be given by:

We wish to simulate the effect of the program intervention on e outcome y as it operates rough changing x. For example, suppose at the variable x represents years of schooling attained and height and at the program intervention increases schooling attainment and height by some amount, i.e. = x + Δx. Suppose also that we have a sample of size n drawn from e unconditional density, f(x). If we know Δx we can generate for each individual i = xi + Δxi . We can simulate e post-program earnings density (y) at a point y0 by the average:

where (y, xi) and (xi) arthe nonparametric estimators of e unconditional densities computed from e original (pre-program) sample:

αy and αx are bandwids at are assumed to satisfy e usual requirements for consistent kernel density estimation4.

The MxFLS data are a stratified sample, so sampling weights are required to reweight e sample back to population proportions. Incorporating sampling weights into e simulation method is straightforward. Assume each observation has a sampling weight, ωi, and at the weights are scaled so that ∑ωi= n. e weights can bthe incorporated into e estimation of ƒ (y) as follows:

and also into e estimation of e unconditional kernel densities:

For expositional clarity, we suppress e weights in e remainder of the discussion, alough wthe incorporate em in e estimation.

2.2.      Accounting for probability mass at zero

Kernel density estimation can approximate well the distributions of continuous random variables, but in our data many people (especially women) report zero earnings. the program intervention could increasthe earnings among workers as well as change the probability of having positive earnings. We accommodate e mass point at zero in the earnings distribution by thestimating the density of earnings as a mixture, where wi somthe probability individuals earn zero and with the remaining probability ey earn income drawn from the density of income conditional on its being positive, fy>0(y). Bo the probability of having positive earnings and the magnitude of earnings are potentially affected by the program.

Let be e random variable representing the distribution of income implied by the counterfactual distribution of . Again, we assume the distribution of y conditional on x stays constant; in other words at the density of earnings conditional on schooling attainment and height is the same wheer or not the program is in place. is implies at

The stability assumption implicitly rules out general equilibrium effects, because it assumes at increases in the population in schooling attainment or height do not affect the earnings premium for ose characteristics.

We can obtain the probability of zero earnings, Pr( = 0), wi the program intervention (under e counterfactual (x)) using e following:

where X is e support of xi and where

In e last equation, 1 (yi = 0) is an indicator at denotes wheer e individual has zero earnings.

Let (y) be the density of income conditional on its being positive. e counterfactual distribution of y conditional on y being positive is given by:

We estimate e conditional density by:

We estimate e conditional densities at a point (y0, x0) using e standard kernel density estimator applied to e subset of data for which income is positive:

We now have all the ingredients to simulate e post-intervention earnings distribution. Earnings is 0 wi probability Pr ( = 0) and is drawn from (y)wi Pr ( = 0) .

2.3.      Measures of poverty and inequality

After simulating the distribution of earnings wi and wiout actual and hypoetical program impacts, it is possible to examine the effect of at the program intervention has on poverty and inequality using standard measures considered in the poverty measurement literature. Below, we briefly summarize e measures at we use in the empirical analysis as functions of e estimated densities, taking into account at densities may havthe probability mass at zero. For a recent discussion of the relative merits of alternativthe poverty and inequality measures, see Foster and Szekely (2007).

Headcount Ratio. The headcount ratio is e fraction of the population below a predefined poverty line. Denote e value of the poverty line by L

Averagthe poverty Gap Ratio. The averag the poverty gap ratio is e mean shortfall between an individual's income and the poverty line (wi ose above the poverty line having no shortfall) expressed as a fraction of the poverty line:

Foster-Greer-orbecke (1984) Index. The Foster-Greer-orbecke (1984) index is a weighted version of the averagthe poverty gap ratio at gives more weight to poorer individuals:

Coefficient of variation. The coefficient of variation is a common measure of dispersion of a distribution, defined as

Inter-quantile ranges. Another common measure of the dispersion of a distribution is the interquartile range. The differences between quantiles of y can be computed directly from the empirical cdf:

Gini Coefficient. The Gini coefficient is widely used as a measure of inequality of a distribution of income. Its values range between 0 and 1, wi 0 corresponding to perfect equality and 1 corresponding to perfect inequality (one person has all the income).

Theil Entropy Coefficient. Thee theil entropy coefficient can be computed from a set of observations by:

If everyone has the same (i.e., mean) income, en e index equals 0. If one person has all the income, en e index equals ln n.

Taking e limit as n → ∞ , we get e following formula in terms of the density, conditional on y > 0:

Generalizing is measure to e case where ere can bthe probability mass at 0 gives e following:

Below, we report how the program affects each of ese alternative measures of poverty.

2.4.      Applying e simulation method to evaluation of Oportunidades

We next describe how the nonparametric simulation method is applied in e context of evaluating Oportunidades. y represents labor earnings, and is modeled as a function of ree covariates: e denotes years of schooling attainment, h denotes height in centimeters (a measure of long-term nutritional status), and x denotes years of potential labor market experience5. e conditional density of labor market earnings is

The overall income distribution integrates over the observed schooling attainment, height and experience distribution in the population:

the Oportunidades program is known to impact schooling attainment levels (e) and height (h) and we want to assess how these impacts translate into changes in the earnings distribution. If participation in the program was universal, we could nonparametrically simulate the effect of the program on the income distribution simply by augmenting schooling attainment and height values by the expected program impacts. Let Δe denote e expected impact on schooling attainment and Δh e impact on height.

Becausthe nonparametric estimation methods do not extrapolate beyond the observed support (A), is simulation can only be performed for e subset of people for whom (e + Δe, h + Ah, x) ∈ A, which wthe denote by S.

e above equation assumes at theveryone experiences a program effect of the magnitude (Δe, Δh), but Oportunidades was targeted to a subset of the population based on poverty-related criteria at arthe discussed in detail below. Let D = 1 for e subset of individuals targeted by the program and D = 0 for ose not targeted. e overall income distribution at results, g(y), reflects at of e combined targeted and nontargeted subgroups:

Suppose the nontargeted subgroup experiences no effect of the program6. e larger e subgroup targeted by the program (Pr(D = 1)), e larger will be e potential effect on the overall earnings distribution.

Using is methodology, we can explore the relative contribution of impacts on schooling attainment and height in changing e overall income distribution, by considering e case where (i) we set Δe = 0 and e only effect is rough Δh, and (ii) where Δh = 0 and e only effect comes rough Δe. Implementing e simulation estimator of the previous section requires nonparametrically estimating e conditional density ƒ (y ǀ e, h, x) and e unconditional density fe,h,x (e,h,x). We estimate e latter using a rethe dimensional kernel density estimator:

where ɑe, ɑh and ɑx are e bandwid choices. In our analysis below, we use a Gaussian kernel and apply Silverman's rule for univariatthe distributions to each dimension of e data (Silverman, 1986). We also experimented wi other bandwid choices and found our main results were quite robust. To estimate e conditional density ƒ (y ǀ e, h, x), observe at the conditional density can be expressed as e ratio of two joint unconditional densities:

each of which can bthe nonparametrically estimated by standard kernel density estimators.

THe convergence rate of pointwisthe nonparametric density estimators slows down as the dimensionality increases, a problem known as e curse of dimensionality. However, the proposed estimators average over the nonparametric estimates and erefore converge at a faster rate.

3.      DESCRIPTION OF THE ANALYSIS SUBSAMPLES

We analyze data from e Mexican Family Lifthe survey (MxFLS-1), which conducted interviews wi 8,440 households in 150 communities in 2002. Every household member age 15 or older was interviewed, yielding about 38,000 individual interviews. 16 of Mexico's 32 states/districts are represented (roughly 10% of the population resides in ese states). Weights arthe provided to make e sample nationally representative. The survey includes comprhensive information on employment and income for bo primary and secondary jobs in e formal and informal sectors. the survey also includes information on household structure, education, and health. The key variables used in simulating counterfactual outcomes arthe income, employment, education level, height and labor market experience. Appendix A describes in morthe detail how we construct each of ese variables from e data.

Table 1 presents descriptive statistics for our two main analysis samples: Adult men and women age 25 to 40. About 10% of men and 64% of women report zero labor income. Mean monly earnings for males are 3,945 pesos and for women 1,140 pesos, where e means include zeros for nonworkers7. e average education level for men is 8.8 years, which is about one year higher an e average for women of 1.1 years. Men are on average 166 centimeters tall, and women are on average 153 centimeters tall. The Gini coefficient for earnings of men is 0.483 and for earnings of women is 0.819. e higher coefficient for women reflects e fact at a large fraction of women in Mexico do not work, so the earnings distribution for women is more unequal an at for men8.


4.      EMPIRICAL RESULTS

We use e methods described in section two to simulate the effect of the Oportunidades program on the earnings distribution as it operates rough changing education and height levels of the younger generation. We infer the relationship between earnings, education, height and labor market experience from information on adults who are age 25 to 40 population and en use at thestimated relationship to draw inferences about how increases in schooling and height would affect earnings distributions. Experimental evaluations of the Oportunidades program (as well as of its predecessor, the PROGRESA program) have found at the program increases schooling attainment levels by 0.6 years on average and adds about one cm to height for both men and women9. We consider e following hypoetical combinations of impacts and their effect on the earnings outcome distribution: (a) an increase in schooling attainment of 0.6 years, (b) an increase in height of one cm, (c) a combined increase in schooling attainment and height in the magnitudes specified in (a) and (b), (d) an increase in schooling attainment by ththree years, and (e) an increase in height of ththree centimeters. An increase of three years of education or an increase in height of three centimeters is a very large impact at is much greater an what was observed under the program, but wthe include these hypoetical impacts simply for purposes of comparison.

4.1.      Program targeting

Our goal is to simulate e longer-term effects of Oportunidades on earnings inequality and poverty. Ideally, we would compare two groups: e "treatment" group would be the population targeted as children by the program observed 20 years later and e "control" group would be the same people in a world where the program did not exist. Unfortunately, we cannot currently observe either group. The program was implemented relatively recently (in e late 1990's), so many of the children who participated are still too young to observe their longer-term labor market outcomes. Additionally, alough we can observe which families are currently participating in the program, it is likely at children from today's Oportunidades households may not emselves meet the program eligibility criteria as adults. In fact, one of the primary goals of the program is to reduce the intergenerational transmission of poverty.

Our simulation is erefore based on a synetic cohort approach at assumes stability in earnings relationships for neighboring cohorts. In particular, it assumes at individuals age 25 to 40 can be used to represent e futurthe earnings of children in families currently participating in the program. We simulate the effects of Oportunidades by identifying e 40% of current 25-40 year-olds at would have been most likely to be targeted when young had the program been available, making use of the observed family background characteristics. We analyze the effects of the program by changing is group's observed characteristics (education, height, and potential experience) in a way at is consistent with the impacts at have been measured in recent program evaluation studies.

THe MxFLS-1 dataset does not contain information on all the criteria used to determine eligibility for Oportunidades, and in fact the exact eligibility criteria are not public. However, from program officials we have learned e approximate criteria and use the most closely related variables from e MxFLS-1 dataset to approximate eligibility. Specifically, we estimate a probit model for program participation using data on children (age 9 to 12) who are currently participating in Oportunidades as a nonlinear function of several variables: mother's education, father's education, wheer e household has indoor plumbing, and e number of children age 0-10 residing in e household.

Table 2 presents descriptive statistics for ese variables. In e sample, 31% of children participate in Oportunidades. the program is most active in e poorer souern states (Chiapas, Oaxaca, Guerrero, Michoacan, and Puebla), where 31% of the children live. On average, the children in e sample have mother s with 4.1 years of schooling attained and faers with 5.2 years. Only 46% of esthe children live in households wi indoor plumbing. Table 3 shows e estimated coefficients from the probit model for program participation10. As expected, parental education, indoor plumbing, and the presence of young children in the household are highly significant predictors of program participation.



Next, we compute a propensity score (the predicted probability of being eligible and participating in the program) for each adult age 25 to 40 using e estimated probit model coefficients and measures of their family background (parental education, characteristics of e household when ey were age 12, and an approximation of e number of children age 0 to 10 in the household at at time). Alough the actual targeting of Oportunidades is based on several additional variables, we have to restrict e analysis to the subset available in e dataset for both children and adults, which fortunately includes e major determinants of program eligibility. We classify e 40% with the highest predicted probabilities of participation as e "target group" and e remaining 60% as e "non-target group".

Table 4 compares e characteristics of e target and non-target groups, separately for men and women. For both men and women, the target group has much lower maternal and paternal education levels. Individuals in the target groups also grew up wi morthe young children in households at were much less likely to have indoor plumbing. For both men and women, there is a two year schooling attainment gap between e target and non-target groups as well as a two cm difference in height. e labor market experience measure we use is Mincer potential experience, which equals age minus years of education minus six. The target group has more experience under is measure, mainly because of having lower schooling attainment11.


The mean levels in Table 4 shows at the target group is less advantaged an the non-target group. In particular, mean monly earnings are 3,300 pesos per mon for targeted men and 4,300 pesos per mon for non-targeted men. Targeted women can expect about half (100 pesos per mon) the labor income of non-targeted women (1,500 pesos per mon). But there is still substantial overlap in the two earnings distributions, as shown in Figure 1. The top panel describes men's labor income while e bottom panel describes women's. The solid line in each panel is a nonparametric estimate of the density of positive earnings, while the two dashed lines correspond to the densities of positive earnings in the target and nontarget groups12. Again, e mean of the target subsample is clearly lower an at of the nontarget, but a substantial proportion of the target group can expect to receivthe earnings above the population mean and a large proportion of the nontarget group receives very littlthe income13.


4.2.      Simulating counterfactual distributions

We next use e estimated earnings-schooling-height relationships to simulate e longer-term effects of the Oportunidades program on labor income. Figures 2 and 3 illustrate the nonlinearities in the density of non-zero labor income, conditional on schooling attainment and height; Figure 2 graphs e conditional density for men and Figure 3 for women. It is evident from the figures at higher levels of schooling attainment are associated with relatively larger increases in marginal earnings. The marginal earnings benefit is more homogeneous with respect to increases in height, but ere also appears to be somthe nonlinearity near e upper end of the height distribution.



Tables 5a and 5b show the results of our main simulation experiments for men and women. The first column displays characteristics of the income distribution wiout any program impacts. is income distribution is equal to the original income distribution with the addition of a small amount of error introduced by the nonparametric smooing. The other columns of Table 5a and 5b each represent a different set of hypoetical program impacts, given by (a)-(e), where we give e stated program impact to each individual in the target group and calculate e implied income distribution for e combined target and non-target groups. For example, case (a) augments each individual's education level by 0.6 years. We use the nonparametric simulation method described above to simulate a coun-terfactual earnings distribution whose features can be compared to the original no-program earnings distribution. As previously noted, we simulate changes in employment along wi changes in the distribution of positive earnings. at is, the earnings distribution includes a mass point at zero for nonworkers and e fraction of nonworkers can be affected by the program. Monly earnings are measured in ousands of pesos.

Table 5a indicates at the program would not significantly affect e fraction of men participating in e labor market, which remains around 90% across all the simulations. Also, impacts (a)-(c) have modest effects on mean earnings for men and almost no effect on earnings inequality, regardless of the measure. The effect of a 0.6 year impact on schooling attainment (in columns (b) and (c)) is larger for women an it is for men; however, the changes in income inequality arthe relatively minor for both men and women. e hypoetical three year increase in schooling attainment, shown in column (d), leads to substantially higher mean earnings and a reduction in poverty as measured by the Headcount ratio and the averagthe poverty gap. Whilthe income inequality actually increases slightly for men, it declines somewhat for women due to the large induced increase in female employment. A one cm increase in height leads to about a 30 peso increase in mean monly earnings for men but no substantial difference for women. The height impact has almost no influence on earnings inequality, but a large hypoetical increase in height of three centimeters slightly increases mean earnings and inequality, wiout having much effect on poverty.

Even ough Oportunidades significantly increased the human capital of children from disadvantaged families and substantially raised mean earnings, we have found its effects on earnings inequality to be minimal for two main reasons. First, returns to schooling in is environment are highly non-linear and in particular, we observe increasing returns at higher schooling levels. Those individuals in the target group at would have higher educational at-tainment in e absence of the program experiencthe relatively larger increases in income as a result of the program, so it is not e case at the poorest of the target group experience the largest benefit. The second factor at dampens the program's effect on inequality is at targeting children from poor backgrounds only imperfectly targets future low-earning adults, because of substantial inter-generational mobility.

We measure e influence of the nonlinearity in returns to schooling by thestimating and simulating parametric models of the employment and earnings processes and comparing ese results to ose found in our nonparametric simulations. Tables 6a and 6b present estimated coefficients for a probit model of employment and a linear regression model of log earnings. Each model contains a linear term for years of schooling and quadratics for height and potential experience. To simulate the employment process, we augment schooling attainment, height, and/or experience under e samthe program scenarios evaluated above and predict employment using draws from the probit error distribution at are consistent wi the observed choices. To simulatthe earnings, we make the same augmentations to e human capital variables and for each worker incorporate e original earnings residual if it was observed and draw from the earnings residual distribution for ose who were not working in e original sample. These simulation results are shown in Tables 7a and 7b.



Because there is near universal employment of men across e human capital distribution we find littlthe effect of schooling and height on employment wi a small positive effect of experience. The story is quitthe different for women where a year of schooling has a strong and significant positive effect on the probability of employment. When schooling is constrained to have a linear effect on log earnings, a year of schooling increases earnings by 8.7% for men and 15.4% for women. e linear and quadratic terms for height are jointly significant (a < 0.05 for both men and women while e experience terms are only jointly significant for women.

A comparison of Table 7a with Table 5a shows at for men, the parametric simulation approach tends to predict small reductions in inequality relative to e small increases in inequality predicted by the nonparametric approach. These differences are almost entirely due to the fact at the parametric model constrains log earnings to be a linear function of schooling and does not capture the fact at schooling has increasing returns. For women (Tables 5b and 7b) the parametric model predicts a smaller reduction in inequality because of the differences in how schooling affects employment. In particular, the nonparametric model predicts at targeted increases in schooling will increase women's employment more an a parametric model at includes schooling as a linear term. This difference in the effect on employment overpowers e impact of imposing constant returns to schooling in the earnings process.

The second major factor explaining Oportunidades' modest effect on inequality is at it targets children from poor families and esthe children are not necessarily e future poor adults. That is, the program increases the completed schooling of some children from an already high level to an even higher level. To explore the importance of targeting, we performed another set of simulations using our nonparametric earnings model where we target the same fraction of individuals wi the program but choose em on the basis of low adult education levels. Specifically, we give the program to those who would other wise form e bottom of the education distribution. This targeting is of course not feasible in practice, because it is impossible to know which children would eventually complete the least amount of schooling. Nevereless, the simulation results reported in Tables 8a and 8b give an upper bound for improving earnings and inequality rough morthe precise targeting. A comparison wi Tables 5a and 5b shows at targeting individuals at the bottom of e education distribution would be morthe effective in reducing inequality an e current targeting mechanism, but at the cost of lower mean earnings, because it does not take advantage of e larger program impacts at higher schooling levels.

In addition to e simulations we report, which assume constant treatment effects, we also carried out all the simulations under alternative scenarios of heterogeneous program effects. For example, we assigned half e target population impacts at were twice as high and half zero impact, keeping e average treatment effect the same. e half of e target group at received e double impact was alternatively chosen to be e less or more advantaged subgroup. Our findings wi regard to effects on mean earnings and earnings inequality under the heterogeneous treatment impact simulations were very similar to osthe discussed previously, so we omit em for e sake of brevity14.

5.      CONCLUSIONS

The Oportunidades program aims to reducthe poverty of the current generation rough transfers and to alleviate poverty of e next generation rough human capital investment. A number of experimental and nonexperimental evaluation studies have documented at the program significantly improves schooling attainment, health and nutrition over e short-term. is paper develops and applies a nonparametric simulation method for e purpose of studying how increases in schooling attainment and in height, as a measure of long-term nutritional status, will affect the distribution of earnings in e next generation.

Our empirical findings suggest at the human capital investment in today's you will increase their mean earnings levels, but will have only a modest effect on earnings inequality. Behrman (2006) comes to a similar conclusion in a survey of human capital policies and from an empirical study of how increasing education affects earnings inequality in Chile. The key factors underlying e modest effects on inequality at wthe observe are the difficulty in predicting which children will become future low earning adults and nonlinearities in how health and education arthe priced in e labor market. With regard to the first factor, childhood poverty is a strong predictor of future low earnings, but there is also substantial intergenerational mobility at makes it difficult to target low adult earners on the basis of childhood characteristics. With regard to e second factor, we found evidence of important nonlinearities in how height and education influence earnings. Most notably, an additional year of secondary school has a higher monetary return an an additional year of primary school. Because of these nonlinearities, people at the upper deciles of the targeted population tend to benefit more from the program intervention.

We conclude by considering some limitations of e simulation method studied in is paper. First, e method assumes at the observed relationship between earnings and e covariates of education, height, and work experience is causal. is raises concern about potential bias due to unobserved ability, which is e subject of a large labor economics literature. Previous attempts to control for ability bias have relied mainly on instrumental variables or natural experiments (e.g. twins wi different levels of schooling).15 Alough there is variation in reported estimates, most estimates of e rate of return to schooling at purport to control for ability bias through the use of instrumental variables exceed those obtained by ordinary least squares. The variation in estimates is partly accounted for by heterogeneity in returns to education on earnings at requires a LATE (local average treatment effect) interpretation of the instrumental variables estimates.16 Estimates at account for ability bias using variation in twin pairs, onth e other hand, tend to be somewhat lower an cross-sectional OLS estimates. Because the literature finds at OLS estimates do not necessarily overstate the causal effect, we have no reason to believe at our nonparametric procedure necessarily overstates the true return to schooling attainment. Also, much of the instrumental variables literature operates wiin a parametric framework and does not easily allow for the nonlinearities in the earnings-schooling-height-experience relationship at we find to be quantitatively important. Nevereless, furer exploration of how e simulation method could be modified to account for unobserved ability and endogenous covariates would be useful.

A second critical assumption of the simulation method is the usual synetic cohort assumption, namely at the characteristics of today's 25 to 40 year olds, observed in 2002, are representative of the future adulood of today's children. Extrapolating from current trends, children today would likely attain more education an current 25 to 40 year olds in the absence of the program intervention. Our estimates indicate at the marginal effect of education on earnings is increasing in years of education, so overall rising education levels could lead e simulation to understate somewhat the impact of Oportunidades on earnings. ird, e simulation method does not account for the general equilibrium effects of increasing the education levels of a large segment of e future labor force, which would tend to decrease returns to education. Any decline, ough, is at least partially mitigated by the fact at Mexico is an open economy. Four, is study focused on individual level earnings for men and women, alough household-level earnings inequality may be more relevant to policy makers. It is also not clear how to interpret high income inequality in a group (like women) where a large proportion choose not to work, because they have a partner who provides enough money for the household. The simulation method could be extended to model household formation by incorporating a marriage outcome, where marriage opportunities and outcomes potentially also depend on variables influenced by the program. Our method could similarly be extended to account for the influence of improving human capital on internal and external migration.

Lastly, improvements in futurthe earnings are only one of the long-term benefits expected from the program. For example, there is a substantial literature documenting how upgrading mother's education increases child test scores (e.g., Rosenzweig and Wolpin, 1994). Female program beneficiaries who choose not to work may be morthe effective mother's and may choose to have fewer children and to invest more in them.The simulation methodology in is paper could conceivably be extended to examine changes in fertility.

 

NOTES

*       This paper was presented at the 2008 UNDP (United Nations Development Program) Conference on Inequality in Latin America and at the 2008 annual meeting of the Population Association of America. We thank Jere Behrman and Estela Rivero-Fuentes for helpful comments.

1       Some similar programs also have been introduced in Asian countries, such as Bangladesh and Pakistan.

2       See, e.g., Schultz (2000,2004), Gertler (2000), Behrman, Sengupta and Eodd (2005), Parker and Skoufias (2000), Buddelmeyer and Skoufias (2003), Eodd and Wolpin (2006) and Freije, Bando and Arce (2006).

3       See, e.g., Freije, Bando and Arce (2006) and Bourguignon, Ferrtheira and Leite (2003).

4       ax → 0, ay 0, as n → ∞ and ayaxn → ∞.

5       The MxFLS data contain information on recent labor histories, but these are not long enough to construct a measure of actual experience. For this reason, we use the standard Mincer potential experience measure: Age minus years of schooling minus 6.

6       This assumption rules out spillover effects of the program onto the nontargeted population. See Angelucci et al. (2008) for a discussion of potential spillover effects of Oportunidades.

7       In 2002 the average daily exchange rate was 1 USD equals 9.68 pesos. Because a small number of the the earnings values seemed to be outliers, we implemented a trimming procedure and omitted all individuals who reported income higher than 40,000 pesos/ month. This corresponded to 9 of 5,180 observations or the top 0.2%.

8       As a point of reference, most developed European nations tend to have Gini coefficients for household income between 0.24 and 0.36. For household income, the United States Gini coefficient is around 0.45 and for Mexico is 0.55 (in 2003).

9       See Behrman and Hoddinott (2005) for discussion of e impacts of PROGRESA on height, and Schultz (2000, 2004), Behrman, Sengupta and Todd (2005) and Todd and Wolpin (2006) for discussion of impacts on years of schooling.

10       The participation model is estimated only for children in rural and semi-urban areas, because in 2002 (the time of our data collection) the program had not been significantly extended to urban areas. The data contain information pertaining to interviews with the parents of 1,910 children age 9-12 in rural areas. After dropping observations with missing variables, we are left with 1,699 observations.

11       The MxFLS data do not include years of actual labor market experience.

12       The target density has been scaled by a factor of 0.4 and the non-target by a factor of 0.6 so that togeer ey add up to equal e total population density.

13       The fraction of men receiving no labor income differs very little between the target (11.4%) and nontarget (10.6%) groups, but the difference is actually quite large among women where 11% of targeted women receive no labor income compared to 58% of the non-target group.

14       The estimates are available from e auors on request.

15       e.g., Behrman, Rosenzweig and Taubman (1994), Ashenfelter and Krueger (1994), Ashenfelter and Rouse (1998), Card (1995, 1999).

16       Card (1999, 2001).

 

REFERENCES

Angelucci, M., G. De Giorgi, M. Rangel, and I. Rasul (2008). "Family Networks and School Enrollment: Evidence from a Randomized Social Experiment", Working Paper.         [ Links ]

Ashenfelter, O. and A. B. Krueger (1994). "Estimates of the Economic Return to Schooling from a New Sample of Twins", Quarterly Journal of Economics, 84 (5): 1157-73.         [ Links ]

Ashenfelter, O. and C. Rouse (1998). "Income, Schooling and Ability: Evidence from a New Sample of Identical Twins", Quarterly Journal of Economics, 113 (1): 253-84.         [ Links ]

Behrman, J. and J. Hoddinott (2005). "Programme Evaluation with Unobserved Heterogeneity and Selective Implementation: the Mexican PROGRESA Impact on Child Nutrition", Oxford Bulletin of Economics and Statistics,Vol. 67, N° 4, pp. 547-569.         [ Links ]

Behrman, J., P. Sengupta and P. Todd (2004). "Progressing rough PROGRESA: An Impact Assessment of a School Subsidy Experiment", Philadelphia: University of Pennsylvania, forcoming in Economic Development and Cultural Change.         [ Links ]

Behrman, J. and E. Skoufias (2004). "Evaluation of PROGRESA/Oportunidades: Mexico's Anti-Poverty and Human Resource Investment Program", in Jere R. Behrman, Douglas Massey and Magaly Sanchez R, eds., The Social Consequences of Structural Adjustment in Latin America, book manuscript.         [ Links ]

Behrman, J. (2006). "How Much Might Human Capital policies Affect Earnings Inequalities and Poverty?", Working Paper.         [ Links ]

Behrman, J., M. Rosenzweig and P. Taubman (1994). "Endowments and e Allocation of Schooling in e Family and in e Marriage Market: e Twins Experiment", in Journal of Political Economy, volume 102, N° 6, 1131-1174.         [ Links ]

Bourguignon, F., F. Ferrtheira and P. Leite (2003). "Conditional Cash Transfers, Schooling and Child Labor: Micro-Simulating Brazil's Bolsa Escola Program". World Bank Economic Review 17 (2): 229-54.         [ Links ]

Buddelmeyer, H., and E. Skoufias (2003). "An Evaluation of the Performance of Regression Discontinuity Design on PROGRESA", IZA Discussion Paper N° 827 (July), Institute for e Study of Labor (IZA), Bonn, Germany.         [ Links ]

Card, D. (1999). "The Causal Effect of Education on Earnings". In Handbook of Labor Economics, Vol. 3A, edited by Orley Ashenfelter and David Card. (Amsterdam: Nor Holland).         [ Links ]

Card, D. (2001). "Estimating e Return to Schooling: Progress on Some Persistent Econometric Problems", Econometrica 69 (5), 1127-1160.         [ Links ]

Chakravarty, S. (1990). Ethical Social Index Numbers. New York: Springer-Verlag.         [ Links ]

DiNardo, J., N. Fortin, T. Lemieux (1996). "Labor Market Institutions and the distribution of Wages, 1973-1992: A Semiparametric Approach", Econometrica, Vol. 64, N° 5, pp. 1001-1044.         [ Links ]

Foster, J., J. Greer and E. orbecke (1984). "A Class of Decomposablthe poverty Measures", in Econometrica, Vol. 52, N° 3, pp. 761-766.         [ Links ]

Freije, S., R. Bando and F. Arce (2006). "Conditional Transfers, Labor Supply, and Poverty: Microsimulating Oportunidades", in Economia, Vol. 7, Issue 1, 73-124.         [ Links ]

Gertler, P. (2000). "Final Report: e Impact of PROGRESA on health", International Food Policy Research Institute, Washington, D.C.         [ Links ]

Lopez-Acevedo, G. (2004). "Mexico: Evolution of Earnings Inequality and Rates of Returns to Education (1988-2002)", Work Bank Report N° 19945.         [ Links ]

Parker, S. and E. Skoufias, 2000, "The impact of PROGRESA on work, leisure and time allocation", October. Report submitted to PROGRESA. International Food Policy Research Institute, Washington, D.C. <http://www.ifpri.org/emes/progresa.htm>         [ Links ]

Persico, N., A. Postlewaite, and D. Silverman (2004). "the effect of Adolescent Experience on Labor Market Outcomes: e Case of Height", Journal of Political Economy.         [ Links ]

Rosenzweig, M. and K. Wolpin (1994). "Are erthe increasing Returns to the intergenerational Production of Human Capital? Maternal Schooling and Child Intellectual Achievement", The Journal of Human resources, Vol. 29, N° 2, Special Issue: Women's Work, Wages, and Well-Being (Spring, 1994), pp. 670-693.         [ Links ]

Schultz, T. P. (2000). "Impact of PROGRESA on school attendance rates in the sampled population", February. Report submitted to PROGRESA. International Food Policy Research Institute, Washington, D.C.         [ Links ]

Schultz, T. P. (2004). "School subsidies for e poor: Evaluating a Mexican strategy for reducing poverty", Journal of Development Economics, (Revision of June 2000 Report submitted to PROGRESA. International Food Policy Research Institute, Washington, D.C. <http://www.ifpri.org/themes/progresa.htm>         [ Links ]

Silverman, B. (1986). Density Estimation for Statistics and Data Analysis. Chapman & Hall/CRC.         [ Links ]

Skoufias, E. and B. McClafferty (2001). "Is PROGRESA working? Summary of the results of an Evaluation by IFPRI", Report submitted to PROGRESA. Washington, D.C.: International Food Policy Research Institute, <http://www.ifpri.org/emes/progresa.htm>         [ Links ]

Strauss, J. (1986). "Does Better Nutrition Raise Farm Productivity?" Journal of Political Economy, 94, April, p. 297-320.         [ Links ]

Strauss, J. and Duncan T. (1998). "Health, Nutrition and Economic Development", Journal of Economic Literature, Vol. 36 (2), 766-817.         [ Links ]

Todd, P. and K. Wolpin (2006). "Assessing e Impact of a School Subsidy Program in Mexico", with Kenne I. Wolpin, American Economic Review, 2006, December.         [ Links ]

Todd, P. (2004). "Technical Note on Using Matching Estimators to Evaluate the Oportunidades Program For Six Year Follow-up Evaluation of Oportunidades in Rural Areas", Philadelphia: University of Pennsylvania, mimeo.         [ Links ]

 

APPENDIX A

This appendix describes how each of the variables for the empirical analysis was constructed. The data analysis has ree parts. First, we estimate a probability of participating in the Oportunidades program and use the estimated model to simulatthe program targeting for men and women between age 25 and 40. Second, we estimatthe nonparametrically the relationship between income, education, height, and work experience for men and women between age 25 and 40. ird, we compute the counterfactual income distribution under assumptions of how the program affects education, height, and work experience at are consistent wi recent evaluations of short-term program impacts.

Sample Construction

The initial sample of MxFLS respondents between age 25 and 40 contains 6,564 observations. When we drop the individuals who worked but did not report their income, the number goes down to 5,871. It drops furer to 5,180 (79% of e original sample) when we drop those individuals who did not report their education or whose height was not measured. Finally, we drop an additional 9 outlier observations for individuals who report receiving more an 40,000 pesos in the previous mon. is leaves a final sample size of 5,171.

Construction of Variables

Income. Income is measured as total labor income earned (including net profits for e self-employed) in the previous mon. It includes zeros for those individuals who don't work. About 6% of individuals who reported working in the previous week are recorded as being "peasants on their plot". 40% of these individuals report zero income in e last mon. This seems plausible for subsistence farmers. Only 2% of other individuals who report working report zero income. Income is measured in ousands of pesos and in 2002 the average daily exchange rate was 1 USD = 9.68 pesos.

We do not use proxy reports on income, because it is not clear how to combine is data with the first-person reports and weight the data correctly. the proxy reports also have more missing data.

Schooling. The MxFLS collects e type of e last school attended and, for most individuals, the number of years at the individual completed at that level. We do not include years of "technical education" in our measure, because wage returns to technical education (based on our own linear regressions) are much lower an the returns of conventional schooling.

Height. Height is not self-reported but instead is measured by trained survey personnel.

Experience MxFLS did not collect information on actual labor force experience, so we use the standard Mincer measure of potential experience equal to age minus years of schooling minus six.

Creative Commons License All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License