Sustaining Impacts When Transfers End: The Role of Women Leaders for Aspirations and Human Capital Investments Karen Macours* and Renos Vakis** September 17, 2016 Abstract Conditional cash transfer (CCT) programs have been shown to change households’ investments in their young children, but there are many open questions on how such changes can be sustained after transfers end. This paper analyses the medium-term impacts of a productive CCT and focuses on the role of interactions with local female leaders in sustaining poor households’ investment response after the end of the program. The causal effect of interactions with leaders is identified through the randomized assignment of leaders and other beneficiaries to different interventions aimed at increasing human capital and productive investments. Interactions with leaders that received the largest package were found to augment program impacts on households’ investments in education and nutrition, and to affect households’ attitudes towards the future during the intervention. This paper shows that the strong multiplier effects from leaders’ treatment persisted two years after the end of the program. Households randomly exposed to successful female leaders sustained higher investments in human capital and reported higher aspirations and expectations for the future of their children. These results suggest that program design features that help enhance ownership of a program’s objectives by local leaders may be key design element that can shift norms and sustain higher levels of human capital investments. Keywords: social interactions, leaders, human capital investments, randomized trial, policy evaluation, aspirations, productive transfers * Paris School of Economics and INRA; ** World Bank We are grateful to the program team at the Ministerio de la Familia and in particular Carold Herrera and Teresa Suazo for their collaboration during the design of the impact evaluation, as well as the Centro de Investigación de Estudios Rurales y Urbanos de Nicaragua (in particular Veronica Aguilera, and Enoe Moncada) for excellent data collection. We are indebted to Ximena Del Carpio, Fernando Galeana, and Patrick Premand for countless contributions to the wider research project. Financial support for this research has been received from BASIS-AMA (under the USAID Agreement No. EDH-A- 00-06-0003-00 awarded to the Assets and Market Access Collaborative Research Support Program) and the World Bank (ESSD trust funds, the RRB grant, as well as the Government of the Netherlands through the BNPP program). The views expressed in this paper are those of the authors and do not necessarily reflect those of the World Bank or any of its affiliated organizations. All errors and omissions are our own. 1 1. Introduction Low levels of investment in human capital are often considered a key constraint for households to escape poverty. Many development interventions aim to increase the investment of poor households in the education and nutrition of their children. Conditional cash transfer programs in particular have been found to augment households’ investment in human capital in many settings (Fiszbein and Schady, 2009; Ganimian and Murnane, 2014). A key question is whether such impacts can go beyond the immediate impact of relieving liquidity constraints and result in a sustainable shift towards higher levels of investment in nutrition and education by the poor. Only a few studies have focused on whether the impacts on households’ human capital investments persist even after such programs end (Baird, McIntosh and Ozler, 2016; Macours, Schady and Vakis, 2012) and the evidence is mixed.1 Even less is known about the possible mechanisms underlying the potential persistence. Understanding the mechanisms is key in order to derive lessons regarding optimal design of new programs, or even to potentially adjust existing ones. For programs to have a persistent effect on households’ human capital investments, they need either to permanently lift existing liquidity constraints, or change the value households attribute to investments in education and nutrition when considering the trade-offs between different budget allocations. The latter may occur if the interventions increase the perceived returns to such investments by addressing information asymmetries or through shifting preferences. Nguyen (2008) and Jensen (2010) show that changes in the perceived returns to schooling through information can indeed lead to educational gains. Recent evidence also suggests the potential of external interventions to shift preferences by changing 1 There is a somewhat larger and growing literature on whether the impacts on human capital outcomes or overall welfare outcomes (as opposed to investments) persist on the longer run (Gertler, Martinez, and Rubino-Codina, 2012; Parker, Behrman and Todd, 2009, 2011; Barham, Macours, and Maluccio, 2013a, b; Macours, Premand and Vakis, 2013; Barrera-Osorio, Linden and Saavedra 2015; Filmer and Schady, 2015; Araujo, Bosch and Schday, 2016; see Molina-Millan et al. 2016 for an overview). 2 parental aspirations for their children (Beaman et al., 2012; Bernard et al., 2014).2 Understanding how to design interventions to maximize such shifts hence becomes an important policy question. Several design features of CCT programs could be contributing to shifts in investment behavior, even if there is little causal evidence on the importance of specific features. Many programs include heavy social marketing and conditionalities enforcing attendance to regular meetings in which the nutritional, health and educational objectives are discussed. To the extent that such messages get internalized, one could expect increased human capital investments to persist. Targeting the transfers to women in the household could lead to a shift in the gender norms regarding decision making within the household, that possibly persist once transfer stop. Many programs further assign specific roles to key women in the community to re-enforce the messages but causal evidence on their specific role is rare.3 This paper provides evidence on the key role of social interactions with such local leaders for the persistence of program impacts. It builds on Macours and Vakis (2014), where we showed that social interactions with successful female leaders substantially increased program impacts on nutritional and educational investments while the program was operating. We use data collected two years after the program ended, to show that these social multiplier effects persist to a remarkable degree. Two years after the transfers stopped, households who live in the proximity of successful leaders still show significantly higher investments in both education and nutrition of their children. We further extend our earlier work by analyzing the impact of leaders on a rich set of questions specifically collected to measure parental aspirations and expectations for their children. This allows showing that proximity to successful leaders also led to significant shifts in parental expectations and aspirations regarding their children’s future, 2 External interventions in addition can change aspirations of children’s themselves (Wydick, Glewwe and Rutledge, 2013) or aspirations of adults for themselves (Lybbert and Wydick, 2016). 3 The key role of these local female leaders has been recognized in several qualitative evaluations of CCT programs (Adato, 2000; Adato et al, 200). In Colombia, an independent ECD intervention specifically targeted the “madre voluntaries” of the CCT program in recognition of their local leadership role (Attanasio et al, 2015). 3 consistent with the persistent shift in human capital investments. As in the previous paper, we rely on the two-staged randomized design of a short-term transfer program in Nicaragua to identify the social interaction effects. The program combined conditional cash transfers (CCT) with interventions aimed at increasing households’ productive potential. Because it targeted the vast majority of households in each community and explicitly encouraged group formation, it provides a unique opportunity to analyse the role of social interactions. The experiment varied the nature and the size of the benefit packages households received, and as such creates random variation in whether beneficiaries live close to the leaders that received the largest package. Moreover leaders that received the largest package outperform other leaders in terms of economic outcomes, and have higher human capital investments than other beneficiaries. We analyse whether these successful examples of leaders affected human capital investments of other beneficiaries, taking advantage of the exogenous variation created by the random assignment of different packages to both leaders and other households. Macours and Vakis (2014) show that social interactions with nearby leaders positively affected human capital and productive investments as well as the future-oriented attitudes of other beneficiaries during the program. Interactions with leaders may have affected other households’ aspirations by setting good examples and sharing their experiences. The earlier work does not establish whether these shifts are sustainable, and a priori the answer is not obvious. Indeed, increasing aspirations in the presence of many other remaining constraints, may only lead to short term gains, and households could well quickly revert back to pre-program behaviour when the transfers stop. On the other hand, if interactions with successful leaders successfully changed norms and beliefs regarding human capital investments, the increased investment levels might persist even after the end of the program. Macours, Schady and Vakis (2012) show that the Nicaraguan CCT indeed had persistent effects on parental investments in early childhood. This paper shows that social interactions with local leaders were crucial for the persistence in the educational and nutritional investment. 4 Our results further show that interactions with leaders changed both beneficiaries’ aspirations and expectations about their children’s educational and occupational future, which can help explain the sustained higher levels of human capital investments. As such, the paper contributes to the developing literature in economics and the wider social sciences on the role and the formation of aspirations (Genicot and Ray, 2014; Besley, 2016). Appadurai (2004) and Ray (2006) argue that upward mobility might be difficult for the poor when they lack the capacity to aspire, i.e. when their own experiences and the experiences of those that are close to them suggest that escaping poverty is not a feasible option. Yet learning about the positive experiences from others that are sufficiently “close” can help open their aspiration window. Hence facilitating social interactions can be instrumental in changing aspirations and shaping positive attitudes to the future, and in turn lead to investments in children’s future.4 Empirical evidence of such mechanisms is rare due to the reflection problem, which is addressed in this paper through the double randomization. This paper hence contributes by providing empirical evidence on the role of social interactions in changing aspirations of the poor. More broadly it relates to recent empirical work on the potential of social interactions to shift norms and behaviour (Feigenberg, Field, and Pande, 2013; Paluck and Shepherd, 2012) and to the emerging literature about mental models and attitudinal changes (e.g. Jensen and Oster 2009, La Ferrara, Chong, and Duryea 2012, World Bank, 2014). Finally, by focusing on the local female leaders, the paper relates to the literature on female reservation for local leadership position in India (Chattopadhyay and Duflo, 2004; Beaman et al., 2009) and in particular to Beaman et al. (2012) who show that a law reserving leadership positions for women in India affected girls’ educational aspirations. 4 Appadurai (2004) describes how mobilization by social movements can expand the capacity to aspire, in part through regular social gatherings and sharing ideas and experiences about future-oriented activities among the poor. 5 The paper is organized as follows: in the next section we discuss the key features of the program and the relevance of social interactions. Section 3 discusses the data and the empirical strategy. Section 4 then shows that social interactions with successful leaders led to persistent impacts on other beneficiaries’ human capital investments and also shows impacts on parental expectations and aspirations. Section 5 concludes. 2. Program information and design 2.1. Program description and treatment packages5 The Atención a Crisis program was a one-year pilot program implemented in 2006 by the Ministry of the Family in Nicaragua. In the treatment communities, three different treatments were randomly allocated among 3000 eligible households. All selected households were eligible for the basic CCT, which included cash transfers conditional on children’s primary school and health service attendance. The transfers came with a strong social marketing message reinforcing the importance of investing in children’s education and in diversified nutrition. Take up of the CCT was 95%. In addition to the CCT, one third of the eligible households received a scholarship for a vocational training (with take-up of 89%). Another third of eligible households received, in addition to the basic CCT, a 200 US$ lump sum grant for productive investment aimed to develop a small non-agricultural business (with take-up of 99%). This treatment variation was perceived by the beneficiaries as the most attractive and as it involved the largest cash amount we call it the largest package. Given the high take-up rates, we henceforth refer to eligible households in treatment communities as beneficiaries. The program design aimed to change household’s investment behaviour through several mechanisms. The 5 More details about the program and its different components are provided in the online appendices of Macours, Schady and Vakis (2012) and Macours and Vakis (2014), as well as the following website: http://go.worldbank.org/VUYJAQ3UN0 6 level of transfers was substantial, ranging from 18 per cent of average annual household income for those receiving the basic CCT package to 34 per cent for those receiving the productive investment package. The conditionalities and social marketing on education, health and nutrition aimed at changing households’ perspectives about investment in long-term human capital. The program design also created many opportunities for enhanced interactions between beneficiaries. More than 90 per cent of the households in treatment communities were eligible for the program, increasing the opportunities for information sharing and interactions, possibly resulting in higher motivation and program ownership. Program participants were also required to participate in a number of local events ranging from discussions on nutrition and health practices to workshops on the importance of education, business development and labour market skills. The program put in place a system of volunteer local promotoras to further enhance information flows and compliance with program requirements. The promotoras met frequently with a small groups of (about 10) beneficiary women to talk about these requirements and the program’s objectives. As such, the program created a lot of new leadership positions for women in these communities.6 Women self-selected for these positions, but subsequently were randomly allocated to one of the three program packages (see below). Qualitative interviews during and after the intervention showed that most of the promotoras had taken strong ownership of the messages and objectives of the program, and were committed to convey and remind other beneficiaries that the purpose of the cash transfers was to invest in the nutrition and education of their children. 2.2. Program randomization The program targeted 6 municipalities in the Northwest of Nicaragua, and a first lottery randomly selected 56 intervention and 50 control communities. Baseline data were used to define household 6 Before the program, leadership positions for women were limited mostly to positions as teachers and health coordinators. 7 program eligibility using proxy means methods for both treatment and control.7 In the treatment communities, the main female caregiver from each eligible household was then invited to a registration assembly. If there were more than 30 eligible households in a community, several assemblies were organized at the same time, and households were assigned to one of the assemblies based on the geographic location of their house. In total, there were 134 assemblies (hence on average 2.4 per community). During the assemblies, the program objectives and its various components were explained and women were asked to volunteer for the promotora positions. Volunteers were approved by the assembly and each promotora was assigned a group of approximately 10 beneficiaries living close to her, based on a joint decision. At the very end of each assembly, all the beneficiaries - including the promotoras - participated in a second lottery process through which the three packages described above were randomly allocated among the beneficiaries, with each of the three packages assigned to one-third of households in the treatment communities. As a result of the two lotteries, households hence were randomly assigned to the control group (in the control communities), or to one of three packages: the CCT, the CCT plus training, or the CCT plus productive investment grant (the largest package). Since promotoras and other existing female leaders in the treatment communities were randomly allocated to one of the three treatment groups, all other beneficiary households will be randomly exposed to leaders with different treatment packages. In particular, as there are on average four leaders in each assembly, some beneficiaries will be randomly exposed to several leaders that got the largest package, while others may not have any leaders with that package. This is the main exogenous variation that we exploit in this paper. 3. Data and empirical strategy 7 As more than 90% of all households were eligible, the analysis in this paper is limited to the eligible households. 8 In treatment communities, data were collected from all households. In control communities, a random sample of households was selected at baseline so that the control group was of equal size as each of the three intervention groups (1000 households). The data analysed in this paper was collected between August 2008 and May 2009, approximately two years after the last transfer. Individuals that had migrated out of the area were tracked to different locations in Nicaragua resulting in a very low attrition rate (3 per cent a the household level), which is uncorrelated with treatment. The survey instrument was modeled after the Nicaraguan Living Standard Measurement Survey (LSMS), with modules on education, health, and detailed household expenditures, among others. To measure educational investment we use an indicator of whether the child was attending school, the number of days the child has been absent from school in the last month, and the amount spent on school expenditures since the start of the academic year. Nutrition investment is measured by the share of food expenditures for animal products and for vegetables and fruits, reflecting the emphasis of the program about the importance of such nutrients for children. The education and nutrition investment indicators are the same as those used in Macours and Vakis (2014). For child level outcomes, we consider all children between 7 and 18 years olds, to facilitate comparability with the earlier results. A specific module was added in 2008 to ask for mothers’ expectations and aspirations for all children between 7 and 15 years old. Mothers were asked both what they desired and what they realistically expected for their children in terms of final educational attainment, occupation, and future monthly earnings. To proxy for future living standards, we also ask for the number of rooms in the house the mothers desired and expected for their children in 30 years time. For occupation, we consider two possible definitions, and define a dummy indicating whether the mother expected or desired a professional job for the child, or alternatively a professional or other skilled salary job. For monthly earnings, and taking into account the highly skewed nature of the distribution of this variable, we follow Athey and Imbens (2016) and use an indicator of the rank in the earnings distribution. We also show a 9 second measure with monthly earnings winsorized at 95%. Finally, to account for multiple hypotheses testing, we use an aggregate indicator for both aspirations and expectations, which is the average of the standardized measures for educational attainment, occupation, monthly earnings and living standard (number of rooms in the house), following Kling et al. (2007).8 All standardized measures were obtained by subtracting the mean and dividing by the standard deviation of the control group. Identification relies on the randomized allocation of beneficiaries to one of the three program packages or the control, and the random allocation of these same packages among leaders. We consider both the leadership positions created in the treatment communities by the program (the promotoras) and other women with leadership positions because they are not mutually exclusive (many health coordinators and teachers volunteered to be promotoras). Female leaders tend to be younger and more educated than the average female beneficiary. While beneficiaries on average have completed 3 years of education, leaders have on average 5 years. Other indicators of socio-economic status at baseline are relatively similar between leaders and non-leaders. In Macours and Vakis (2014) we show that the randomization worked well and that the short-term returns to the largest package for the leaders were higher than for the other beneficiaries. During the intervention, leaders with the largest package also had higher non-agricultural and total income than leaders with other packages, reflecting the additional cash they had received to start new activities. As the income level and the income sources of these leaders at baseline were similar to those of the other beneficiaries, it seems plausible that beneficiaries could identify with their success during the program and that this might have motivated and inspired them. 8 As an important share of parents desire professional jobs for their children, but few expect their children to get such jobs, we use the dummy for professional job in the aggregate index for aspirations, and the dummy for professional or skilled wage job in the aggregate index for expectations. 10 The largest package is also the intervention that created sustainable gains in income and consumption levels two years after the end of the program (Macours, Schady and Vakis, 2012; Macours, Premand and Vakis, 2013). Focusing on the outcomes of leaders, Table 1 shows that the leaders that received this package continue to stand out. Two years after the end of the intervention, leaders with the largest package still have higher incomes from non-agricultural self-employment than other leaders. And their nonagricultural income and total income is significantly higher than for other beneficiaries who received the same package, even if their income from agricultural wages is lower. This suggests they may have been better in maintaining their new commercial activities and likely continue to be seen as successful leaders in the community. Interestingly, we observe the same patterns when considering answers to questions about parents’ expectations for their children’s future. Expectations of leaders who received the largest package are significantly higher than for other leaders on a number of dimensions. They notable expect their children to achieve higher schooling levels and earn higher wages, and are 11 percentage points more likely to expect their children to become a professional or skilled salary earner. There are also large differences in the expectations of these leaders and other beneficiaries who received the same package, with leaders expecting their children to obtain 1.5 years more education, and 21 percentage points more likely to become a professional or skilled salary earner. While these latter differences are consistent with the program identifying natural leaders through self-selection, they also imply that these leaders may be seen as local success stories - both in current achievement and in their attitudes towards the future - that others could aspire to emulate. A similar pattern is found for differences between leaders and others in their reported aspirations for their children, even if differences in aspirations are smaller than differences in expectations. Comparing mean values between the indicators of aspirations and expectations further shows large gaps between the two sets of outcomes, with expectations for educational attainment, for instance, 5 years less than aspirations 11 and similarly large differences for earnings, occupation and living standards. Interestingly, these gaps are smaller for leaders than non-leaders. The patterns suggests that both leaders and other households internalize their constraints when reporting their expectations, but they also confirm a capacity to aspire a much better life for their children. Finally, and in line with the other results, leaders’ investments in education and nutrition of their children are higher than those of others beneficiaries. The significant differences in human capital investment between leaders and non-leaders in Table 1 mirror similar findings from the baseline and the midline survey (Macours and Vakis, 2014). Leaders with the largest package hence continued to provide positive examples for others, in line with the program objectives and this two years after the transfers ended. We therefore focus on whether higher exposure to those leaders continued to affect education and nutrition investments of other beneficiaries. Specifically, we calculate the share of leaders that was randomly allocated to the largest package in each registration assembly. The average numbers of leaders in an assembly is four, so that there is substantial variation in the share of leaders that got the largest package in an assembly. There is much less variation in the share of other beneficiaries that got the largest package as the number of households in each assembly was relatively large and the number of leaders is small, so that the share of non-leaders with the largest package in each assembly is close to one-third in all cases. Our general specification is: Yia = δ0+δ1Tia + δ2(Tia* Sa) + δ3 Sa +εia (1) where Yia is an outcome indicator for eligible household (or a child of household) i who was invited in assembly a, Tia is assignment of i to any of the three treatment groups, and Sa is the share of leaders (over all leaders in the assembly) that randomly received the largest package in i’s registration assembly. Given that households were invited to particular assemblies based on geographic proximity, Sa will capture the 12 share of leaders with the largest package that live in the proximity of i.9 Since Sa is always 0 in the control communities, and since all eligible households in the treatment communities receive one of the three intervention packages, the term δ3 Sa cancels out of the estimation. The coefficients of interest are δ1 and δ2. A finding, for example, that δ1 and δ2 are both positive would imply that while assignment to the treatment group increases the outcome of interest (δ1), there is an additional impact of the program that comes from the social interactions (δ2). We also explore how the share of leaders with the largest package affects impacts for beneficiaries of each of the three packages separately. All regressions are estimated on the sample of eligible households (or their children) that are not leaders themselves. 4. Social interaction effects on human capital investments 4.1 Main results We first pool households across treatment packages and investigate whether there is a general relationship between program impacts and the proximity to leaders who received the largest package. Table 2 presents the results for 2008, the main focus of this paper, in the top panel. In the bottom panel, the table reports the findings for 2006 from our earlier work for comparison. The interaction terms in top panel of Table 2 suggest that social interactions are crucial to sustain program impacts on education and nutrition investments after the end of the intervention. Indeed the estimates indicate that there are no significant sustained impacts on human capital investments when no leader was assigned the largest package, in contrast with the findings during program implementation (bottom panel). 9 Location of one’s house might be endogenous, and people living in the proximity of leaders might also be more likely to be their family members, or otherwise have similar characteristics. The identification in this paper does not depend however on the proximity to the leader per se, but instead it depends on the random allocation of certain packages to those leaders. 13 Still, the interaction terms suggests that the higher the share of leaders with the largest package, the less likely children are absent in school and the more households invest in education, in animal proteins and in fruit and vegetables. The multiplier effects are not only statistically significant but also large. For example, school expenditures increase with 49% when all the leaders in one’s assembly got the largest package, while school absences decline with 21%. Strikingly, the magnitude of the multiplier effects two years after the end of the program are similar, if not larger, than those found while the intervention was in place. Table 3 shows the social interaction impacts on human capital investments by treatment group. The effects are strongest for beneficiaries of the largest package. For instance, the school expenditures doubles for beneficiaries of the largest package in the extreme case that the share of female leaders with the same package changes from 0 to 1. The impacts are about half the size for the beneficiaries of the training packages (and even smaller for those with the basic package) for most outcomes and many of the interaction terms are not significant. Nevertheless, as in our earlier findings, the P-values indicate that we cannot reject that the social effects are different for the three groups for most variables. In fact, when pooling the basic and the training packages together, the interaction effects for the school expenditures, and expenditures for fruit and vegetables are significant (not shown). Note that while the coefficients of the interaction effects are large, their interpretation needs to account for the fact that there are on average about 4 leaders in a registration assembly. The estimates hence indicate that having one additional leader with the largest package in one’s assembly reduces school absences with 0.4 days per month and increases school expenditures by about 16 percent. And for households who themselves have the largest package, one additional leader with the same package increases school attendance by 2.5 percentage points and increases school expenditures by 25 percent. These are not only large effects, but are similar or even higher than the effects found in 2006. Hence, 14 interactions with leaders had a remarkable persistent impact on other households’ investment behaviour, and those are particularly important for households who themselves received the largest package. 4.2. Robustness checks The results are robust to several alternative specifications.10 A first concern could be that the results are driven by extreme values in the independent variable. While the average share of leaders with the largest package is 0.33, for 95% of the observations, it is between 0 and 0.67. A first robustness check in Table 4 therefore excludes the observations with values above 0.67. This does not substantially alter any of the results, even if, as expected, the standard errors increase. The results are also robust to clustering the standard errors at the level of the registration assembly, as opposed to at the community-level, and to not excluding outliers. The next two specifications show that the results are further robust to controls for the total number of people in an assembly, or the total number of peers (defined as beneficiaries that are not leaders) in an assembly. Finally, the results remain generally robust when including a community fixed effect, with the exception of the food expenditures for animal products, even if the variation in the independent variable is reduced. Table 4 further shows alternative specifications using the number of leaders with the largest package instead of the share. These specifications also control for the total number of leaders in the registration assembly. The coefficient on the number of leaders with the largest package is consistent with the main results in terms of sign, size, and magnitude. We can then also compare the coefficient of the number of leaders with the largest package, with the coefficient of the number of peers with the largest package (last specification in Table 4). The results suggest that social interaction effects from peers might be more limited: the coefficients are generally not significant and smaller than the coefficients for the number of 10 Table 4 presents robustness checks for the beneficiaries with the productive investment package. Results pooling all beneficiaries are similarly robust. 15 leaders, with the exception of the expenditures for animal products. The coefficients for leaders and peers are significantly different for school attendance, absences and spending on fruit and vegetables. Note however that these results should be interpreted with caution, given that they could be driven by the fact that there is less variation to identify the social effects of peers. 4.3. Multiplier effects on aspirations and expectations While the identification strategy in our paper allows to clearly demonstrate the importance of the social interaction effects with local female leaders, it does not necessarily help to understand how exactly leaders might be influencing other households’ investments. Indeed, as with other papers on social learning, one can wonder whether the interaction effects reflect that other households mimic the behaviour of leaders or whether they reflect actual shifts in aspirations and expectations of non-leader households for the future of their children.11 We investigate this question by analysing the data regarding mothers’ expectations and aspirations regarding children’s final educational levels, future occupation, earnings and living standards. Table 5 shows results of the main specification for these outcomes, and also shows the spillovers on the educational level attained by 2008. We restrict the sample for this analysis to children less than 15 years old, as older children are more likely to already have reached their final education levels. We also exclude children less than 9, as they would not have benefitted directly from the educational component of the intervention. The top panel shows the impacts on the expectations mothers reported for their children, while the lower panel focuses on their aspirations. Table 5 shows first of all that the large spillover effects in investments found in table 3 are reflected in spillovers in educational attainment by 2008. Indeed 2 years after the end of the intervention, having one 11 One might also wonder whether the results are driven by more direct communication between leaders with the largest package and other beneficiaries. Two years after the intervention, this does not appear to be the case as we find no significant multiplier effects on the probability of talking to a leader, a teacher or health promotor. 16 additional leader in one’s registration assembly increases children’s school attainment with 0.2 years of schooling. Considering the spillovers on beliefs, we note that parents expect these gains to persist and potentially slightly increase in the future. Parents’ expectations about their children obtaining professional jobs or skilled salary jobs are also strongly affected by the proximity with successful leaders. Having one more leader in one’s registration assembly increases expectations of parents for their children to become (white-collar) professionals with almost 50% (starting from a very low level in the control group). Strong multiplier effects are also found for expectations regarding children’s future earnings and living standards. The social interaction effects for mothers’ aspirations follow a similar pattern, even if the magnitude for most indicators is lower. Indeed, averaging over the different indicators, we find that the difference between no exposure and full exposure to leaders with the largest package changes expectations regarding children’s future with 0.32 standard deviations, while it increases aspirations with 0.22 standard deviations. This difference could suggest that the social multiplier is instrumental in helping to narrow the aspiration gap.12 Overall, these findings show that interactions with female leaders changed beneficiaries’ expectations and aspirations regarding their children’s educational and occupational future, consistent with the sustained higher levels of human capital investments. 5. Conclusions Many development interventions aim, through a variety of mechanisms, to shift the investment behaviour of beneficiary households. Conditional cash transfer programs in particular have an implicit or explicit objective to change households’ attitudes and the social norms toward investment in the education, health 12 It is also possible however that measurement error for the aspiration variables is higher than for the expectation variables, as reporting on desired outcomes might be more difficult than reporting on expectations. If that is the case, the differences in effect size may reflect stronger attenuation bias for the aspiration results. 17 and nutrition of their children. When programs are designed to only last for a limited period, the sustainability of the impacts might crucially depend on whether changes in investment behaviour persist after the end of the program. Yet, the mechanisms through which such change in attitudes can be reached and reinforced are not always clear. This paper shows that social interactions with local female leaders can contribute to changing parental aspirations and expectations and lead to sustainable changes in educational and nutritional investment. The results suggest natural leaders living in people’s close proximity can be important vehicles for change by motivating and encouraging others and by providing examples that people aspire to follow. We find these effects for a program in which both female leaders and other female beneficiaries received sizable transfers and social effects are particularly large when leaders and beneficiaries received the same package. Hence the results do not suggest that interventions should be primarily targeted to such leaders, but rather that examples of positive experiences of nearby leaders can help open parents’ aspiration window and increase human capital investments when they are provided with resources to follow those examples. The evidence in this paper draws attention to the positive role local leaders can play, which contrasts with a common focus on the possible negative roles of leaders through elite capture in many policy discussions. It points in particular to the importance of assuring that development program designs leave room, and possibly enhance the role of natural leaders to help shift attitudes towards development objectives. More generally, the results have implications for the debate on the sustainability of using cash or asset transfer programs in low-income countries. The evidence in this paper suggests that designing such programs in ways that facilitate and encourage social interactions may be important to create sustainable change. 18 References Adato, Michele, 2000. Final report: The impact of Progresa on community social relationships International Food Policy Research Institute, Washington, DC. Adato, Michelle and Terry Roopnaraine, 2004. Final report: A Social Analysis of the Red de Proteccion Social (RPS) in Nicaragua, International Food Policy Research Institute, Washington, DC. Appadurai, Arjun, 2004. “The Capacity to Aspire”. In Rao, V., and M. Walton (eds.) Culture and Public Action. Palo Alto, CA: Stanford University Press. Araujo, Caridad, Mariano Bosch and Norbert Schady, 2016. “Can Cash Transfers Help Households Escape an Inter-Generational Poverty Trap?” Paper presented at “NBER conference on the The Economics of Asset Dynamics and Poverty Traps”, Washington DC. Athey, Susan and Guido W. Imbens, 2016. “The Econometrics of Randomized Epxeriments”, Banerjee, A. and E. Duflo, (eds.), Handbook of Economic Field Experiments. Volume 1. Elsevier. Attanasio, O., E. Fitzsimons, C. Fernandez, S. Grantham-McGregor, C. Meghir, and M. Rubio-Codina. “Using the infrastucture of a conditional cash transfer program to deliver a scalable integrated early child development program in Colombia: cluster randomised controlled trial.” British Medical Journal 249 (2014): 5785. Baird, Sarah, Craig McIntosh and Berk Ozler, 2016. “When the Money Runs Out: Do Cash Transfers have Sustained Effects?” mimeo. Barham, Tania, Karen Macours, and John A. Maluccio, 2013a. “Boys' Cognitive Skill Formation and Physical Growth: Long-term Experimental Evidence on Critical Ages for Early Childhood Interventions.” The American Economic Review Papers and Proceedings 103(3), 467-471. Barham, Tania, Karen Macours, and John A. Maluccio, 2013b. “More Schooling and More Learning?: Effects of a Three-year Conditional Cash Transfer Program in Nicaragua after 10 years.” IADB Working Paper Series IDB-WP-432. Barrera-Osorio, Felipe, Linden, Leigh L., and Saavedra, Juan E., 2015. “Long Term Educational Consequences of Alternative Conditional Cash Transfer Designs: Experimental Evidence from Colombia.” Unpublished manuscript. Beaman, L., Chattopadhyay, R., Duflo, E., Pande, R. and Topalova, P. (2009). ‘Powerful women: does exposure reduce prejudice’, Quarterly Journal of Economics, vol. 124(4), pp. 1497–540. Beaman, Lori, Esther Duflo, Rohini Pande and Petia Topalova, 2012. “Female Leadership Raises 19 Aspirations and Educational Attainment for Girls: A Policy Experiment in India,” Science, 355: 582-586. Behrman, Jere R., Susan W. Parker, and Petra E. Todd, (2009b). “Schooling Impacts of Conditional Cash Transfers on Young Children: Evidence from Mexico.” Economic Development and Cultural Change 57(3): 439–477. Behrman, Jere R., Susan W. Parker, and Petra E. Todd, (2011). “Do Conditional Cash Transfers for Schooling Generate Lasting Benefits? Five-year Follow-up of Progresa/Oportunidades.” Journal of Human Resources 46(1), 93-122. Bernard, Tanguy, Stefan Dercon, Kate Orkin, and Alemayehu Seyoum Taffesse, 2014. “The Future in Mind: Aspirations and Forward-Looking Behaviour in Rural Ethiopia“, CSAE Working Paper WPS/2014-16. Besley, Timothy, 2016. “Aspirations and the Political Economy of Inequality”, Oxford Economic Papers , forthcoming. Chattopadhyay, R. and Duflo, E. (2004). ‘Women as policy makers: evidence from a randomized policy experiment in India’, Econometrica, vol. 72(5), pp. 1409–43. Feigenberg, Benjamin, Erica Field, and Rohini Pande. 2013. “The Economic Returns to Social Interaction: Experimental Evidence from Microfinance.” Review of Economic Studies 80 (4): 1459–83. Filmer, Deon and Schady, Norbert, 2014. “The Medium-Term Effects of Scholarships in a Low-Income Country.” Journal of Human Resources, 49(3), 663–694. doi:10.1353/jhr.2014.002. Fiszbein, Ariel and Schady, Norbert R., 2009. “Conditional Cash Transfers: Reducing Present and Future Poverty.” World Bank Policy Research Report. Washington, DC, United States: 24, World Bank. Ganimian, Alejandro J., and Richard J. Murnane, 2014. “Improving Educational Outcomes in Developing Countries: Lessons from Rigorous Evaluations”, NBER Working Paper 20284. Genicot, Garance, and Debraj Ray, 2015. “Aspirations and Inequality”, NBER Working Paper No. 19976 Gertler, Paul, Martinez, Sebastian and Rubio-Codina, Marta 2012. "Investing Cash Transfers to Raise Long Term Living Standards." American Economic Journal: Applied Economics 4(1): 164–92. Kling, Jeffrey, Jeffrey Liebman, and Lawrence Katz. 2007. “Experimental Analysis of Neighborhood Effects.” Econometrica 75(1): 83-119. 20 Jensen, Robert, and Emily Oster. 2009. “The Power of TV: Cable Television and Women’s Status in India.” Quarterly Journal of Economics 124 (3): 1057–94. La Ferrara, Eliana, Alberto Chong, and Suzanne Duryea. 2012. “Soap Operas and Fertility: Evidence from Brazil.” American Economic Journal: Applied Economics 4 (4): 1–31. Lybert, Travis and Bruce Wydick, 2016. “Poverty, Aspirations, and the Economics of Hope: A Framework for Study with Preliminary Results from the Oaxaca Hope Project”, Paper presented at “NBER conference on the The Economics of Asset Dynamics and Poverty Traps.” Macours, Karen, Norbert Schady and Renos Vakis, 2012. “Cash Transfers, Behavioural Changes, and Cognitive Development in Early Childhood: Evidence from a Randomized Experiment” American Economic Journal: Applied Economics, 4(2): 247–273. Macours, Karen and Renos Vakis, 2014. “Changing Households’ Investment Behaviour through Social Interactions with Local Leaders: Evidence from a randomized transfer program”, Economic Journal 124 (May), 607–633. Macours, Karen, Patrick Premand and Renos Vakis, 2012. “Transfers, Diversification and Household Risk Strategies: Experimental Evidence with Lessons for Climate Change Adaptation.” CEPR Discussion Paper 8940. London, United Kingdom: Center for Economic Policy Research. Molina-Millan, Teresa, Tania Barham, Karen Macours, John A Maluccio, and Marco Stampini. 2016. “Long-Term Impacts of Conditional Cash Transfers in Latin America: Review of the Evidence” Inter-American Development Bank Technical Note No. IDB-N-923. Paluck, Elizabeth Levy, and Hana Shepherd. 2012. “The Salience of Social Referents: A Field Experiment on Collective Norms and Harassment Behavior in a School Social Network.” Journal of Personality and Social Psychology 103 (6): 899. Ray, Debraj, 2006. “Aspirations, Poverty and Economic Change”, in Banerjee, Abhijit V., Roland Benabou, and Dilip Mookherjee (eds.), Understanding Poverty, Oxford University Press, Oxford. World Bank, 2014. World Development Report 2015: Mind, Society and Behavior, Washington DC. Wydick, Bruce, Paul Glewwe and Laine Rutledge. 2013. "Does International Child Sponsorship Work? A Six-Country Study of Impacts on Adult Life Outcomes." Journal of Political Economy, 121(2), pp. 393-436 21 Table 1: Comparison of follow-up outcomes of leaders with largest package with other leaders and non-leaders Non- P-value P-value P-value Leader Leader Leader leader Leaders Leaders Leaders T3- T1 T2 T3 T3 T3-T1 T3-T2 NonleaderT3 Economic activities (in cordoba per capita) Income from non-agricultural self-employment 489.1 546 810 557 0.04** 0.09* 0.04** Income from commercial activities 190.9 156 404 222 0.05** 0.02** 0.05** Income from agricultural wages 602.8 749 679 973 0.51 0.55 0.01** Value animal stock 1630 2104 2191 1631 0.14 0.84 0.13 Total income 11707 12049 12272 10925 0.51 0.78 0.05* Expectations for children's future average index 0.15 0.17 0.37 0.00 0.01*** 0.02** 0.00*** expected years of education attained 9.72 9.82 10.39 8.64 0.06* 0.07* 0.00*** expected occupation: professional 0.03 0.04 0.04 0.05 0.43 0.77 0.90 expected occupation: professional or skilled empl. 0.35 0.36 0.47 0.26 0.04** 0.06* 0.00*** expected number of rooms in house 2.64 2.76 2.87 2.74 0.05* 0.40 0.28 expected monthly earnings 2132 2047 2332 1976 0.12 0.02** 0.00*** expected monthly earnings (winsorized 95%) 1703 1637 1898 1593 0.13 0.04** 0.00*** Aspirations for children's future average index 0.06 0.19 0.17 0.02 0.15 0.83 0.06* desired years of education attained 14.05 14.39 14.49 13.56 0.12 0.78 0.00*** desired occupation: professional 0.53 0.63 0.60 0.53 0.13 0.62 0.16 desired occupation: professional or skilled empl. 0.90 0.94 0.92 0.88 0.49 0.56 0.09* desired number of rooms in house 5.18 5.35 5.221 5.25 0.83 0.49 0.88 desired monthly earnings : rank 2140 2239 2273 2037 0.30 0.79 0.06* desired monthly earnings (winsorized 95%) 4532 4655 4831 4274 0.38 0.64 0.10* Human capital investment Attending school 0.863 0.84 0.82 0.77 0.49 0.96 0.00*** Number of days absent from school 4.329 5.16 5.71 6.35 0.44 0.70 0.02** School expenditures 767 683 636 518 0.51 0.67 0.00*** Share of food expenditures for animal products 0.17 0.17 0.18 0.16 0.37 0.97 0.07* Share of food expenditures for vegetables&fruit 0.07 0.07 0.08 0.07 0.47 0.51 0.16 Note: Sample includes intent-to-treat households in treatment communities. Economic outcomes and food expenditures are household level data. Data on education, expectations and aspirations are child-level data. Highest and lowest .5% outliers of income and expenditures data trimmed. Expectation and aspirations questions refer to children 9-15 years old. Education questions refer to children 7-18 years old. Average expectation index is average of standardized outcomes for expected years of education, professional or skilled employment, number of rooms in the house and monthly earnings rank. Average aspiration index is average of standardized outcomes for expected years of education, professional employment, number of rooms in the house and monthly earnings rank. Earnings ranks are calculated by converting the absolute monthly earnings to the rank in the earnings distribution, combining answers of leaders and non-leaders. All monetary values are in Cordoba (1 US$ =~ 20 Cordobas). P-values account for clustering at the community level. *** p<0.01, ** p<0.05, * p<0.1 22 Table 2: Social interaction effects on human capital investments Education Nutrition Attending Number of School Share of food Share of school days absent expenditures expenditures food (7-18 year from school (7-18 year for animal expenditures olds) (7-18 year olds) products for olds) vegetables and fruit 2008 Intent-to-treat* 0.045 -1.506* 310.9*** 0.0387** 0.0221*** share of leaders with largest package (0.040) (0.88) (118) (0.017) (0.008) Intent-to-treat -0.008 0.197 -68.80 -0.005 0.0008 (0.026) (0.58) (62.5) (0.010) (0.004) Mean dependent variable in the control 0.777 6.341 493.4 0.154 0.0581 Observations 5228 5228 5205 3214 3214 2006 Intent-to-treat* 0.062* -1.760*** 191.7*** 0.022 0.014** share of leaders with largest package (0.032) (0.669) (70.9) (0.017) (0.006) Intent-to-treat 0.050*** -1.352*** 188.6*** 0.055*** 0.019*** (0.019) (0.405) (34.8) (0.010) (0.004) Mean dependent variable in the control 0.761 6.209 300.9 0.152 0.066 Observations 5176 5169 5153 3278 3279 Note: The share of leaders measures the share of female leaders with the productive investment package over all female leaders in a beneficiary's registration assembly. Individual level data for education, household level data for food expenditures. Excluding households with female leaders. Intent-to-treat estimators. Highest and lowest .5% of outliers in expenditures trimmed. Robust standard errors in parentheses, corrected for clustering at the community level. *** p<0.01, ** p<0.05, * p<0.1 23 Table 3: Social interaction effects on human capital investments by intervention group Education Nutrition Attending Number of School Share of food Share of school days absent expenditures expenditures food (7-18 year from school (7-18 year for animal expenditures olds) (7-18 year olds) products for olds) vegetables and fruit 2008 Largest package* 0.0926* -2.676** 485.4** 0.0500** 0.0338*** share of leaders with largest package (0.050) (1.09) (200) (0.019) (0.011) Largest package -0.0339 0.764 -114.0 -0.00403 -0.000783 (0.032) (0.69) (72.0) (0.0099) (0.0046) Training package* 0.0293 -1.017 246.2 0.0381* 0.0227** share of leaders with largest package (0.061) (1.38) (165) (0.021) (0.011) Training package 0.00673 -0.0413 -36.92 -0.0120 -0.00227 (0.030) (0.69) (77.2) (0.013) (0.0056) Basic package* -0.000652 -0.538 192.8 0.0315 0.0111 share of leaders with largest package (0.053) (1.15) (154) (0.020) (0.012) Basic package 0.0107 -0.299 -46.06 0.000742 0.00536 (0.031) (0.69) (69.4) (0.011) (0.0053) Mean dependent variable in the control 0.777 6.341 493.4 0.154 0.0581 Observations 5228 5228 5205 3214 3214 P-value test social effect on T1 vs T2 0.671 0.744 0.779 0.743 0.350 P-value test social effect on T3 vs T1 0.109 0.116 0.193 0.252 0.069* P-value test social effect on T3 vs T2 0.360 0.291 0.348 0.575 0.373 2006 Largest package* 0.097** -2.579*** 291.6*** 0.044** 0.019* share of leaders with largest package (0.047) (0.975) (102.5) (0.019) (0.011) Largest package 0.045** -1.107** 174.3*** 0.049*** 0.020*** (0.022) (0.458) (39.5) (0.011) (0.005) Training package* 0.047 -1.356 145.6* 0.017 0.008 share of leaders with largest package (0.041) (0.844) (81.9) (0.021) (0.007) Training package 0.049** -1.438*** 181.4*** 0.057*** 0.018*** (0.023) (0.479) (39.4) (0.011) (0.004) Basic package* 0.045 -1.293 149.3* 0.006 0.016 share of leaders with largest package (0.052) (1.128) (82.8) (0.021) (0.010) Basic package 0.057** -1.574*** 211.8*** 0.058*** 0.020*** (0.026) (0.584) (42.2) (0.011) (0.005) Mean dependent variable in the control 0.761 6.209 300.9 0.152 0.066 Observations 5176 5169 5153 3278 3279 P-value test social effect on T1 vs T2 0.964 0.959 0.964 0.603 0.518 P-value test social effect on T3 vs T1 0.306 0.238 0.124 0.0325** 0.810 P-value test social effect on T3 vs T2 0.434 0.360 0.151 0.174 0.327 Note: The share of leaders measures the share of female leaders with the productive investment package over all female leaders in a beneficiary's registration assembly. Individual level data for education, household level data for food expenditures. Excluding households with female leaders. Intent-to-treat estimators. Highest and lowest .5% of outliers in expenditures trimmed. Robust standard errors in parentheses, corrected for clustering at the community level. *** p<0.01, ** p<0.05, * p<0.1 24 Table 4. Robustness checks and alternative specifications : beneficiaries of largest package Attending Number of School Share of Share of school days absent expenditures food food (7-18 year from school (7-18 year expenditures expenditures olds) (7-18 year olds) for animal for olds) products vegetables and fruit Base specification 0.0926* -2.676** 485.4** 0.0500** 0.0338*** (0.050) (1.09) (200) (0.019) (0.011) Robustness checks Excluding extreme values independent variable 0.0638 -2.087* 319.5* 0.0609*** 0.0373*** (0.057) (1.25) (186) (0.022) (0.012) S.e. clustered at level of assembly 0.0926 -2.676** 485.4*** 0.0500** 0.0338*** (0.061) (1.29) (177) (0.022) (0.010) Not excluding outliers 726.0** 0.0510** 0.0403*** (325) (0.020) (0.012) Controlling for number of people in assembly 0.0928* -2.681** 485.8** 0.0510*** 0.0338*** (0.049) (1.06) (201) (0.019) (0.011) Controlling for number of peers in assembly 0.0950* -2.745** 495.4** 0.0513*** 0.0339*** (0.048) (1.05) (203) (0.019) (0.011) With community fixed effects 0.0959* -2.668** 350.2 0.00916 0.0218** (0.051) (1.16) (226) (0.019) (0.010) Alternative specifications with # number of leaders # leaders with productive investment grant 0.0186 -0.599* 94.64** 0.0128** 0.0088*** controlling for total nr leaders (0.015) (0.33) (47.2) (0.0052) (0.0029) # leaders with productive investment grant 0.0328** -0.855** 72.98 0.00122 0.00620** controlling for total nr leaders and community f.e. (0.016) (0.36) (53.8) (0.0051) (0.0031) # leaders with productive investment grant 0.0313 -0.833* 100.3* 0.00934 0.00686** controlling for total nr leaders (0.019) (0.42) (55.2) (0.0069) (0.0031) # peers with productive investment grant -0.00284 0.0283 32.76 0.0123** 0.00107 controlling for total nr peers and community f.e. (0.020) (0.45) (36.9) (0.0054) (0.0027) P-value test social effect leader = social effect peer 0.064* 0.046** 0.265 0.567 0.100* Note: See notes table 3. Every line corresponds to a separate specification, with the exception of the last specification where the number of leaders and peers are included in the same specification. Peers are defined as all beneficiaries with the same package that are not leaders. Specification with extreme values of independent variable excluded: excludes observations for which the value of the share is in the upper 5% of the distribution. 25 Table 5: Social interaction effects on educational attainment, and parental expectations and aspirations Expected Expected monthly Expected occupation: Expected Expected earnings : Years of years of Expected professional number of monthly Cordobas Average education education occupation: or skilled rooms in earnings: (winsorized expectation attained attained professional empl. house rank 95%) index Intent-to-treat* 0.777*** 0.936* 0.0419** 0.162*** 0.353*** 309.3** 349.7** 0.316*** share of leaders with largest package (0.22) (0.49) (0.020) (0.059) (0.13) (119) (137) (0.10) Intent-to-treat -0.251 -0.217 0.00303 -0.0335 -0.0353 -68.93 -63.99 -0.0604 (0.16) (0.28) (0.0086) (0.031) (0.081) (75.2) (82.9) (0.056) Mean dependent variable in the control 3.686 8.612 0.022 0.254 2.627 1968 1535 -0.0298 Observations 3348 3329 3323 3323 3329 3281 3281 3331 Desired Desired monthly Desired occupation: Desired Desired earnings: years of Desired professional number of monthly cordobas Average education occupation: or skilled rooms in earnings: (winsorized aspiration attained professional empl. house rank 95%) index Intent-to-treat* 0.956** 0.134** 0.0272 0.229 240.3* 428.2 0.224* share of leaders with largest package (0.41) (0.065) (0.034) (0.46) (143) (324) (0.12) Intent-to-treat -0.162 -0.0326 0.0116 -0.0632 -8.006 -64.33 -0.0383 (0.26) (0.049) (0.021) (0.17) (102) (230) (0.071) Mean dependent variable in the control 13.42 0.513 0.871 5.224 1960 4168 -0.0213 Observations 3330 3328 3349 3330 3321 3321 3331 Note: Sample of Children 9-15 years old. The share of leaders measures the share of female leaders with the productive investment package over all female leaders in a beneficiary's registration assembly. Individual level data for children 9-15 years old in 2008. Excluding households with female leaders. Intent-to-treat estimators. Robust standard errors in parentheses, corrected for clustering at the community level. *** p<0.01, ** p<0.05, * p<0.1. Average expectation index is average of standardized outcomes for expected years of education, professional or skilled employment, number of rooms in the house and monthly earnings rank. Average aspiration index is average of standardized outcomes for expected years of education, professional employment, number of rooms in the house and monthly earnings rank. Earnings ranks are calculated by converting the absolute monthly earnings to the rank in the earnings distribution, combining answers of leaders and non-leaders.