WPS5137


Policy Research Working Paper                         5137




                    Impact Evaluation Series No. 41



       Changing Households' Investments
    and Aspirations through Social Interactions
          Evidence from a Randomized Transfer Program

                                  Karen Macours
                                   Renos Vakis




The World Bank
Latin America and the Caribbean Region
Poverty Sector Unit
November 2009
Policy Research Working Paper 5137


  Abstract
  Low aspirations can limit households' investments and                            beneficiaries to three different interventions within each
  contribute to sustained poverty. Vice versa, increased                           treatment community. Social interactions are found
  aspirations can lead to investment and upward mobility.                          to affect households' attitudes toward the future and
  Yet how aspirations are formed is not always well                                to amplify program impacts on investments in human
  understood. This paper analyzes the role of social                               capital and productive activities. The empirical evidence
  interactions in determining aspirations in the context of                        indicates that communication with motivated and
  a program aimed at increasing households' investments.                           successful nearby leaders can lead to higher aspirations
  The causal effect of social interactions is identified                           and corresponding investment behavior.
  through the randomized assignment of leaders and other




  This paper--a product of the Poverty Sector Unit, Latin America and the Caribbean Region--is part of a larger effort in the
  region to evaluate the effectiveness of asset transfer programs in Latin America and the Caribbean. Policy Research Working
  Papers are also posted on the Web at http://econ.worldbank.org. The author may be contacted at rvakis@worldbank.org.




         The Impact Evaluation Series has been established in recognition of the importance of impact evaluation studies for World Bank operations
         and for development in general. The series serves as a vehicle for the dissemination of findings of those studies. Papers in this series are part
         of the Bank's Policy Research Working Paper Series. The papers carry the names of the authors and should be cited accordingly. The findings,
         interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the
         International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of
         the World Bank or the governments they represent.


                                                      Produced by the Research Support Team
Changing Households' Investments and Aspirations through Social Interactions:
                        Evidence from a randomized transfer program1


                                  Karen Macours (Johns Hopkins University)

                                                          and

                                            Renos Vakis (World Bank)




 1
   We are grateful to Ximena Del Carpio for countless contributions during data collection and preparation, the
 program team at the Ministerio de la Familia and in particular 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, Carold Herrera, Enoe Moncada, Carlos Obregon and the entire field team) for
 excellent data collection. We also thank Fernando Galeana for research assistance. The paper benefited from
 comments from Abhijit Banerjee, Michael Carter, Esther Duflo, Marcel Fafchamps, Alain de Janvry, Robert Moffitt,
 Ben Olken, Rohini Pande, Martin Ravallion, Elisabeth Sadoulet, Rob Townsend, and participants at seminars at
 Georgetown University, Johns Hopkins University, MIT, Oxford University, the Paris School of Economics,
 Toulouse School of Economics, UC Berkeley, UC Davis, IFPRI, the World Bank, the Massachusetts Avenue
 Development Seminar, and MIEDC, NEUDC, and PacDev conferences. 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. Contacts: kmacours@jhu.edu and rvakis@worldbank.org
1. Introduction


Low levels of investment in human capital and productive activities are often considered a key
constraint for households to escape poverty. Many development interventions hence aim to
increase investment by the poor. Conditional cash transfer programs aim at augmenting
households' investment in human capital in at least 29 countries (Fiszbein and Schady, 2009)
and microfinance programs provide means for productive investment for more than 100 million
households around the world (Microcredit Summit Campaign, 2009). A key question related to
such programs is whether their impact on households' investments can go beyond the immediate
impact of relieving liquidity constraints and result in upward mobility. The answer is not
obvious, as other economic, social, and behavioral constraints for wealth accumulation may
often remain.2 Indeed, Appadurai (2004) and Ray (2006) have argued that upward mobility
might be difficult for the poor when they lack the capacity to aspire, which itself can be caused
by poverty. The poor might have low aspirations, and no orientation to the future, in part because
their own experiences and the experiences of those that are close to them may suggest that
escaping poverty is not a feasible option. But they also discuss how learning about the positive
experiences from others that are sufficiently "close" through social interactions can be
instrumental in changing aspirations and shaping positive attitudes to the future, and in turn lead
to investments in education and income generating activities.3


This paper provides empirical evidence of the positive impact of social interactions on
aspirations in the context of a transfer program. It shows in particular that social interactions with
nearby leaders can play a role in affecting investments and positive attitudes of other
beneficiaries. Aspirations are analyzed by considering both attitudes and investments in human
capital and productive activities, as the latter are the economic outcomes that can result from the
shift in aspirations. We focus on female leaders that live in the proximity of other beneficiaries

2
  Banerjee and Duflo (2007) indicate, for instance, that the consumption patterns in a wide set of countries suggest
that many among the poor don't seem interested in accumulating wealth.
3
  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 (such as savings
or investment in new housing technologies) among the poor. Ray (2006) proposes the concept of an aspiration
window, with an individual, through social comparisons, drawing her aspirations from the achievements and
attitudes of her peers and near-peers. He indicates that a widening of this aspiration window then can result in
deliberate action through investments that can help raise future standards of living.


                                                         2
and have similar socio-economic backgrounds, which make their experience relevant for the
other beneficiaries. We hypothesize that such leaders may affect aspirations by setting good
examples and that their experiences can help open the aspiration window of others. This could be
further enhanced by talking about these experiences and otherwise motivating others.


Psychologists have long emphasized that social comparisons can influence attitudes as well as
actions (Festinger, 1954, Karlson et al. 2004). Psychological evidence has also shown that goals
or aspirations can affect performance in various ways (Locke and Latham, 2002), and that this
can be further enhanced by leaders who communicate an inspiring vision and behave
supportively (Latham and Saari, 1979; Bass, 1985). In economics, the role of aspirations for
decision-making has been incorporated in the theoretical micro literature (Borgers and Sarin,
2000; Diecidue and van de Ven, 2008) and aspiration-based learning has also been introduced in
game theory (Bendor, Mookherjee, Ray, 2001). Aspiration levels are modeled to be constant, to
evolve over time, or to be based on experiences of others (Bendor, Diermeier and Ting, 2008;
Dixon, 2000; Karandikar et al., 1998). In the theoretical macro literature, aspirations have been
incorporated in overlapping generation models with trade-offs between human capital investment
and current consumption. In these models, aspirations often depend only on own or parents' past
consumption (Alonso-Carrera, et al. 2007; De la Croix, 2001). But in Mookherjee, Napel and
Ray (2009) aspirations are based on achievements of one's neighbors and the resulting social
externality can lead to equilibria with segregation. Aspirations in Genicot and Ray (2009) also
depend on average outcomes of others, or alternatively only on outcomes of those that are doing
better than one. Individuals react to aspirations through investment in own self-improvement or
in their children, and the model predicts a lot of upward mobility when other people's
experiences result in aspirations that are challenging, yet attainable.


Hence, while some theoretical models assume aspirations are formed based on own experiences,
others assume that aspirations are being shaped in part by social interactions with peers or with
people that are doing (slightly) better. These differences in assumptions lead to very different
predictions. This paper provides empirical evidence that helps shed light on these assumptions.
Such evidence is rare in part due to the reflection problem that complicates most empirical




                                                  3
analysis of social interactions (Manski, 1993). We use a randomized experiment to address the
identification challenge.


In particular, we rely on the two-staged randomized design of a short-term transfer program
aimed at protecting and augmenting the asset base of the rural poor in a shock-prone area in
Nicaragua. The program combined conditional cash transfers 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 analyze the role of social interactions. Households were randomly assigned to three different
intervention groups within randomly selected treatment communities. Leaders were also
randomly allocated to one of the three interventions. This implies that there is random variation
in whether beneficiaries live close to leaders with a particular benefit package. The largest of the
three packages offered resources for productive investments, and, as we will show, leaders who
randomly received this package successfully started new productive activities. This provides us
with the exogenous source of variation for our analysis, as we consider the impacts of proximity
to the random group of leaders that had received this package and analyze whether the successful
examples affected investments and attitudes of other beneficiaries.4


We hypothesize that such social interactions affect households' attitudes and amplify the
program impacts on investments in human and productive capital. Evidence based on other
randomized experiments suggests that the role of social interactions to further positive
development outcomes is not always straightforward. Duflo et al. (2006) find no evidence of
social learning for the adoption of fertilizer in Kenya, despite its demonstrated profitability. And
Kremer and Miguel (2007) find negative social effects for the adoption of de-worming drugs in
the same setting.5 This further indicates the importance of understanding the consequences of
social interactions when considering the impacts of any particular intervention.


When analyzing social interactions and learning, and following Besley and Case (1994) and

4
 In focusing on leaders, this paper also relates to the literature on the importance and the role of (female) leaders in
developing countries (Beaman et al., 2009; Chattopadhyay and Duflo, 2004; Jones and Olken, 2005).
5
  Evidence on social interaction effects based on randomized assignment in the Moving Out of Poverty experiment
is also mixed (Kling, Liebman and Katz, 2007).


                                                           4
Foster and Rosenzweig (1995), a growing number of studies consider geographic neighbors
(Munshi 2004), networks of friends and neighbors (Bandiera and Rasul, 2006), or use additional,
detailed information about communication patterns between households to identify "information
neighbors" (Conley and Udry, 2008). In this paper, we primarily consider geographical
neighbors by using proximity to female leaders, and analyze communication patterns to further
shed light on the relationship between these leaders and the other beneficiaries.


Much of the literature on social interactions focuses on social learning related to the use or the
benefits of new or existing technologies, and its impact on technology adoption.6 In relation to
conditional cash transfer programs, social interactions have also been studied with regard to
consumption and human capital investment. The focus has primarily been on spillovers through
direct transfers (Angelucci and De Giorgi, 2008; Angelucci et al., 2009) or peer effects in
education (Bobonis and Finan, 2008; Lalive and Catteneo, 2006). All of these papers find
substantial spillover effects.7


In this paper, we focus on social interactions between leaders and other beneficiaries and
similarly find large effects. Yet, by focusing on aspirations, we highlight a mechanism that is
different than technical learning, economic spillovers, or peer effects in school. In the empirical
analysis, we distinguish between these various mechanisms. Overall the empirical evidence
supports the hypothesis that examples of, and communication with, motivated and successful
leaders led to higher aspirations and corresponding investment behavior by other beneficiaries,
while it does not provide support for the alternative mechanisms.


The paper is hence organized as follows: in the next section we discuss the key features of the
program and the relevance of social dynamics and households' attitudes towards the future in the

6
  A closely related literature looks at how information from role models (Nguyen, 2008) or product endorsement by
trusted sources (Cole et al., 2009) affects investment and adoption behavior.
7
  Angelucci and De Giorgi (2008) find that direct transfers caused food consumption increases for the ineligible of
10% (compared to approximately 20% among eligible), Bobonis and Finan (2008) focus on peer effects in
secondary school enrolment and find that a 10% point increase in enrollment of a child reference group, increases
probability of enrollment with 5% points. Lalive and Catteneo (2006) find that enrollment of non-eligible increased
about 1/3 of the increase for the eligible. Angelucci et al. (2009) find no impact on secondary school enrollment
among eligible without extended families, while there is an 8% point increase among households with extended
families.


                                                        5
context of the program. Section 3 discusses the data and the empirical strategy. We show that the
randomization worked and establish that female leaders who randomly received the largest
package have better economic outcomes and more positive attitudes than others. Section 4 and 5
contain the main results of the paper and show that social interactions with these successful
leaders increased program impacts on human capital, productive investments and attitudes, and
particularly so for beneficiaries that themselves also received the largest package. Changes in
attitudes are shown to be related to the investment outcomes and attitudes of leaders. We also
shed light on the distinction between interactions with peers and interactions with local leaders.
Section 6 further investigates the mechanisms underlying the social interaction effects. We show
evidence of higher levels of communication and motivation among leaders and beneficiaries of
the productive investment package. We also investigate alternative mechanisms in the form of
economic spillovers or technical learning and do not find empirical support for those. Section 7
concludes.


2. Program design, social interactions, and aspirations


2.1.    Description of the program and the three transfer packages
The Atención a Crisis program was a one-year pilot program implemented between November
2005 and December 2006 by the Ministry of the Family in Nicaragua.8 The program was
implemented in the aftermath of a severe drought and had two objectives. First, it aimed to serve
as a short-run safety net by providing cash transfers to reduce the need for adverse coping
mechanisms, such as taking children out of school or reductions in food consumption. Second,
the program intended to promote long-run upward mobility and poverty reduction by enhancing
households' asset base and income diversification capacity. A total of 3,000 households were
selected to participate in the program. These households were allocated one of three different
packages through a participatory lottery: (i) the basic CCT; (ii) the basic CCT plus a scholarship
for an occupational training; and (iii) the basic CCT plus a grant for productive investments.
While the basic CCT's aim was to protect investments in human capital, the two additional
components aimed at strengthening households' ex-ante risk management.

8
 The pilot design built on the already existing and successful conditional cash transfer (CCT) model in Nicaragua
Red de Protección Social, evaluated by Maluccio and Flores (2004).


                                                       6
All selected beneficiary households received the basic CCT, which included cash transfers
conditional on children's primary school and health service attendance during the one-year time
period. Households received a transfer of US $145 even if they did not have children.
Households with children between 7 and 15 enrolled and attending in primary school received an
additional US $90 per household and an additional US $25 per child (with all amounts referring
to the total transfer received over the year). The school enrollment and attendance requirement
was carefully monitored. However, due to implementation problems, children's visits to the
health centers were not monitored by the program (see Aguilera et al., 2006).


In addition to the CCT, one-third of the beneficiary households also received a scholarship that
allowed one of the adult household members to choose among a number of vocational training
courses offered in the municipal headquarters. The scholarship was conditional on regular
attendance to the course. The courses aimed at providing participants with new skills for income
diversification outside of subsistence farming. These beneficiaries were also offered labor-
market and business-skill training workshops organized in their own communities.


Finally, another third of the beneficiary households received, in addition to the basic CCT, a
grant for productive investments aimed at encouraging recipients to start a small non-agricultural
business activity with the goal of asset creation and income diversification. This grant was
conditional on the household developing a business development plan, outlining the objectives of
the business and proposed investments in new livestock or non-agricultural income generating
activities. Beneficiaries received technical assistance to make a business plan and also
participated in business-skills training workshops organized in their own communities.9


Due to implementation delays, the vocational training courses had not started at the moment the
data of the follow-up survey, used for this paper, were collected. At the time of the survey, the

9
  Take-up of the overall program among eligible households was 95%. As for the different components: 89% of the
households eligible for the vocational training grant had enrolled one of its household members in a course. Take-up
of the productive investment grant among eligible households in the program was near 100%. About 10% of the
business development plans had initially been refused by the ministry, but these were sent back to the households
and virtually all of them developed a new plan, with the help of technical assistance (with the few exceptions being
the households that migrated out).


                                                         7
difference between the vocational training beneficiaries and those of the basic CCT package was
hence that vocational training beneficiaries had participated in a number of meetings with other
beneficiaries of the same package with the objective to select the courses they were going to
take. They might also have had, off course, other expectations about future skills, about related
future income and/or expectations about compensation for the time spend in training. The
beneficiaries of the productive investment package, on the other hand, had received the largest
amount of benefits: 2-3 months before being surveyed they had received $175 to invest in a
nonagricultural activity.10 In addition they had received technical assistance to select the activity
and develop a business plan, help which they were still receiving during the follow-up survey.
Given these insights and the enthusiasm observed about the productive investment package
during qualitative fieldwork, we focus on whether social interactions with leaders that received
the productive investment package affected investments and attitudes of other beneficiaries.


2.2.       Attitudes about the future and social interactions in the context of the program
The main economic activity of most of the beneficiaries of the program is the cultivation of corn
and beans, mainly for subsistence purposes. With the frequent occurrence of droughts, this
livelihood is quite precarious as harvests are often completely lost. Many households attempt to
cope with these shocks through seasonal migration. Despite the frequent re-occurrence of
weather shocks, few households seem to rely on ex-ante risk management strategies. Data from
control communities show that 38 percent of households planned to invest more in agriculture,
which - given that they reside in a drought-prone region - arguably increases their exposure to
future shocks. And 20 percent of control households report that they would do nothing or only
pray to God to prevent negative impacts of future shocks.

In qualitative interviews, informants often said that households do not really plan ahead and
instead live from "day to day". But they also testified that they had started to think more about
the future because of the program, and in particular the productive investment package.
Beneficiaries also described the importance of learning about other's experiences in changing
these attitudes.11 This is interesting as the productive interventions of the program aimed at

10
     The remaining $25 was to be paid on the next payment day (after survey completion).
11
 A beneficiary, who received the productive investment package, noted: "Before the program, I just thought about
working in order to eat from day to day. Now I think about working in order to move forward through my business.


                                                          8
increasing the households' risk management. Anecdotal evidence hence suggests that aspirations
and perspectives towards the future may have been a key part of program impact, and also
indicates the potential role of social interactions in changing attitudes.


The program could have an impact on household's investment behavior through several of its
design features. First, the level of transfers was substantial, ranging from 18 percent of average
household income for those receiving the basic CCT package to 34 percent for those receiving
the productive investment package. Such large relative transfers may facilitate shifts in asset
accumulation and behavior. Second, the conditionalities and social marketing on education,
health and nutrition aimed at changing households' perspectives about investment in long-term
human capital.      This was emphasized through repeated communications during program
enrollment, pay-days and other capacity training activities. Similarly, the vocational training and
productive investment interventions specifically aimed at increasing households' risk
management through income diversification, and these messages were also repeatedly conveyed
during program implementation.


The program's design also created many opportunities to enhance interactions between
beneficiaries, and in particular between women, who were the recipients of the cash transfers and
were given a leading role as the main participants in the implementation of the different
components of the program. Moreover, more than 90 percent of the households in treatment
communities were eligible for the program, increasing the opportunities for information sharing
and interactions, and possibly resulting in higher motivation and program ownership.


Program participants were also required to participate in a number of local events and talks
ranging from discussions on nutrition practices to workshops on business development and labor
market skills. The division of program beneficiaries in three distinct groups with different
benefits created an exogenous channel (via the random allocation of the three packages) by




Through experiences, one learns and opens up towards the future. By talking to others, one understands and
learns."




                                                    9
which beneficiaries of the same program package would have more opportunities to interact
among each other.


Finally, and as part of the program design, a subset of beneficiary women were selected during
the registration assemblies to serve as promotoras or leaders of small groups of beneficiary
women (approximately 10 per group) in order to further enhance information flows and
motivation and to enhance compliance with the various program requirements and
conditionalities. Specifically, the promotoras were expected to frequently meet with the
beneficiaries in their groups to talk about the objectives and the conditionalities of the program.
While these women self-selected to lead these groups, they were randomly allocated to one of
the three program packages (see below). The program hence created a lot of new leadership
positions for women in these communities, whereas before the program, leadership positions for
women in these communities were limited mostly to positions as teachers and health
coordinators.


2.3.     Program randomization
The program was targeted to 6 municipalities in the Northwest of Nicaragua. These were
municipalities that met both criteria of having been affected by a drought the previous year and
by the high prevalence of extreme rural poverty based on the national poverty map.12 From the
list of all communities in the 6 municipalities, 56 intervention and 50 control communities were
randomly selected through a lottery to which the mayors of the 6 municipalities were invited to
attend and participate.13 Baseline data were then used to define program eligibility based on


12
   The budget for the pilot allowed targeting 3000 households for a one-year period, which was much smaller than
the population of the 6 municipalities. The program was therefore allocated randomly. Households were notified that
funding of the project implied that the program would last 1 year, and would only cover the treatment communities.
Households in the control communities did not receive any program benefits. They were notified that if there was a
decision to scale up the program after the initial year, the control communities would be incorporated. People in the
treatment communities understood the program was only to last for a year, and people in the control communities
knew that there was a possibility they may receive the program the next year, but they also knew it was likely to
depend on the result of the national elections that were to be organized at the end of that year. In that election, the
government changed and the project was not scaled up.
13
  Before the lottery, all communities in the 6 municipalities were grouped in pairs based on similarity in road
access, infrastructure, (micro-) climate, crop mix, and proximity. Through the lottery, one community of each pair
was selected as a treatment community, the other as control. In case of uneven number of communities, a "pair"
consisted of the largest community and the combination of the two other communities. The identification of
communities and community pairs was based on maps and discussions with municipality technical personnel.


                                                         10
poverty and vulnerability, resulting in the identification of 3000 households to participate in the
program.14 From each eligible household, the female household member that was reported as the
primary caregiver was then invited to a registration assembly.15 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.16 In total, there were 134 assemblies.


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 beneficiaries were allocated to a promotora based on a joint decision, typically
based on proximity. At the very end of each assembly, all the beneficiaries - including the
promotoras - participated in a lottery process through which the three packages described above
were randomly allocated among the eligible households. Specifically, each beneficiary -
including the promotoras, any other women with already existing leadership positions in the
community, and all other beneficiaries - was asked to randomly draw a ball with 1 of 3 colors
from a black, nontransparent bag. For each assembly, the bags contained an equal number of
balls from each color, and the total number of balls matched the total number of beneficiaries in
the assembly. At the end of the day each color was matched to an intervention package through
another lottery to which all beneficiaries from the community attended. Hence at the moment of




Communities tend to be geographically separated from each other, which reduces the potential for spillover effects
from the treatment on the control communities.
14
   The eligibility criteria were determined using the proxy means. Additional discussions with local leaders from
each intervention community were conducted to identify possible exclusion or inclusions errors. Based on the
discussions with leaders, 3.72% of all the households considered were re-assigned from non-eligible to eligible, and
3.65% from eligible to non-eligible. To avoid any possible selection bias resulting from the re-assignment by the
leaders, all estimates in this paper are intent-to treat estimates, using the intent-to-treat as defined by the proxy
means methodology.
15
  Only in the few cases that there was no adult female in the household, an adult man was selected as the program
recipient.
16
   During baseline data collection, existing maps that identify the location of each house in the communities were
updated, and each house received a number based on the location on these maps. Each community has a community
leader who is the mayor's contact person in his community and who helped the survey teams identify the borders of
the community, as used for all other administrative purposes. Neighboring houses were given subsequent numbers.
Invitations to the registration assemblies were based on these house numbers, so that the beneficiaries invited to the
same assembly are likely to live in relative proximity to each other (though the geographical distance differs from
community to community, as some communities are more disperse than others).


                                                         11
promotora selection, nobody knew which intervention package the promotoras or anybody else
would end up receiving. 17


3. Data and empirical strategy


3.1.    Data
The data come from a household panel in the control and treatment communities. In treatment
communities, data were collected from all households. In control communities, a random sample
of households was selected so that the sample size in each control community was equal to one-
third of the population in the intervention community that belonged to the same pair. This
resulted in a control group of equal size as each of the three intervention groups (of about 1000
households). 18 The follow-up data was collected 9 months after the start of the CCT component
of the program. The attrition rate of the second round was very low - 1.3 percent of the original
households ­ due to tracking of households that had migrated. Attrition is uncorrelated with
treatment - in a regression of attrited households on a dummy for treatment the coefficient is -
.004, with a standard error of .005.


A number of survey instruments were used for data collection. The main household survey,
collected in both rounds, contains household and individual level data on various socio-economic
indicators for approximately 4400 households. In the follow-up survey, additional modules were
added to specifically capture information about social dynamics and information sharing. A
community survey was also collected to track, among other things, possible price changes and
the presence of new programs.


During the follow-up survey, a separate team of female enumerators administered an additional
questionnaire on early childhood development and women's socio-emotional state and attitudes.

17
   Due to the transparency of the process, the lottery process was widely perceived as fair. Participation by the
invited beneficiaries to the assemblies and lotteries was near 100%.
18
  If the number of households in the control community was less than one-third of the population of the treatment
community, additional control households were sampled in a nearby control community. About 1100 households
were sampled in the control communities. Out of those about 1000 are identified as "eligible" households using the
same proxy means method as used for the treatment. These are the control households included in the intent-to-treat
estimates.


                                                        12
This data were collected for all women who were the primary caregivers of children between 0
and 8 in treatment and control households. Mental health was measured using the Center for
Epidemiological Studies Depression scale (CESD). The CESD is a widely-used measure of
depression (Radloff, 1977), and consists of 20 questions on self-reported depression. Besides
calculating an aggregate depression score, these questions also allow to look separately at
positive and negative feelings, including expectations about the future. Finally, the questionnaire
also included direct questions about aspirations and perceptions on upward mobility. This allows
us to consider the relationship between attitudes towards the future and social interactions.


In addition to the quantitative data, qualitative work preceded each round of data collection. The
qualitative work consisted of focus groups and semi-structured interviews with a wide set of
beneficiaries and other local actors in treatment and control communities, and in municipal
headquarters, and explored qualitative evidence of the program's impacts as well as issues
related to program implementation (Aguilera et al., 2006).


3.2    Identifying social interactions
Our identification strategy relies on two key program design elements, namely the randomized
allocation of beneficiaries to one of the three program packages, and the random allocation of
these same packages among local female leaders. This allows us to explore whether program
impacts depend on social interactions between beneficiaries and leaders.


Table 1 presents the randomization results for the full sample of eligible households. The
baseline differences between treatment and control communities are generally small and not
statistically significant. The last three columns in Table 1 show the P-values for differences
between the three intervention packages (i.e. the result of the participatory lotteries in the
communities). As expected, there are only a small number of statistically significant differences
and the direction of the differences suggests there is no systematic bias.


The variables used to identify social interaction effects rely on the random allocation of female
leaders to one of the three intervention groups. We consider both promotoras and other women
with leadership positions. Specifically, information was collected for each household member on



                                                 13
leadership responsibilities in the community. About 17% of eligible households are households
with a female leader. The share of households with female leaders is higher (19%) in the
treatment than in the control communities (11%), given that about half of the leadership
positions in the treatment communities were directly created by the program in the form of the
program promotoras. Other female leadership positions in the communities are mainly
responsibilities as health coordinators and teachers, which already existed before the program.19
The lottery process described above implied that the distribution of female leaders across the
three different program packages is random across registration assemblies.


Table 2 confirms this by presenting randomization results for the subgroup of households with
female leaders. There are no systematic significant baseline differences between leaders with the
three packages. Comparing Table 1 and Table 2 also sheds some light on the characteristics of
the female leaders. Female leaders tend to be younger and more educated than the average
beneficiary. While beneficiaries have on average 3 years of education, leaders have on average 5
years. For the leaders that received the productive investment package, these higher education
levels, and in particular higher levels of numeracy, are likely to be important for a more
successful management of their new nonagricultural activities. Leaders' outcomes for human
capital investments in their children also are generally somewhat better than those of others. On
the other hand, income level and the income structure at baseline were similar to those of the
other beneficiaries. The similar socio-economic status of leaders and others might make it easier
for the others to identify with the positive experiences of the leaders. 20


For the analysis of the social interactions, we consider all female leaders that participated in each
registration assembly, and calculate the share of female leaders that was randomly allocated
productive investment packages. The share varies between 0 and 1. On average, 32 percent of
female leaders received the productive investment package, which further confirms the

19
  We consider both types of female leaders together, in part because they are not mutually exclusive (as many
health coordinators and teachers volunteered to be promotoras). Empirically, we cannot reject that proximity to
promotoras and other leaders has the same impact.
20
  Descriptively, the data further show that living standards are very low. Only 13% of households have access to
water in their homes, 40% have access to electricity, about a quarter do not have latrines, and average per capita
income is less than a dollar per day. The data also show that monetary income from self-employment or local
agricultural wage work is very limited, with households depending to a large extent on self-consumption of their
agricultural production, migration income, and transfers.


                                                       14
randomization (Table 1). Yet, in some assemblies, the share will be relatively high while in
others it can be low. Given that leaders participated in the same lotteries than other beneficiaries,
it can randomly occur, for instance, that in one assembly 0 out of 3 leaders got the productive
investment package, while in another assembly 2 out 3 did. It is this random variation that we
rely on to identify social interaction effects.


Based on the above, our general specification is of the following form:


                                       Yia = 0+1Tia + 2(Tia* Sa) + ia                                   (1)


where Yia is an outcome indicator for beneficiary i who participated in assembly a, Tia is
assignment of beneficiary i to any of the treatment groups, and Sa is the share of female leaders
(over all female leaders in the assembly) that randomly received the productive investment
package in beneficiary i's registration assembly.21 Given that households were invited to
particular assemblies based on geographic proximity, Sa will capture the share of leaders with the
productive investment package that live in the proximity of beneficiary i.22 A finding, for
example, that 1 and 2 are both positive would imply that while assignment to the treatment
group increases the outcome of interest (1), there is an additional impact of the program that
comes from the social interactions (2). We also explore how the share of leaders with the
productive investment package affects impacts for beneficiaries of each of the three packages
separately. All regressions are estimated on the sample of beneficiaries that are not leaders
themselves.


3.3. Outcomes for households with female leaders
Before turning to the social interaction results, this section describes the follow-up outcomes of
female leaders in order to further motivate the focus on leaders with the productive investment
package. In particular, we show how the outcomes of these leaders differ both compared to other


21
     Sa is always zero for those in the control group and hence collinear with Tia* Sa.
22
   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.


                                                             15
leaders and compared to other beneficiaries with the same package. For human capital
investments, we focus on education and nutrition because the social marketing of the program
was heavily focused on these investments and because the education conditionalities were
rigorously monitored by the program (in contrast to the health conditionalities). Specifically, we
consider school attendance and spending on education for children between 7 and 18, as well as
the share of food expenditures used for nutrient-rich food (animal proteins and fruit and
vegetables) at the household level.23 In terms of economic activities, we focus on income from
local wage and self-employment. Specifically, we consider increases in income in
nonagricultural activities and especially commercial activities, and shifts away from agricultural
wages. And given that investments in livestock take a longer period to result in income, we
consider the value of the animal stock.


Households with female leaders who received the productive investment package are, at follow-
up, more likely to have higher income from commercial activities, and more generally, from
non-agricultural self-employment, than other leaders (Table 3). By the time of the survey, they
had earned about 16 US$ more per capita in non-agricultural employment than the other leaders.
They also have larger animal stocks (about 26 US$ per capita more). On the other hand, we do
not find significant differences between the three types of leaders with regard to human capital
investments. Considering the attitudes of the female leaders towards the future, positive feelings
are generally the strongest for leaders with the productive investment package, who are for
example 20 percentage points more likely to report that they are moving ahead in life than
leaders with the basic package.24 Overall, outcomes for leaders that received the productive
investment package hence appear to differ from other leaders and they generally seem to be
doing better. This is particularly true for economic activities and for their attitudes regarding
upward mobility.




23
  Focusing on cognitive development in early childhood and related investments, Macours, Schady, and Vakis
(2008) show that the program caused an overall shift in parental investment and spend proportionally more on
animal protein, fruit and vegetables.
24
   Leaders with the vocational training intervention have more positive expectations than those with the basic
package (consistent with the expectations about the delayed intervention).


                                                     16
Compared to the other beneficiaries with the productive investment package, leaders appear to be
investing more in human capital and productive activities (Table 3). Their income from
commercial activities is more than twice as high compared to other beneficiaries, and the value
of their animal stock is higher. Leaders also have higher overall income levels, and have a
stronger feeling that they are moving forward in life. The distribution of the type of activities
leaders and other beneficiaries invested in with the productive grant is also different, with leaders
investing more in commercialization of specialized products or general corner stores, while other
beneficiaries invest more in small livestock (Table 4). This difference in the type of activity may
in part help explain why leaders had higher outcomes 2-3 months after receiving the grant, as
income gains from livestock activities take more time. More generally, it is not surprising to find
better outcomes for leaders, since they might be more likely to be entrepreneurial and/or have
other unobserved characteristics correlated with leadership. Independently of the mechanism, the
data indicate that human capital and productive outcomes, as well as positive attitudes, were
higher for leaders, leaving room for positive effects on other beneficiaries.25


4. Social interaction effects on human capital and productive investments


4.1 Main results
We first consider all eligible households, excluding the leaders themselves, and investigate
whether there is a relationship between program impacts and the presence or proximity to female
                                                                             26
leaders who received the productive investment package.                           The interaction terms in Table 5
show that social interaction effects increased program impacts on human capital investments.
The higher the share of leaders with the productive investment package in one's proximity, the
higher the impacts on various education and nutrition investments of other beneficiaries.27 The
effects are not only statistically significant, but are also quite large. For example, while school
attendance is estimated to increase with 5 percentage points when no leader in one's assembly

25
  Note that any positive social effects would reduce the observed differences between leaders and non-leaders,
implying that the differences in Table 3 are an underestimate of the potential for social effects.
26
   The average treatment effects for all eligible households are in line with results from other conditional cash
transfer program with strong impacts on education, consumption, and nutrition. Also average income from
commercial activities and more generally from non-agricultural self-employment increased significantly for the
beneficiaries who received the productive investment package (results available from authors).
27
     Similar results are obtained with variables capturing individual food intake of young children (0 to 8 years old).


                                                            17
received the productive investment package, it increases with an estimated additional 6
percentage points if all the leaders in one's assembly got the productive investment package.


Table 6 shows the social effects on human capital investments separately for each treatment
group. We find that social interactions are the strongest for beneficiaries of the productive
investment package. For instance, the share of food expenditures used for nutrition-rich food is
estimated to double for beneficiaries of the productive package in the extreme case that the share
of leaders with the same package changes from 0 to 1. The social effects are smaller for the
beneficiaries of the training packages and the basic packages (about half the size) and many of
the interaction terms are not individually significant. Nevertheless, the P-values indicate that we
can generally not reject that the social effects are different for the three groups. When pooling
the basic and the training packages together, the interaction effects for the number of days
absent, school expenditures, and expenditures for fruit and vegetables are significant.


Turning to shifts in economic activities, there are no significant social effects on all beneficiaries
together (Table 7). But Table 8 shows however that the higher the share of leaders with the
productive investment package in one's assembly, the higher the income from nonagricultural
activities, and in particular from commercial activities, for beneficiaries with the productive
investment package. This seems to be compensated with a shift away from agricultural wages.
There are also substantial social effects on livestock holdings.


Social interactions hence enhanced investments in nonagricultural activities (commerce and
livestock), and as such augmented the program's objective of income diversification. This is the
case for the beneficiaries of the productive investment grants, but not for the other beneficiaries.
The point estimates of the interaction effects are small for the other beneficiaries, and the P-
values indicate that the social interaction effects for economic outcomes are significantly
different between the beneficiaries with the productive package, when compared with the other




                                                 18
beneficiaries. This is consistent with the fact that the other beneficiaries had not received any
extra resources that would allow for such investments.28


To put the magnitude of the social interaction effects in perspective, it is useful to know there are
on average about 4 leaders in a registration assembly. Hence the coefficients suggest that having
one additional leader with the productive investment package in one's assembly increases school
attendance with approximately 2.5% points, and reduces absences with about .85 days per
month. Looking at the social effects on productive investments, one additional leader increases
income from nonagricultural activities with about 60 cordobas (about 3.3 US$) per capita, and
value of the animal stock with about 220 cordobas (about 12US$) per capita, on average (with
average household size about 5). Given that the beneficiaries had received 175$ of the
productive investment grant 2 to 3 months before the survey, and different households invested
in different types of activities, these are substantial effects.29


4.2. Robustness checks
The results on social interaction effects on human capital and productive investments are robust
to several alternative specifications.30 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
productive investment grant is .33, the range varies between 0 and 1. However, for 95% of the
observations, it is between 0 and .67. In the first robustness check in Table 9, we therefore
exclude the observations with values above .67, and show this does not substantially alter any of
the results, even if, as expected, the standard errors increase. The results are also robust to

28
   One could alternatively hypothesize that for beneficiaries of all packages impacts might be larger for if they
received the same packages than the leaders in their proximity. However, we do not find evidence of such social
interaction effects for beneficiaries with the basic or the training grant (results available from the authors).
29
   The results in Table 8 also suggest that in the extreme case that there are no leaders with the productive
investment package in one's registration assembly, there are no significant impacts of the productive investment
package. This could possibly be explained by slower implementation of the new activities in those cases and does
not necessarily mean that on the long-term the intervention does not have an impact on those households. Anecdotal
evidence suggests that some beneficiaries with the productive investment package had delayed implementation of
their plans, even after receiving the cash. On the other hand, some others had started to invest in their
nonagricultural activities even before receiving the actual transfer, drawing on other income sources and possibly
some of the CCT money.
30
  Table 9 presents robustness checks for the beneficiaries with the productive investment grant. The social effects
on human capital investment when all beneficiaries are considered together are similarly robust to these alternative
specifications.


                                                        19
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 control 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. This accounts for the fact that as the share of leaders with the
productive investment package increases in an assembly, the share of peers with that package
automatically decreases. One could hypothesize this might have an effect on its own, but in that
case we would expect the total number of persons in the registration assembly (which varies
because of the different sizes of the communities) to be correlated with such an effect. The
results are robust to including these controls.


The next specification includes a community fixed effect. This sheds light on whether, once we
control for any community-level social effects, we still find social effects of leaders that live in
beneficiaries' proximity (i.e. that were in their registration assembly). While the fixed effects
specification reduces the variation in the independent variable (as 13 of the 56 treatment
communities only had one registration assembly), the results still show significant social
interactions effects for productive activities and educational outcomes.


Table 9 further shows alternative specifications using the number of female leaders with the
productive investment grant instead of the share as independent variable. These specifications
also control for the total number of female leaders in the registration assembly. The coefficient
on the number of leaders with the productive investment grant shows results that are largely
consistent with the ones with the share variable in terms of sign, size, and magnitude. One
drawback of this specification is that the total number of leaders living in one's proximity is
likely to be correlated with many different unobservables. The community fixed effects in the
next specification controls for some, but not all of these unobservables, which is why this is not
our preferred specification.


Nevertheless, an advantage of considering the number of leaders, as opposed to the share, is that
we can compare the coefficient of the number of leaders with the productive investment grant,
with the coefficient of the number of peers with the productive investment grant (last
specification in Table 9). This is interesting as one might hypothesize that social effects might



                                                  20
not only be generated by leaders, but also by peers with the same package. Given that one-third
of the beneficiaries in each registration assembly received the productive investment package,
once we control for the share of leaders with the productive investment package, there is very
little variation in the share of peers with that package.31 Yet, registration assemblies varied in
size, implying that there is more variation when we consider the number of peers as opposed to
the share. The results in Table 9 suggest that social interaction effects from peers might be more
limited, as the coefficients for the number of peers with the productive investment package are
generally not significant, and smaller than the coefficients for the number of leaders. T-tests
confirm that the coefficients for leaders and peers are significantly different for the economic
outcomes (except animal stock) and the educational variables. Nevertheless, 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, and given the concern related to the endogeneity of the
total number of leaders in an assembly.32


5. Social interaction effects and attitudes


We now turn to investigating the relationship between social interactions with leaders and the
attitudes of other beneficiaries. Given the above findings, we focus on beneficiaries with the
productive investment package. As the literature on attitudes and mental health in developing
countries has found strong correlations between an individual's mental health and average
community mental health in different countries (Das et al., 2007), Table 10 shows estimates of
the basic specification and a specification with community fixed effects.33 The results show that
women caregivers in beneficiary households with the productive package are more likely to

31
  There is some variation coming from the fact that the number of people in a registration assembly was not always
a multiple of three, but this is very limited.
32
   An alternative approach to analyze the social effects of leaders versus peers could be to consider the share of
leaders and peers in the smaller groups that were formed around the promotoras during the registration assemblies.
But, while the results of the social effects of leaders by group are qualitatively similar than the results by assembly,
the coefficients are smaller and less precisely estimated. This could be because social effects are not restricted to the
smaller groups, but it is also possible that beneficiaries reorganized the groups after the assemblies, so that the
effective groups do not correspond to the administrative data on groups. Moreover, the administrative information
on group composition itself is less precise than the information on who participated in which assemblies, introducing
more measurement error in these estimations.
33
  We only have information on positive attitudes and mental health from a subsample of households (those where
the primary beneficiary was the caregiver of small children), which might affect the precision of the basic
specification. Including the community fixed effects helps to increase the precision of these estimates.


                                                          21
express positive attitudes when there are more leaders with the productive investment package in
their proximity: they are more optimistic about the future, are happier in life, and have lower
indicators of depression. Further, the higher share of leaders with the productive investment
package, the less likely beneficiaries of the productive investment package answered they would
not do anything (or anything else besides praying) to reduce the impact of future shocks. The
social effects are large: given an average of 4 leaders per assembly, an additional leader with the
productive investment package in one's assembly implies that beneficiaries are about 6.5
percentage points more likely to be optimistic about the future, and about 2.5 percentage points
less likely to report they will "do nothing" to manage future risks.


A possible interpretation of these findings is that the successful outcomes of the leaders in their
new activities helped to open the aspiration window of the other beneficiaries. We investigate
this in Table 11 by regressing the attitudes variables of the beneficiaries with the productive
investment package on outcomes of the leaders. In particular, we consider whether the maximum
values of leaders' income in nonagricultural activities, their total income, and their perceptions
on whether they are moving forward in life are positively correlated with the attitudes of other
beneficiaries that were part of the same registration assembly. While there are clearly many
unobservables that could affect both the leaders' outcomes and the beneficiaries' attitudes, we
control for many of these by including 3 variables to separately capture outcomes of leaders with
the basic package, the vocational training package and the productive investment package. We
also add a community fixed effect to further control for unobservables. While these results are
more suggestive because of possible remaining concerns about unconfoundness, they are quite
striking. Table 11 shows a positive relationship between the economic outcomes of leaders with
the productive investment package and the attitudes of the other beneficiaries with the productive
package. We do not find a similar relationship between the outcomes of the other leaders and
attitudes (and in fact, some of the results suggest an opposite relationship for the leaders with the
basic package). 34




34
  Similar regressions to look at whether investment outcomes of leaders with the productive package are directly
correlated with the same investment outcome of others show that these correlations are much weaker.


                                                      22
Returning to the question on whether the social effects of peers are similar to the social effects of
leaders, we estimate a similar regression as above, but now also include the outcomes of peers
with the three different packages separately. These results show that in general the positive
correlations between positive attitudes and outcomes of leaders with the productive investment
package are larger and more significant than the correlations with the outcomes of peers with the
same package (Table 12). However the differences between leaders and peers are only
significant for some of the outcomes. Taking these results together with the results on peers in
Table 9, we interpret this as tentative evidence that the outcomes of leaders generate more social
effects than those of peers.


6. Underlying mechanisms


The results in sections 4 and 5 show that the proximity to female leaders who received the
productive investment package considerably strengthened program impacts on households'
investments and attitudes. There is also a strong relationship between leaders' outcomes and
other beneficiaries' attitudes. This is consistent with our interpretation that leaders are affecting
others by increasing their aspirations. We now explore this mechanism further, before turning to
possible alternative mechanisms.


6.1     Social interactions and motivation of leaders and beneficiaries
First, we analyze whether there is evidence on whether leaders and beneficiaries with the
productive investment are more motivated and whether they share more information with each
other. One indication of their motivation is the effort that either of them uses to communicate
with the other. We therefore use information about the location of different houses and define a
proximity metric based on the distance of a beneficiary's house to the closest female leaders'
house.35 We expect people to communicate less with each other as distance increases. This could
be the case either because of increased transaction costs (time) that come with distance, or purely
because they might be less likely to be close family or friends, and more likely to be different
from each other along various dimensions, when they live farther from each other. We therefore
35
   Since we do not have information about the exact physical distance between the different houses we use
information on the order by which dwellings were numbered as an indicator of proximity. This information was
obtained from detailed community maps that were updated during baseline data collection.


                                                    23
analyze whether there is a difference in the impact of distance for the different types of leaders
and beneficiaries, taking advantage again of the randomized allocation of the three packages to
both beneficiaries and leaders.


Table 13 shows the relationship between distance and communication with leaders. The
dependent variable in the first three columns measures whether the beneficiary talked to her
promotora in the last week. The dependent variable in column 4, 5, and 6 measures whether the
beneficiary talked to her promotora, a health coordinator, or a teacher.36 The results in column 1
and 4 first establish that, in general, distance to a leader does indeed reduce the likelihood of
talking to the leaders. Column 2 and 5 show the evidence on motivation. In particular, we find
that distance does not significantly reduce communication between leaders of the productive
investment package and other beneficiaries, while the effect of distance is the largest for leaders
with the basic package. The point estimate is in between the two others for the distance to leaders
with the vocational training package (column 2 and 5).37 The contrasts between the distances to
different leaders are even bigger when only considering beneficiaries of the productive
investment grant (column 3 and 6). P-values of the t-test for equality of the coefficients show
that the differences between leaders with different packages are highly significant. These results
could mean that leaders with this package, and, to a lesser extent, leaders with the vocational
training transfer put in more effort to go talk to the beneficiaries, or that beneficiaries put in more
effort to go talk to those leaders. Overall, it is clear however that the leaders and beneficiaries of
the productive investment package are communicating more.


It also appears that they specifically talked about the program while meeting as we find similar
patterns in beneficiaries' knowledge about the program. Specifically, a short test with nine yes-
no questions on program knowledge was applied. The test dealt with issues related to targeting,
program objectives and conditionalities, and general program rules. Comparing the last 3
columns in Table 13 with the previous results shows that the relationship between


36
  As indicated above, promotoras, teachers and health coordinators represent the vast majority of female leaders.
Unfortunately, we don't have information about whether the beneficiaries talked to other types of female leaders.
37
   As expected because of the randomization, the main distances to leaders with specific packages are three times
larger than distance to any leader, explaining why the point estimates on distance in column 2 and 5 are overall
smaller than in column 1 and 4.


                                                       24
communication and distance between different types of beneficiaries and promotoras is
generally reflected in the knowledge score. While the differences are less stark than before, the
results show a negative relationship between program knowledge and distance to leaders with the
basic or the vocational package, and not for distances to leaders with the productive investment
packages. The results on knowledge make it unlikely that the frequent communication merely
reflect pure business transactions. Hence, overall distance to leaders with the productive
investment package does not reduce communication with leaders or program knowledge, while
distance to other leaders does. And these patterns are stronger when we only consider beneficiary
with the productive investment package.


It is of course possible that beneficiaries and leaders with the productive investment package
make more effort to talk to each other because as they start to engage in their new activities, they
have a greater need to share experiences. And the results might also partly capture "mechanical"
program impacts. Specifically, because of the different activities related to the program,
beneficiaries with productive investment package regularly attended meetings and workshops
together. Yet, whether the reported communication is directly linked to program activities or
not, overall the results show that there is frequent communication between leaders and
beneficiaries of the productive investment package.38 Such communication with leaders may have
facilitated and enforced the possibility for beneficiaries to get inspired by the positive examples
of these leaders. It is hard to separate out to what extent the results are driven by the motivation
of leaders, and to what extent they are driven by the positive examples these leaders provide,
exactly because these two mechanisms are likely to reinforce each other.39 More positive leaders
likely shared their positive experiences and reflect their enthusiasm and positive attitudes in their
interactions with other beneficiaries, and by observing and learning about these successes, others
became more excited and optimistic. Hence the positive attitudes appear to have been
contagious.

38
   The program also influenced communication more generally. Reported communication with the community
leader, the health coordinators and the teachers is between 25 and 50 percent higher in treatment than in control
communities. Moreover, in the treatment communities, people are 31 percent more likely to talk about food prices
and 200 percent more likely to talk about businesses than in the control communities, and these impacts are stronger
for beneficiaries of the productive investment package than other beneficiaries.
39
  The evidence does suggest that the social effects do not primarily result from better monitoring and enforcement
of the program rules by the more motivated leaders, as this would be unlikely to lead to positive effects on attitudes.


                                                         25
6.2        Alternative explanations: technical learning or economic spillovers
The social effects through aspirations discussed in this paper are different from other social
interaction effects in the literature. In particular, it is distinct from social learning related to
technical issues and/or information asymmetries. The mechanism is also different from a purely
economic spillover effect through monetary transfers. We now discuss why, in the context of
Atencion a Crisis, it seems unlikely that these alternative types of social effects can explain the
results.


First, one could hypothesize that it was "technical" social learning, together with a relaxation of
the liquidity constraints that changed households' perspectives about the future because it gave
them access to new opportunities for wealth accumulation. This then, arguably might not reflect
a change in aspirations, but rather a change in expectations, more narrowly defined. Yet the
scope for learning-from-others on business management in the productive investment group was
limited, as there were different types of businesses, and the type of activities in which leaders
tended to specialize were somewhat different than those of others (see Table 4).40 To shed more
empirical light on whether beneficiaries were likely to imitate leaders, or were learning about
specific activities from their leader, we consider commercial, noncommercial, and livestock
activities separately and analyze whether a beneficiary has (higher) income from each of those if
their leaders are active in that same activity. Specifically, Table 14 shows the relationship
between the beneficiaries' incomes from a certain activity and variables capturing the income or
participation of the leaders in their proximity in that same activity, controlling for community
fixed effects. We find no evidence of a positive correlation between the activities of the leaders
with the productive investment package and the type of activities of the other beneficiaries with
the same package. This makes it unlikely that the observed social effects resulted from technical
learning or imitation of specific activities. 41

40
  Also, households were receiving ongoing technical assistance from technical staff related to the program, whom
they could ask more technical advice on the specific activity they were undertaking.
41
  It is further unlikely that the results can be explained by households pooling resources for a joint project. In fact,
households were given the explicit option to pool resources with other beneficiaries and submit a joint plan but less
than 5% of households did. Given that few beneficiaries decided to submit joint plans, and given that the
implementation of the plans was monitored by the technical staff of the program, it is unlikely that households
pooled resources ex-post.


                                                          26
Second, we investigate whether the identified effects of leaders instead may reflect pure
economic spillover effects. This could be because people with leadership positions might also be
better entrepreneurs. Indeed, Table 3 shows that leaders did better than non-leaders, especially on
outcomes related to productive activities. If more of them received the productive investment
package in a given community, this might then have caused a larger boost to the local economy
than when other people received the transfer. This in turn might have economic spillover effects
on other beneficiaries, as successful leaders now buy from other businesses or provide better
access to (food) products for other households. However, the data do not provide support for this
alternative hypothesis. First, we analyzed the availability and prices of 20 products (basic food
and household products) in the community. The share of leaders with the productive investment
package only increases the availability of one out of the 20, and does not affect any of the prices.
Consistent with these findings, the share of leaders with the productive investment package does
not seem to induce people to buy their products in their own community (P-value =0.97 for all
beneficiaries, P-value = 0.80 for beneficiaries with productive investment package). And, similar
to other beneficiaries, leaders report that they buy the majority of their products outside of the
community. Leaders with the productive investment package are also not more likely to buy food
in their community than other leaders. Further, it seems likely that any type of economic
spillovers from increased supply or demand by leaders would affect the larger community. But
Tables 9 and 10 showed that the social interaction results are largely robust to the inclusion of a
community fixed effect, further making it unlikely that they are driven by such economic
spillovers.


The social effects in this context also do not result from monetary transfers as increasing the
share of leaders with the productive investment package in a registration assembly does not lead
to increased transfer income for other beneficiaries (P-value is .66).42 And, as shown in Table 8,
it does also not lead to increases in agricultural wage earnings by other beneficiaries (in fact the



42
  Results for the CCT program in Mexico (Angelucci and De Giorgi, 2008) show considerable spillovers through
monetary transfers on non-beneficiaries. Unfortunately, the design of Atencion a Crisis does not allow for a rigorous
analysis of spillover effects on non-beneficiaries, given that 90% of all households were eligible, and 3.72 % of the
remaining 10% had received the program because of reallocation by the leaders (see footnote 14).


                                                        27
coefficient is negative).43 Overall, these results suggest that economic spillover effects cannot
explain the social interaction effects identified in this paper.


7. Conclusions


Many development programs aim, through a variety of mechanisms, to change the attitudes and
aspirations of beneficiary households. Indeed, when programs are designed to only last for a
limited period, the sustainability of the impacts might crucially depend on their success to affect
attitudes toward the future, and as a consequence lead to long-term changes in investment
behavior. For example, an implicit or explicit objective of many conditional cash transfer
programs is to change households' attitudes and/or the social norms toward investment in the
education, health and nutrition of their children. Yet, the mechanisms through which such change
in attitudes can be reached and reinforced are not always clear. The evidence in this paper
suggests that social interaction effects can contribute to such changes in attitudes and investment
behavior.


While it appears intuitive that social interactions can be key for changes in aspirations, it can be
difficult to identify the causality of this relationship. This paper uses a unique experiment with
two levels of randomization to address this identification problem. We find that there were large
and significant social interaction effects in the program studied. Social interactions with leaders
who randomly received the largest program package substantially increased program impacts on
both human capital investments and income diversification and affected households' attitudes
toward the future. While social effects may result from a variety of mechanisms, the evidence
suggests changes in aspirations facilitated by increased communication and motivation by female
leaders were important in this context. The evidence does not support alternative mechanisms,
such as technical learning or economic spillovers.


These results suggest that witnessing local success stories of upward mobility can be important
to change households' investment behavior. While the design of the experiment analyzed does

43
  Agricultural wage jobs are typically the only wage jobs available in the communities. The new businesses were
small and did not employ others.


                                                      28
not allow shedding definite light on possible social effects generated by peers, the results show
that natural leaders living in people's close proximity can be important vehicles for such changes
in behavior, both because they may motivate and encourage such changes, and because they may
provide the examples that people aspire to follow.


In contrast with a common focus on the possible negative roles of leaders through elite capture,
the evidence in this paper hence draws attention to the positive role local leaders can play. In
particular, it points to the importance of assuring that development program designs leave room,
and possibly enhance, a positive role for natural leaders. It does not however suggest that
interventions should be primarily targeted to such leaders. Indeed, we find that social effects are
particularly large when leaders and beneficiaries received the same package, and find much
smaller effects on households that received smaller packages. Overall, these findings suggest that
examples of positive experiences of nearby leaders can help open people's aspirations window,
and particularly so when they are provided with resources to follow those examples.


More generally, the results have implications for the debate on the feasibility and sustainability
of using cash or asset transfer programs in low-income countries. Some argue that such countries
can simply not afford to distribute transfers to all poor households for long periods of time. The
question then becomes whether and how short-term transfer programs can be designed to launch
households on a sustainable pathway out of poverty. Sustainability of short-term interventions
may depend on whether they manage to change households' attitudes toward the future and the
related social norms. The evidence in this paper suggests that designing such programs in ways
that facilitate and encourage social interactions may be important to create such shifts in
aspirations.




                                                29
References

Alonso-Carrera Jaime, Jordi Caballe, and Xavier Raurich, 2007. "Aspirations, Habit Formation,
      and Bequest Motive", The Economic Journal, 117: 813-836.
Aguilera, Verónica, Ximena Del Carpio, Carold Herrera, Karen Macours, Miriam Enoe Moncada
       Carlos Obregón, and Renos Vakis, 2006. "Estudio Cualitativo del Componente de
       Atención a Crisis (CAC) del Ministerio de la Familia, Nicaragua", mimeo, Johns
       Hopkins University.
Angelucci, Manuela, and Giacomo De Giorgi, 2008. "Indirect Effects of an Aid Program: How
      do Cash Transfers Affect Ineligibles' Consumption", American Economic Review,
      forthcoming.
Angelucci, Manuela, Giacomo De Giorgi, Marcos Rangel, and Imran Rasul, 2009. "Family
      Networks and School Enrollment: Evidence from a Randomized Social Experiment",
      mimeo, University of Arizona.
Appadurai, Arjun, 2004. "The Capacity to Aspire". In Rao, V., and M. Walton (eds.) 2004.
      Culture and Public Action. Palo Alto, CA: Stanford University Press.
Bendor, Jonathan, Daniel Diermeier, and Michael M. Ting, 2008. "Social Comparisons in
      Games". Available at SSRN: http://ssrn.com/abstract=1086204.
Bendor, Jonathan, Dilip Mookherjee, and Debrj Ray. 2001. "Aspiration-Based Reinforcement
      Learning in Repeated Games: An Overview." International Game Theory Review 3: 159-
      174.
Bandiera, Oriana, and Imran Rasul. 2006. "Social Networks and Technology Adoption in
      Northern Mozambique." Economic Journal, 116(514): 862­902.
Banerjee Abhijit V., and Esther Duflo, 2007. "The Economic Lives of the Poor", Journal of
       Economic Perspectives, 21(1): 141-67.
Bass, Bernard M., 1985. Leadership and performance beyond expectations. New York: Free
       Press.
Beaman, Lori, Raghabendra Chattopadhyay, Esther Duflo, Rohini Pande, and Petia Topalova,
     2008. "Powerful Women: Does Exposure Reduce Prejudice", Quarterly Journal of
     Economics, forthcoming.
Besley, Tim, and Anne Case, 1994. "Diffusion as a Learning Process: Evidence from HYV
       Cotton." Discussion Paper. Princeton University, Woodrow Wilson School of Public and
       International Affairs, Princeton, NJ.
Bobonis, Gustavo, and Frederico Finan, 2008. "Endogenous Social Interaction Effects in School
      Participation in Rural Mexico", Review of Economics and Statistics, forthcoming.
Borgers, Tilman, and Rajiv Sarin. 2000. "Naïve Reinforcement Learning with Endogenous
       Aspirations." International Economic Review 41, 921-950.
Chattopadhyay, Raghabendra and Esther Duflo, 2004. "Women as Policy Makers: Evidence
       from a Randomized Policy Experiment in India", Econometrica, 72, 1409-1443.



                                             30
Cole, Shawn, Xavier Gine, Jeremy Tobacman, Petia Topalova, Robert Townsend, and James
       Vickery, 2009. "Barriers to Household Risk Management: Evidence from India", Federal
       Reserve Bank of New York Staff Reports, no 373.
Conley, Timothy G., and Christopher Udry. 2008. "Learning About a New Technology:
      Pineapple in Ghana." American Economic Review, forthcoming.
Das, Jishnu, Quy-Toan Do, Jed Friedman, David McKenzie, and Kinnon Scott, 2007. "Mental
       health and poverty in developing countries: Revisiting the relationship", Social Science
       and Medicine 65 (2007): 467-480.
De la Croix, David, 2001. _"Growth dynamics and education spending: the role of inherited
      tastes and abilities." European Economic Review (45): 1415­38.
Diecidue, Enrico, and Jeroen van de Ven, 2008. "Aspiration Level, Probability of Success and
       Failure, and Expected Utility", International Economic Review, 49(2): 683-700.
Dixon, Huw David. 2000. "Keeping Up with the Joneses: Competition and the Evolution of
       Collusion." Journal of Economic Behavior and Organization 43: 223-238.
Duflo, Esther, Michael Kremer, and Jonathan Robinson, 2006. "Why Don't Farmers Use
       Fertilizer: Evidence from Field Experiments in Western Kenya." Massachusetts Institute
       of Technology. Cambridge, mimeo.
Festinger, Leon. 1954. "A Theory of Social Comparison Processes." Human Relations 1: 117-
       140.
Fiszbein, Ariel and Norbert Schady, 2009. Conditional Cash Transfers. Reducing Present and
       Future Poverty. The World Bank. Washington DC.
Foster, Andrew D., and Mark R. Rosenzweig, 1995. "Learning by Doing and Learning from
        Others: Human Capital and Technical Change in Agriculture." Journal of Political
        Economy 103(1995):1176­209.
Genicot, Garance, and Debraj Ray, 2009. "Aspirations, Inequality, Investment and Mobility",
      Georgetown University and New York University, mimeo.
Jones, Benjamin F., and Benjamin A. Olken, 2005. "Do Leaders Matter? National Leadership
       and Growth Since World War II", Quarterly Journal of Economics 120 (3), pp. 835-864.
Karandikar , Rajeeva, Dilip Mookherjee, Debraj Ray, and Fernando Vega-Redondo. 1998.
      "Evolving Aspirations and Cooperation." Journal of Economic Theory 80: 292-331.
Karlsson, N., P. Dellgran, B. Klingander, and T. Garling. 2004. "Household consumption:
       Influences of aspiration level, social comparison, and money management." Journal of
       Economic Psychology 25: 753-769.
Kling, Jeffrey R., Jeffrey B. Liebman, Lawrence F. Katz, 2007. "Experimental Analysis of
       Neighborhood Effects." Econometrica 75: 83-119.
Kremer, Michael, and Edward Miguel, 2007. "The Illusion of Sustainability", Quarterly Journal
      of Economics, 2007, 112(3), 1007-1065.
Lalive, Rafael, and Alejandra Cattaneo, 2006. "Social Interactions and Schooling Decisions",
       IZA Discussion Paper No. 2250.



                                              31
Latham, Gary P., and Lise M Saari, 1979. "The Importance of Supportive Relationships in Goal
      Setting." Journal of Applied Psychology, 64, 151-156.
Locke, Edwin A. and Gary P. Latham, 2002. "Building a Practically Useful Theory of Goal
       Setting and Task Motivation. A 35-year Odyssey", American Psychologist. 57(9): 705-
       717.
Macours, Karen, Norbert Schady, and Renos Vakis, 2008. "Cash Transfers, Behavioral Changes,
      and Cognitive Development in Early Childhood: Evidence from a Randomized
      Experiment." World Bank Policy Research Paper, 4759.
Maluccio, J., and R. Flores. 2004. "Impact Evaluation of a Conditional Cash Transfer Program:
      The Nicaraguan Red de Protección Social." Unpublished manuscript, Food and Nutrition
      Division, International Food Policy Research Institute, Washington, D.C.
Manski, Charles F., 1993. "Identification of Endogenous Social Effects: The Reflection
      Problem." Review of Economic Studies 60: 531-542.
Microcredit Summit Campaing, 2009. http://www.microcreditsummit.org/
Mookherjee, Dilip, Stefan Napel, and Debraj Ray. 2009. "Aspirations, Segregation and
     Occupational Choice." Journal of the European Economic Association, forthcoming.
Nguyen, Trang. 2008. "Information, Role Models and Perceived Returns to Education:
      Experimental Evidence from Madagascar". mimeo, MIT.
Munshi, Kaivan, 2004. "Social Learning in a Heterogeneous Population: Technology Diffusion
      in the Indian Green Revolution." Journal of Development Economics. 73(1): 185-215.
Radloff, L.S. 1977. "The CES-D scale: A self-report depression scale for research in the general
       population." Applied Psychological Measurement, 1, 385-401.
Ray, Debraj, 2006. "Aspirations, Poverty and Economic Change", in Banerjee, Abhijit V.,
      Roland Benabou, and Dilip Mookherjee (eds.), Understanding Poverty, Oxford
      University Press, Oxford.




                                              32
33
34
35
36
37
38
39
40
41
42
43