This is the first in this year’s series of posts by PhD students on the job market.
A group of recent experimental studies have shown that people take advantage of information asymmetries within households to hide financial decisions from other members of their families (Ashraf, 2009; Schaner, 2012). This behavior may be especially prominent in migrant households that are geographically split (de Laat, 2008; Chen, 2013) but whose finances are linked through the sending and spending of regular remittance payments. Understanding how information imperfections may affect decisions about remittances is important both because remittances have been shown to have positive impacts for development and because they are so substantial, equal to 16 percent of GDP in El Salvador in 2010 (Ratha and Silwal, 2012).
In my job market paper, I examine the direct, causal effect of information asymmetries on decisions about the sending and spending of remittances. I make two main contributions. First, while a number of papers have recognized that migrants might be unable to control how recipients spend the remittances they receive (Ashraf et al., 2011; Chin et al., 2011), these papers do not recognize that information problems may also run in the other direction. I study both sides of the migrant-recipient relationship, examining how information issues can affect both how much the migrants send home in remittances and how the recipients spend what they receive. Second, I examine two different types of information problems and show that both can be important. The first are those that may result in strategic behavior: migrants may underreport their income to reduce the remittances they are expected to send home, and recipients may misrepresent the ways in which they spend the remittances sent by the migrant. The second type consists of those information problems that can inadvertently affect decisions: recipients may be unaware of the migrants’ actual preferences for how remittances should be spent.
I conduct an experiment in a matched sample of migrants from El Salvador and their family members at home. Migrants were recruited at the Salvadoran consulate in Washington, DC and interviewed while waiting for consular services. My experiment was conducted at the end of a baseline survey that included information on demographics, family relationships, and remittances. Because the baseline survey was also used for a separate field experiment on remittances and education, migrants were required to have high school or college aged relative in El Salvador in order to participate. Following the migrant survey, a family member in El Salvador identified by the migrant was contacted by phone to complete a survey, and again, my experiment was conducted at the conclusion of that survey. 82 percent of family members were successfully interviewed, and the final sample is 1,298 migrant-recipient pairs.
The migrant experiment
The migrant experiment consisted of migrants making a remittance sending decision. Given the possibility of winning $600, they were asked how much they wished to keep and how much they wished to send home to the recipient. Some migrants were subsequently chosen by lottery to win the $600 prize and the allocation they chose was implemented. Migrants were randomly allocated into two groups: those who were told that their choice would not be revealed to the recipient and those who were told that their choice would be revealed. If migrants send less money home when their choice is kept secret from the recipient, that is evidence that they take strategic advantage of information imperfections, for example the opportunity to hide income from their family member.
The results show that migrants do indeed send less of the $600 prize home when their choice is not revealed to recipients. Although all migrants send home large amounts (more than half of both treatment groups send the entire $600), those whose choice is not revealed send five percent less than those whose choice is revealed. However, there is strong heterogeneity in the results. The treatment effect is only evident in migrant-recipient pairs where the recipients have the ability to punish migrants for reneging on agreements about how much of their income they should send home. The intuition is that migrants only take advantage of the opportunity to hide income when they are being pressured by recipients to send more home than they would choose to send voluntarily. I represent this punishment ability through the use of five proxy variables from the baseline survey. Although all are imperfect measures, I argue that the consistent pattern of treatment effects concentrated among high punishment ability pairs is convincing.
The recipient experiment
The recipient experiment consisted of recipients making a remittance spending decision. Recipients were told that because the migrant had participated in the study they (the recipients) had the chance to win a remittance worth $300. Recipients were then asked to allocate the $300 among four broad spending categories: restaurant meals, daily expenses, education, and health. Lottery winners received the allocation that they selected in-kind. The recipient experiment included two separate treatments:
· The recipient monitoring treatment: In a parallel treatment to what was performed with migrants, half of the recipients were told that their choice would not be revealed to the migrant and the other half were told that their choice would be revealed. If recipient choices more closely mirror migrant preferences when the recipients’ choices are revealed, that is evidence that recipients may strategically seek to hide their spending behavior from the migrant.
· The recipient communication treatment: During the migrant survey, migrants were asked what their preference for the recipient choice was. At the time that the recipients make their decision, half are informed of those migrant preferences and the other half are not. If recipient choices are closer to the migrant preferences when they are informed of those preferences, it is evidence that recipients may be inadvertently deviating from migrant wishes simply because they do not fully understand them.
The results show that there is no effect of the recipient monitoring treatment on recipient choices. In other words, in this context, recipients do not take strategic advantage of the opportunity to spend a remittance without the migrant knowing. There is, however, an effect of the communication treatment. Recipients told about migrant preferences reduce the difference between their choices and migrant preferences by about 10 percent. This implies that migrant preferences do matter to recipients and that some deviation from those preferences may be inadvertent.
Discussion and policy implications
Although most studies that address information imperfections in migrant households have focused on how recipients spend remittances, the results of my study show that, in this experimental context, it is only migrants who react strategically to whether or not their choice will be monitored. This demonstrates that recipients have important influence in the migrant remittance decision, something that has not been previously shown empirically. Understanding that recipients have this influence is crucial for the design of new initiatives that seek to increase remittances and enhance their development impact. At the same time, I also find that recipient decisions are altered by information about migrant preferences. This result suggests that policies or products that improve communication may be a low-cost way to improve recipient compliance with migrant preferences.
Kate Ambler is a Ph.D. student at the University of Michigan.