Jeremy Shapiro and Johannes Haushofer

NexThought Monday – Cash Transfers, Prejudice and Mental Health: MIT researchers discuss some overlooked aspects of their RCT on GiveDirectly

Many governments around the world make direct transfers to poor people. In the U.S., the SNAP (Food Stamps) program is slated to transfer an average of $127 per month to 46.9 million people in FY 2014, at a cost of just under half a percent of GDP, and the Earned Income Tax Credit disbursed an average of $195 per month to 27 million families. Similarly, many developing countries operate such transfer programs; a prominent example is Oportunidades, a conditional cash transfer program which reaches 6.5 million families in Mexico.

Such “handouts” to the poor are a favorite target of criticism, especially during business cycle contractions. A primary concern that critics voice is that direct transfers, whether in kind or in cash, dilute incentives to work and instill a sense of dependency in recipients. Some also worry cash might be spent on alcohol, tobacco or other potentially harmful activities.

There are two problems with these criticisms. First, they are often based on prejudice rather than fact, and we should ask whether they stand up to scientific scrutiny. Second, even if they were true, they may be outweighed by benefits of cash transfers that we currently do not fully understand. For instance, poverty is strongly associated with stress, depression, and generally poor mental health; if cash transfers can address these problems, this should be added to the list of their virtues in weighing up their merits.

To find out whether the objections to cash transfers are really true in practice, and whether they have mental health benefits, we recently performed a randomized controlled trial on the program of the charity GiveDirectly, which uses cell-phone money transfer technology to make unconditional cash transfers to poor families in Kenya. (Full disclosure: Jeremy Shapiro, one of the researchers, is a co-founder and former director of GiveDirectly.) Families receive an average of $700, or about five months of household consumption, through the mobile money system MPesa, and are told explicitly that they are free to spend the money however they wish. In our study, the flip of a coin decided which families received transfers and which did not, and the comparison of these “treatment” and “control” groups gives us a rigorous understanding of what impact the money has.

To measure impacts, we conducted an extensive survey (up to six hours long) with the treatment households, receiving the transfers, and the randomly selected control households that did not. In the survey, we collected data that allows us to address the economic dimension of the concerns noted above: that transfers cause people to stop working, waste money or not earn as much as they otherwise could. This data sheds light on the question of whether these concerns are valid, and adds to a growing body of research on these aspects of cash transfers.

But importantly, to understand the full breadth of the impact of cash transfers, we went beyond looking just at economic impacts and also collected extensive data on the mental health of members of treatment and control families. We even collected levels of the stress hormone cortisol. This is not only helpful because cortisol levels are good indicators of stress and depression, but also because with cortisol we don’t have to worry about transfer recipients telling us “what we want to hear” when we ask them about their well-being.

Before turning to the mental health results, first let’s cover the economic aspects. The results of our study argue against the view that direct transfers dilute incentives to work or spend money on “temptation” goods. We find that recipient households use 39 percent of the transfer to increase their investment, mainly in livestock holdings, which increase by 51 percent or USD 85, and in the quality of recipient’s homes. Concomitantly, revenue from animal husbandry increases by 48 percent (PPP USD 2 per month), and total revenue from self-employment increases by PPP USD 11 per month (38 percent) as a result of the transfers. Furthermore, we don’t observe changes in labor supply, i.e. recipient households are no less likely to work than non-recipients. Thus, at least in our study, it does not appear to be the case that transfers made people lazy; in contrast, they increased revenue from business activities and agriculture. In addition, we don’t see any increases in spending on alcohol, tobacco or gambling.

That cash transfers appear to do no harm on economic dimensions is encouraging. Our results go beyond this, however, and provide a look at the psychological dimension of these concerns. To start with, consider the worry that cash transfers to the very poor will instill in them a sense of dependency. To understand this phenomenon from the psychological perspective, we borrowed a useful construct from psychologists called locus of control, which measures to what extent an individual feels in charge of their life (internal locus of control) vs. perceives their life outcomes to be determined mainly by events beyond their control (external locus of control). We find that receiving a transfer does not alter the recipients’ locus of control. This suggests that cash transfers to very poor households does not alter their perception of their own ability to control and change their situation, and thus may partly explain the fact that we don’t observe them changing the level of their level of effort in agriculture and business.

On the flipside, we observe large increases in mental health among transfer recipients on standard psychological questionnaires. Transfers lead to a large increase in happiness (0.18 standard deviations on an established measure of happiness), life satisfaction (0.15 standard deviations) and a reduction in stress (0.14 standard deviations). Recipients also show a large and significant decrease in depression scores. Of course we might worry that these improvements are due to people telling us what we want to hear. But we find that large transfers (of approximately $1,500) lead to a reduction in levels of the stress hormone cortisol, which is correlated with depression – this measure does not suffer from the response bias problem. Thus, not only do transfers not appear to create a sense of dependency, they lead to large and highly significant improvements in mental health and associated biomarkers.

These mental health improvements as a result of transfers are, of course, important welfare outcomes in their own right. But their significance goes further: adverse mental health also has large economic costs, both in developing and developed countries. For instance, depression is the leading cause of disability worldwide. Given that dysfunction in the workplace is among the diagnosis criteria for depression, it is not surprising that the costs of depression in terms of absenteeism and lost productivity are about half a percent of US GDP (about $51 billion) – a figure that doesn’t include treatment costs.

Thus, in the discussion over the benefits and drawbacks of direct transfers to the poor, the critics should not forget the indirect benefits of such “handouts” in terms of mental health. Transfers have clear economic benefits by allowing people to function normally, both at home and in the workplace. They must therefore be weighed against the potential disincentive effects of transfers and in the cost-benefit comparison to alternative uses of resources aimed at bettering the lives of the poor.

Jeremy Shapiro is a co-founder and former director of GiveDirectly. Johannes Haushofer is a Prize Fellow in Economics at Harvard University and a Postdoctoral Fellow at the Abdul Latif Jameel Poverty Action Lab at MIT. The views expressed here do not necessarily represent the views of GiveDirectly.

Health Care, Impact Assessment
cash transfers, impact measurement, poverty alleviation, research