New research busts the myth of welfare dependency

Social safety nets worldwide routinely come under attack by critics wielding an argument that is as misleading as it is familiar. Measures such as subsidized health insurance, food and nutrition programs, and targeted cash payments to the poor, it is said, incentivize idleness, encourage freeloading, and create a culture of dependency. In response, policymakers cut funding, allow inflation to erode benefits, and make social programs harder for people to access.

In the United States a generation ago, President Bill Clinton’s promise to “end welfare as we know it” assumed that income support to the needy generates indigence. Accordingly, his administration drastically reduced transfers and benefit durations, and introduced stiffer eligibility requirements. At the same time, social programs began to include mechanisms to compel labor-market participation, by cutting benefits for able-bodied adults who proved unable or unwilling to find work. The very name of one key new program, Temporary Assistance for Needy Families (TANF), emphasized the expectation that support would not be indefinite.

Today, the fixation on dependency and its consequences is no less acute. Following a new directive by the Trump administration, Kentucky, Arkansas, and 14 other US states have announced or introduced work requirements as a condition of eligibility for Medicaid (public health insurance for the poor). According to the administration’s statement on the future of the safety net in the US, the goal is to “lift our citizens from welfare to work, from dependence to independence.”

But the idea that government assistance drives dependency is not unique to any country, even if all countries face unique challenges in providing safety nets for the poor. Moreover, beliefs about dependency are not just common among the rich; one often hears similar complaints from the very people whom social programs are meant to help. It is thus little wonder that such beliefs would translate into policy.

That is why it is important to understand how much citizens’ attitudes about dependency affect the safety net in their respective countries. Using the World Values Survey, my colleagues and I have assessed how much people attribute poverty to laziness, as opposed to social and economic unfairness, and how it relates to beliefs on redistribution. We find that the more people attribute poverty to a lack of willpower, the less generous the transfer system in their country will be.

Safety net or trampoline?

So, beliefs about dependency can have real and tangible implications for the poor and the protections they need. But what if those beliefs are wrong? For example, far from creating dependency, it is possible that welfare programs actually give people the necessary tools to achieve financial independence, provided that the assistance is dependable rather than sporadic and temporary. In that case, the provision of government assistance over an extended period of time could yield high social and economic returns, not least by allowing low-income families to make longer-term investments for the future.

To determine if social programs lead to dependency or independence, my coauthors and I studied the effects of Indonesia’s cash-transfer scheme, Program Keluarga Harapan (“Hopeful Family Program”). With the help of the World Bank, the government of Indonesia launched PKH as a large-scale policy experiment in 2008. The program was implemented in 180 randomly selected sub-districts, which were compared to a control group of 180 sub-districts that did not have the program. All told, 14,000 households were surveyed to assess the program’s outcomes.

PKH provides quarterly cash transfers to the country’s poorest households, roughly meaning those within the bottom 7% of the income distribution. Payments constitute 7-14% of a recipient’s income, so they are not meant to cover all of a household’s needs. Moreover, the program was directed at families, which were encouraged to use the benefits to invest in their children. Only households with children or a pregnant woman could enroll, and a portion of the stipend was made conditional on fulfilling various health- and education-related obligations, such as basic immunization and the completion of at least nine years of school. As in many countries, these conditions are hard to enforce in practice, so many households received full payments despite non-compliance.

One important feature of PKH is that it did not merely provide a few weeks or months of assistance between jobs or in the case of a financial shock. Rather, it focused on the very poor, and was administered for at least six years, with the understanding that climbing out of poverty takes time and requires consistent support and stability.

Putting transfers to the test

In 2011, a study of PKH’s initial effects showed that it had a positive impact on short-term indicators of health and educational outcomes after about two years. For example, the program was found to have increased recipients’ visits to post-natal care facilities, as well as their enrollment of children in elementary and middle school.

Given this initial success, the Indonesian government expanded the program widely over the next few years. By 2013, it was providing assistance to about 2.3 million households in 3,400 sub-districts across the country. Now, however, the government was targeting specific districts, rather than following the previous random-selection process. As a result, many of the sub-districts in the initial control group were left out, and have not received the program yet.

This offered us a rare opportunity to understand what happens when poor households receive continued assistance beyond the scope of their immediate needs, and into what researchers call the “medium run” (in this case, six years). Upon re-surveying the 14,000 households in the original treatment and control groups, we found several interesting outcomes.

The first concerns stunting, or impaired growth, which is one of the most serious child health problems in Indonesia. Because children grow slower when they are malnourished, a child’s height relative to their age can serve as a proxy measure of nutrition. Research has shown a correlation between stunting, lower IQs, and poorer socioeconomic outcomes later in life.

At the two-year mark, PKH had no impact on child stunting. And yet, because height is a measure of health that expresses itself cumulatively over time, it was possible that stunting would start to be reversed only after continued assistance from the program. That was precisely what happened. At the six-year mark, children whose families had regularly received extra assistance from PKH were 23-27% less likely than those in the control group to experience stunting.

Back to school

We found similar effects with respect to education. At the two-year mark, PKH had increased school enrollment for children aged 7-15, but not for those aged 15-17. At the time, we reasoned that older children who had dropped out prior to the program would have a harder time returning to school than would their younger counterparts, even if their family resources had recently improved.

But we suspected that if families could benefit from sustained access to the program, their kids would not drop out at an earlier age, with enrollment among those in the 15-17 age bracket thus improving over time. Again, this turned out to be the case. At the six-year mark, children whose families started receiving PKH benefits when they were around 9-11 years old (meaning they were now 15-17 years old) were about 16% more likely to be enrolled in school.

The importance of these improvements in health and education can hardly be overstated. Leaving aside the obvious moral arguments for guaranteeing children access to nutrition and schooling, these investments in low-income households will likely translate into far-reaching economic benefits, including increased labor-market participation and productivity. And that, in turn, could lead to reduced participation in social programs themselves.

In short, Indonesia’s cash-transfer program yielded significant improvements in some of the most stubborn and problematic areas of public health and education. More to the point, these gains were made possible by a cumulative investment in children over the course of six years. Most likely, these results would not have been achieved under a program providing temporary or sporadic benefits.

Welfare as we know it

But the question remains: do such programs create a “culture of dependency”? In the case of PKH, children who grow up healthier and more educated will arguably be better positioned to earn higher incomes and work longer into the future. Of course, to confirm this empirically, we would need to follow the original sample population into adulthood as they enter the workforce and develop careers.

In the meantime, critics will doubtless argue that the program creates dependency for able-bodied adults today. And they might even contend that the culture of dependency will be reproduced inter-generationally, as parents’ attitudes about not working and “being on the dole” are passed down to their children.

Yet here it is worth taking a step back and exploring where the idea of welfare dependency comes from in the first place. We know that beliefs about dependency, laziness, and voluntary unemployment among the poor are pervasive. But why is that?

The primary culprit seems to be classical economic theory, which predicts that when governments provide benefits, individuals may decide that they can afford to work less (economists refer to this as the “income effect”). Similarly, if recipient worry about losing eligibility for benefits if they earn more, they may abstain from work. It is this moral hazard that supposedly leads people to remain poor and rely on welfare indefinitely.

But the evidence does not always support this theory. In another study, my colleagues and I re-analyzed data from seven different experimental trials of government cash-transfer programs throughout the developing world, from the Philippines to Morocco to Mexico. We found that in most cases, men who received benefits tended to be working already, and that there was no evidence that systematic income support reduced work. In an even more recent study, Sarah Jane Baird, David J. McKenzie, and Berk Özler of the World Bank undertook a systematic review of the economics literature on this topic and came to a similar conclusion.

As for PKH, we did not find that program recipients stopped working, even after six years of receiving cash transfers.

From "dependency" to dependable

To be sure, some social programs might well reduce work. Obviously, policymakers should consider the downstream effects of public benefits on labor markets and other areas of the economy. At the same time, the way we think and talk about “dependency” needs to change. The claim that transfers necessarily reduce work may hold true in Economics 101; in the real world, much depends on context and how policies are designed and implemented in practice.

It’s time to bring our beliefs about dependency into line with the data. There is an extensive and growing body of evidence from around the world showing that even very simple cash-transfer programs need not have adverse effects on work.

More broadly, we should focus less on policing “freeloaders,” and more on giving poor families the type of dependable financial assistance that will allow them to make substantial long-run investments in the health and education of their children. As we have seen, the returns on such investments are real, and will accumulate over time.

If we are willing to cast aside the prevailing beliefs about poor people’s “dependency” and “laziness,” we can start to make the social safety net a springboard of upward social mobility.