Spotlight Exclusives

Moms receiving cash benefits spend no more on cigarettes, alcohol, and opioids, study finds

Greg J. Duncan Greg J. Duncan, posted on

The potential impact of unrestricted cash aid for Americans with low incomes has long been a topic of debate in policy circles, and the COVID-19 pandemic forced an unexpected experiment through programs such as the expanded Child Tax Credit. Offering crucial context for the cash aid debate was the first set of results from the Baby’s First Years project, the first study in the United States to assess the impact of poverty reduction on family life and infant and toddler’s cognitive, emotional, and brain development. BFY has released a new paper looking at whether mothers receiving cash aid through the study spent more on cigarettes, opioids, and alcohol. One of the authors, Greg J. Duncan, an economist at the University of California, Irvine, spoke to Spotlight recently about the results. The transcript of that conversation has been lightly edited for length and clarity. 

We’ve featured other studies from Baby’s First Years, but quickly give us some general background on the project

Baby’s First Years (BFY) seeks to understand the impact of poverty reduction on child development in the early years of life as well as impacts on families. It is a clinical trial, with random assignment, so it avoids the problem in correlational studies where you don’t really know whether income is the active ingredient driving changes you might observe.

We recruited one thousand low-income mothers in four cities: New York, the Twin Cities, Omaha, and New Orleans. We approached them shortly after they had given birth and were still in the hospital. After they consented to participate and responded to our baseline interview, we offered them a chance to receive cash gifts. Those who agreed were then randomized in two groups—one receiving $333 a month ($4,000 a year) on a debit card and the other group receiving $20 a month ($240 a year), also on a debit card. These payments are continuing to the present day.

We’re now in the process of completing our data collection around the time of the children’s third birthdays. We look forward to continuing the study through age four and, if we secure the funding, through age 6 and beyond. Our key focus is on whether children in the “high cash gift group” (those in families receiving $333 per month) differ from children in the “low-cash gift group” (the $20-per-month recipients) in terms of their socio-emotional development, health, and cognition. We measured electrical brain activity at 12 months of age and will be doing so again – plus gathering a host of other developmental measures – when the children turn four.

As I mentioned before, we’re also very interested in group differences in family processes that might explain whatever child impacts we observe. So, for example, we’re asking mothers about their spending on things that might enrich their child’s early learning—buying books and toys and center-based childcare, for example. And we’re looking at whether the poverty-reducing cash payments reduce mothers’ reports of stress, depression, and anxiety, which are thought to affect parenting.

The paper we’re talking about today is based on data on family spending on things not related to child enrichment. In particular, the article we just published examined group differences in expenditures on alcohol and cigarettes, as well as responses to our questions about mothers’ opioid use. It is part of our attempt to take a comprehensive look at family spending patterns.

And this study Greg, as you point out, deals with one of the enduring stereotypes about people getting cash benefits. In fact, Sen. (Joe) Manchin (D-W.Va.) allegedly raised this as a key reason for opposing the extension of the expanded Child Tax Credit.

Right, although framing it that way might help to create or reinforce that stereotype in people’s minds. The most straightforward way of thinking about the article is whether families in the high- and low-cash gift groups differ in their expenditure patterns, including on things like cigarettes and alcohol. We found that the high-cash gift group spent just 39 cents more per week on cigarettes and alcohol, combined – such a small difference that it’s attributable to chance.

And how do you gather that data? Are you getting it through interviews, or can you see how they’re using the money on the debit card?

We asked for permission from the mothers in our study to track their expenditures on the debit cards. Most of them agreed to that but we soon discovered that about one-third of the money deposited every month on their debit cards is taken out in cash from ATM machines, and we don’t know from the transaction records how that’s spent. Moreover, the transaction records record where but not how the money is spent, for example in Big Box store or on We have found that the most reliable way of tracking expenditures is to ask our study participants directly for their estimates of how much money they spent in a typical week on alcohol, cigarettes, and other things. People probably can’t recall those numbers to the penny, but other studies have found that this method gets pretty close to what expenditure actually are. And what’s important in our study is not whether the reports might contain a bit of over- or under-reporting, but whether the expenditures reported by the high- and low-cash gift groups differed, with the composition of the two groups determined by a coin flip.

The last time we spoke, you were still doing most of the interviews, if not all the interviews, virtually. Is that still the case? Are you now more able to talk to people in person?

We’re finishing up our age three interviews by phone in June. In July, when some of the children in our study turn four, we’ll be bringing them and their parents into university labs that we have set up in our four different cities. And in those labs, we be conducting child assessments – including another look at electrical activity in their brains – and interviewing the mothers as well. We’re delighted to be able to get back to in-person assessments, and we hope it stays that way.

Have you found any differences in the data between cities?

We enrolled one thousand mothers in our study. That sounds like a big number, but once you divide them up into the high- and low-cash gift groups, and then divide them further across our four cities, it is very difficult to determine whether site differences we might observe are really there. Started another way, we don’t have very much statistical power to detect differences across sites.

What’s next? What’s the next paper you’re expecting to publish?

We’re working on two family-process papers. The first is looking for group differences in indicators of the stress pathway I mentioned before. The second looks at the investment pathway—enrichment expenditures on children and time spent by the mothers with their children reading stories and playing. We’ve also have an employment paper, which addresses the hot topic of whether, if you give families money, as in the recent Child Tax Credit expansion, the two groups differ in terms of their employment and earnings.

And to be clear, as you’ve pointed out before, what you’re doing is different than the way the expanded Child Tax Credit worked.

That’s correct. In the American Rescue Plan, an expansion of the Child Tax Credit provided $3,000 or $3,600 per child. Part of this was paid out in six monthly payments, with the rest of it paid in the family’s annual tax refund. BFY payments are monthly as well, but the money comes from private philanthropy rather than from the government. But the main difference is that our payments do not increase with family sizes, whereas the expanded Child Tax Credit was a per-child payment.

And were you surprised by this particular result in this paper, or was that about what you expected?

We were not surprised. Past studies of expansions of the Earned Income Tax Credit examined whether any of the increase was spend on alcohol and cigarettes and didn’t find any evidence of that. And there have been there a number of studies of cash allowances like ours in developing countries, which also find either no differences or reductions in these kinds of expenditures.

« Back to Spotlight Exclusives