The Importance of the Counterfactual

Justin Wolfers, a University of Michigan economist and left-of-center public intellectual, recently tweeted a chart on the number of people lifted above the poverty line by various social programs, including Social Security (25.9 million), refundable tax credits (9.8 million), SNAP benefits (4.7 million), Unemployment Insurance (3.8 million), Supplemental Security Income (2.8 million), and housing assistance (1.3 million). Wolfers observes that “the social safety net raises a ton of people out of poverty,” and many advocates of expanding safety net programs have replied approvingly.

But there is a problem with this line of thinking, which is that it neglects the counterfactual. In the absence of these programs, how might individuals, families, and communities change their behavior, or organize their affairs differently? In the absence of Social Security, it is at least possible that workers would save more for their own retirement, and that older Americans would retire later in life than they do at present. The children of aging parents might also increase their labor supply to help provide for them. It is also possible, however, that there would be no increase in labor supply or savings, and that tens of millions of Americans would find themselves destitute. To note that Social Security raises household incomes doesn’t really tell us very much, because older Americans know that Social Security exists, and they respond accordingly. I’m quite confident that poverty among older Americans would increase substantially if we were to abolish Social Security tomorrow. Would poverty among older Americans be extremely high in a world in which the Social Security program had never been established? Might new savings vehicles have been created? Might tight-knit multigenerational families be more commonplace in this world? These are impossible questions to answer. 

Or consider the fact that less-skilled unauthorized immigrants appear to be more likely to migrate from high-unemployment regions to low-unemployment regions in response to an employment shocks than native-born workers, regardless of skill level. One reason this might be true is that households headed by unauthorized immigrants tend to have more limited access to safety net programs than the native-born poor or authorized immigrants, and so they have little choice but to uproot themselves in search of paid work. This raises another counterfactual possibility: in the complete absence of safety net programs, economic desperation might force people who command low wages to work much longer hours, or to take on more dangerous or demanding jobs that command somewhat higher wages. Would this be an unambiguously good thing? I certainly don’t think so. Yet this is the kind of possibility we have to take seriously when we consider the impact of safety net programs on market incomes. 

There are very few serious people who will deny that some safety net programs can reduce labor supply under some circumstances. Indeed, one could argue that this aspect of the safety net is morally attractive. One of the benefits of Obamacare, according to its advocates, is that it frees some older Americans from having to work to retain their health insurance, a belief that accepts that Obamacare will tend to reduce labor supply (and, by extension, market income), yet that this is ultimately a good thing. Does Obamacare lift people who would work in its absence out of poverty? Does it plunge them into poverty by encouraging them to work fewer hours, or to stop working entirely? 

To liberals, the fact that safety net programs raise household incomes is evidence that these programs are working as they should, and that the chief problem we face as a society is our unwillingness to spend more on these programs. To conservatives, in contrast, the fact that market incomes for so many households are so low suggests that Great Society programs have failed. How can that be? Because to conservatives, the purpose of these programs is not solely to alleviate the effects of poverty, but rather to help people enter the workforce, gain experience, and command higher market wages over time, so that they could support their families and their communities. Left-liberals in particular dismiss this conservative view as moralistic and naive, on the grounds that there is nothing particularly ennobling about work, and that what poor people really need is money, not work, work experience, and the subjective feeling that comes with supporting oneself through work. In responding to Wolfers chart, a left-liberal might say that what we really need to do is write everyone, or rather everyone below a certain income threshold, a check so that we can lift even more people out of poverty. A conservative, in contrast, might argue that Wolfers is failing to understand what it really means to be raised out of poverty — that what it really means is to achieve some modicum of economic independence.

This is why you’ll often find that thinkers on the right maintain a distinction between in-work benefits like the earned-income tax credit (EITC), which is explicitly designed to increase labor force participation and work hours, and other social programs that are primarily designed to meet the basic needs of low-income households, regardless of employment status. There are at least some conservatives, myself included, who believe that it would be better to have a safety net geared towards getting people back on their feet than the safety net we have today, even if it were far more expensive. But to left-liberals, talk of personal responsibility and work requirements seems hateful and punitive. And so we go back and forth in circles. 

P.S. Back in 2004, the researchers Gary Englehardt and Jonathan Gruber (yes, that Jonathan Gruber) investigated the role of Social Security in reducing poverty among older Americans, and they did a careful job of teasing out how individuals responding to increased benefits. Did they work and save less in response to higher benefits? They find that increased benefits account for virtually all of the decrease in elderly poverty over the period they study. They also observe that “there were stark changes in the living arrangements of the elderly over the time period we study, with a large shift in living with others to living independently.” This implies that increased Social Security benefits had a more dramatic effect in reducing poverty among families than among households, as increased benefits led older individuals and couples to establish their own households, which tended to be poorer in relative terms. How we feel this shift in living arrangements depends to some extent on how we feel about multigenerational living arrangements. Some see the turn away from these arrangements as a great moral advance while others see them as having contributed to the erosion of social capital. Around the same time, Michele Boldrin, Mariacistina De Nardi, and Larry E. Jones found that increases in old-age pensions were strongly correlated with reductions in fertility, another dimension of the safety net question that is sure to spark controversy. 

Reihan Salam — Reihan Salam is executive editor of National Review and a National Review Institute policy fellow.

Most Popular


Fire the FBI Chief

American government is supposed to look and sound like George Washington. What it actually looks and sounds like is Henry Hill from Goodfellas: bad suit, hand out, intoning the eternal mantra: “F*** you, pay me.” American government mostly works by interposition, standing between us, the free people at ... Read More
Film & TV

Black Panther’s Circle of Hype

The Marvel Cinematic Universe (MCU) first infantilizes its audience, then banalizes it, and, finally, controls it through marketing. This commercial strategy, geared toward adolescents of all ages, resembles the Democratic party’s political manipulation of black Americans, targeting that audience through its ... Read More