Month: March 2025

Here are three main often-asked questions about the US safety net:

1) If your goal is to expand the social safety net, are you more likely to be successful with a focus on universal programs (like Social Security) or means-tested programs (like food stamps or welfare)?

2) Back in 1996, President Bill Clinton signed what is often called the “welfare reform act,” following up on his campaign promise to “end welfare as we know it.” For better or worse, did that legislation keep Clinton’s promise?

3) Those who need social safety net programs are often at a point where their live are unstable: perhaps an unstable job, unstable housing, unstable family arrangements, unstable income, even unstable health. For people in this situation, finding out about available programs, signing up for them, and keeping track of changes in requirements to that they can remain eligible, may be difficult. Are the administrative burdens that safety net programs place on recipients a necessary evil, or something else?

For answers to these and many related questions, I recommend the three-paper symposium on the US safety net in the Winter 2025 issue of the Journal of Economic Perspectives (where I work as Managing Editor).

• “Two Histories of the Public Safety Net,” by Christopher Howard
• “Did Welfare Reform End the Safety Net as We Knew It? The Record since 1996,” by Lucie Schmidt, Lara Shore-Sheppard, and Tara Watson
• “Administrative Burdens in the Social Safety Net,” by Pamela Herd and Donald Moynihan

Howard discusses the poverty-reduction history of means-tested and broader social insurance programs over the last century or so. He reminds readers of the old saying “programs for the poor are poor programs.”

Given the stark contrasts between means-tested “welfare” and inclusive Social
Security, it is not surprising when advocates for paid family and medical leave (Romig
and Bryant 2021) or long-term care (Powell 2019) endorse a social insurance
model. Nor is it surprising when they endorse Medicare for All (Scott 2019). Part of
me is attracted to these proposals, for reasons that range from lower administrative
costs to social solidarity. I am also attracted by the possibilities of giving extra help
to lower-income households within broadly inclusive programs. However, recent
history does not point decisively in favor of inclusive over means-tested programs.
In retrospect, the biggest achievements of social insurance, politically and
programmatically, occurred between the 1930s and 1970s. Since then, Social
Security and Medicare have had more success preventing retrenchment than accomplishing expansion.

Conversely, the last 30 years or so have seen dramatic increases in spending on a number of means-tested programs: the Earned Income Tax Credit, Medicaid, the Children’s Health Insurance Program (CHIP), the Child Tax Credit, and food stamps (now called the Supplemental Nutrition Assistance Program). Meanwhile, the trust funds that support Social Security and the hospital portion of Medicare are dwindling fast, and those programs face sharp pressures in how to constrain the projected future rise in spending. The evidence of the last three decades or so is that the US political system may be better at expanding means-tested programs, and less willing to expand universal programs, than previously thought.

Schmidt, Shore-Sheppard, and Watson dig into the evolution of specific safety net programs since 1996 in more detail. They point out that when President Clinton signed the welfare reform act into law in 1996, he was fulfilling a campaign promise to “end welfare as we know it.” The legislation in fact did so along several dimensions:

First, spending on conventional “welfare”–that is, the Aid to Families with Dependent Children (AFDC), which was converted into Temporary Assistance to Needy Families (TANF) in the 1996 welfare reform act–has dwindled in real dollars and has pivoted away from cash support to families and toward programs related to reducing poverty (for example, programs to improve high school graduate rates and reduce teen pregnancies). However, as noted a moment ago, total spending on means-tested programs has increased, not decreased.

Second, the earlier “welfare” program was a mixed federal-state program. Indeed, many safety-net programs were designed on a shared federal-state basis, including AFDC, Medicaid, and unemployment insurance. But the new surge of means-tested spending has in substantial part been federal spending–which also has the effect of equalizing support for the poor across states.

Third, many of the expanded means-tested programs have linkages to working parent, or to benefits aimed at children. As a result of this targeting, low-income working families have become better-off. However, low-income single people, or children in families where the adults do not have a job (and thus can be the poorest of the poor), may well be worse off.

Finally, Herd and Moynihan discuss administrative burdens of the safety net. Here’s an illustrative anecdote from the start of their paper:

By the age of five, Abel Sewell had survived cancer and was receiving monthly blood tests to ensure his leukemia had not returned. The tests were covered by TennCare, Tennessee’s Medicaid program, until, at a regular doctor’s visit, his mother discovered that the coverage had lapsed. Abel’s mother spent months fighting to get their coverage restored, ultimately taking out a second mortgage on their home to manage the mounting health care debt that came from being uninsured. The family very much wanted health services, and were willing to endure significant hardships to get it.

The Sewells were not an anomaly. Between 2016 and 2019, almost 250,000 children in Tennessee lost coverage. What happened? The Sewells, like many others, said they never received the TennCare renewal forms. Even those who did receive the renewal packets often struggled to complete the 47 pages (Kelman and Reicher 2019). Failure to return forms accounted for 67 percent of those who lost coverage, and it seems likely that many of them did not receive the forms due to outdated mailing addresses. Late or incomplete forms also resulted in a loss of coverage (Arbogast, Chorniy, and Currie 2022).

Full disclosure: The issue of administrative burdens is personal for me. My wife and I are the legal guardians for one of my adult children, who is disabled, and thus receives federal Disability Insurance (which means he is also on Medicare for health insurance) and state-level support through a Minnesota program. We are deeply grateful that such support is available, and the benefits make it possible for our son to live independently (although we live close by). But the paperwork burden is heavy, even for a family with two college-educated parents. We strongly suspect that a number of people who are eligible for the same benefits as my son are not actually receiving these benefits, because the legal guardians of those people can’t handle the paperwork burden.

As Herd and Moynihan point out, part of the problem here is “sludge,” the gradual accumulation of rules and forms that accumulate in a bureaucracy. As a simple example, imagine that there are two safety-net programs, and a large majority of those who qualify for one will also qualify for the other. One program has a work requirement; one does not. Should we add a work requirement to the second program? As a practical matter, adding the additional work requirement to the seconed is redundant: again, those in the second program already face a work requirement for the first program. But in a bureacracy, there is a natural tendency to reason along the lines, “they already meet the work requirement for one program, so why not add it for the other program, too?” Multiply this tendency, and you end up mailing out 47-page forms for annual renewals. For our family, a single 47-page form per year would be an enormous improvement over the administrative burden we actually face.

On average, those families with the lowest level of resources will have the hardest time meeting them. In that sense, adding to the paperwork burden for safety-net programs is a regressive method of rationing their

For the US population as a whole, 11.1% of the population is below the poverty line, and an additional 15.8% are above the poverty line, but have income of less than twice the poverty line. For those under 18 years of age–for whom the advice to “get a job” has zero practical relevance–15.2% of the population lives in a household below the poverty line, and an additional 19.8% are in a household above the povery line but with an income less than twice the poverty line. At its best, the safety net is about helping this substantial part of the US population to have sufficient resources to persevere through the hard time, and to have the space to envision and work toward better prospects in the future.

One of my long-ago professors–not an economist, and not a political conservative– sometimes said that Adam Smith was just flat out deeper and more interesting than many of his critics, who often try to reduce him to a cardboard cutout disciple of free-market fundamentalism. For example, I’ve heard (uninformed) criticisms of Smith that he assumes everyone wants to buy and sell, when a number of people instead would prefer to dominate and take. Paolo Santori engages with this corner of Smith’s work in “Domination vs. Persuasion: The Role of Libido Dominandi in Adam Smith’s Thought” (The Review of Politics, 2025, 1–18).

Santori seems to prefer the Latin libido dominandi, but as he points out, Smith writes of “love of domination” and “love of domineer.” Here’s Smith’ discussion of desire to dominate, in the context of masters who want to dominate slaves, from The Wealth of Nations (Book III, Chapter 2). Smith wrote:

The experience of all ages and nations, I believe, demonstrates that the work done by slaves, though it appears to cost only their maintenance, is in the end the dearest of any. A person who can acquire no property, can have no other interest but to eat as much, and to labour as little as possible. Whatever work he does beyond what is sufficient to purchase his own maintenance can be squeezed out of him by violence only, and not by any interest of his own. In ancient Italy, how much the cultivation of corn degenerated, how unprofitable it became to the master when it fell under the management of slaves, is remarked by both Pliny and Columella. In the time of Aristotle it had not been much better in ancient Greece. …

The pride of man makes him love to domineer, and nothing mortifies him so much as to be obliged to condescend to persuade his inferiors. Wherever the law allows it, and the nature of the work can afford it, therefore, he will generally prefer the service of slaves to that of freemen. The planting of sugar and tobacco can afford the expence of slave-cultivation. The raising of corn, it seems, in the present times, cannot. In the English colonies, of which the principal produce is corn, the far greater part of the work is done by freemen. … In our sugar colonies, on the contrary, the whole work is done by slaves, and in our tobacco colonies a very great part of it. The profits of a sugar-plantation in any of our West Indian colonies are generally much greater than those of any other cultivation that is known either in Europe or America; and the profits of a tobacco plantation, though inferior to those of sugar, are superior to those of corn, as has already been observed. Both can afford the expence of slave-cultivation, but sugar can afford it still better than tobacco. The number of negroes accordingly is much greater, in proportion to that of whites, in our sugar than in our tobacco colonies.

Santori traces Smith’s ideas about the “love to domineer” across Smith’s other works, like The Moral Sentiments and the Lectures on Jurisprudence. He argues that other authors have sometimes interpreted the “pride” that Smith speaks of as the root of a “love to domineer” as a kind of vanity or a desire for the recognition of others.

Santori argues that a more persuasive interpretation is to think of “pride” in this context as a sin. He quotes Smith in The Moral Sentiments: “”The proud man does not always feel himself at ease in the company of his equals, and still less of that of his superiors.” Santori argues for this kind of pride, there is a pleasure in not needing to spend time or energy persuading or obtaining consent. Indeed, this “love to domineer” is strong enough, in Smith’s argument, that those who hold slaves are willing to give up some of the material benefits they could have from hiring free labor.

In this part of the Wealth of Nations, Smith is discussing the historical transition from feudal to commercial society. In that context, Santori argues:

We read in the Lectures on Jurisprudence (LJ) and Wealth of Nations (WN) that masters’ love of domination is what will make slavery or servitude perpetual, in contrast with masters’ real interest that would be fostered by having free men rather than enslaved people working for them. … Smith argued that the emergence of European commercial society, grounded on free-market exchanges between individuals based on persuasion, marginalized and undermined libido dominandi. However, he knew that commercial society could not eliminate libido dominandi and that, whenever socio-economic circumstances allow, human beings will try to dominate each other. He saw this in the colonies and in specific markets (colliers and salters). …

To Smith, commercial society is a more mature way of conceiving life in common and civil society. In contrast, love of domination expresses a childish wanting to obtain everything without effort. Human beings can flourish when they learn to live in a society where they cannot impose their aims. They must deal with others’ aims and opinions in relations based on persuasion rather than domination. Adult life in a commercial society requires something better than the love of domination. Here, I am expanding Smith’s argument, but hope to have remained faithful to his spirit.

A common complaint about so-called “free markets” is that there are times they don’t feel especially “free,” like when it’s time to go to work in the morning or when the bills are due. Moreover, hierarchies in commercial firms and markets do provide some scope for those who “love to domineer” to do so. But the “love to domineer” doesn’t go away in countries where markets and politics are not free–and can manifest itself in even more distasteful ways.

It’s intuitively obvious that “uncertainty” matters in economic decision-making. If the risks of making a choice–starting a company, making an investment, buying a house–look especially big in the present, then there is reason to postpone that decision. As a result, higher uncertainty can lead to a drop in economic activity. Thus, it’s a concern that, by some measures, economic uncertainty is on the rise.

For example, here’s the Economic Policy Uncertainty Index for the United States, as reported by the FRED website run by the St. Louis Fed. You can see the recent spike on the far right.

Or here’s the Global Economic Policy Uncertainty Index:

These graphs are surely a reason for concern. Whatever the merits of a “move fast and break things” approach in certain contexts, it obviously will increase uncertainty. But how does one measure uncertainty? What is being measured here?

The US uncertainty index is not official government data. It is based on a method developed by three economists, Scott R. Baker, Nick Bloom, and Steven J. Davis. I mentioned their approach here when it was first being developed back 2012. They combine three sources of data: “the frequency of newspaper articles that reference economic uncertainty and the role of policy; the number of federal tax code provisions that are set to expire in coming years; and the extent of disagreement among economic forecasters about future inflation and future government spending on goods and services.” The average value from 1985-2010 is arbitrarily set at 100. Thus, you can see spikes during the Great Recession, the pandemic, and now early in 2025.

For a discussion from Nicholas Bloom about this and other ways of measuring uncertainty, and how they relate to actual economic outcomes, a useful starting point is his article, “Fluctuations in Uncertainty,” in the Spring 2014 issue of the Journal of Economic Perspectives (where I work as Managing Editor).

An obvious concern about measures of uncertainty based (at least partly) on news reports is that there may be a divergence between how the media is covering the news of the day and what actual investors and business people are doing and saying. At least at the moment, there appears to such a divergence.

For example, a standard measure of uncertainty in financial markets is the CBOE Volatility Index from the Chicago Board Options Exchange, commonly called the VIX. The basic idea is to look at expectations of volatility of the stock market, by looking at the options that investors are buying on future values of the S&P stock market index. For example, if more investors are buying options to protect themselves against especially large falls in stock prices, then volatility would be up. But the VIX isn’t showing a rise in uncertainty just now.

Another way to measure uncertainty is ask businesses about their sales and employment forecasts 12 months in the future, and how much uncertainty they feel about those forecasts. The Federal Reserve Bank of Atlanta carries out a Survey of Business Uncertainty with this approach, and it does not show a prominent recent uptick in business uncertainty.

I don’t quite know what to make of these various meausures. The Baker-Bloom-Davis measure of uncertainty has been tested and used in research, and it cannot be casually dismissed. However, other meausures of uncertainty are not spiking int he same way. At the moment, it seems fair to say that there’s uncertainty about uncertainty, which isn’t the same thing as greater uncertainty, but perhaps headed in that direction.

Will workers in an inherently high-stress environment perform better if their bosses seek to defuse that stress? Or if their bosses play up and emphasize the stress? The answer probably depends on specific contexts of the workforce and the boss. But for some evidence and a speculative answer in one context, Oded Stark offers “Stress in the air: A conjecture” (Economics and Human Biology, December 2024). From the abstract:

The 1949 study The American Soldier: Combat and Its Aftermath, Volume II, by Stouffer et al. presents detailed accounts of the attitudes of American fighter pilots toward the stress experienced by them and of the policies and practices of the American Air Force command in addressing this stress during WWII. The 2022 study “Killer incentives” by Ager et al. documents an aspect and a repercussion of the stress of German fighter pilots and can be used to identify the response to that stress by the German Air Force command during WWII. Drawing on these two studies, in this paper I construct fighter pilot stress profiles in the two air forces. The picture that emerges is that there is a stark difference between the approaches of the two commands. This diversity leads me to conjecture that the American Air Force command explicitly sought to forestall and curtail fighter pilots’ stress, whereas the German Air Force command implicitly cultivated and engineered fighter pilots’ stress.

Stark points out that the American Air Force command was very aware of the stress experienced by fighter pilots, and how performance tended to diminish with additional missions. They and tried to address the stress in various ways. One approach was to set a limit: “The limit to the tour of duty of fighter pilots was 300 hours of combat flying, which was typically achieved in six or seven months of active combat duty.” This limit was pre-announced and socially sanctioned. In addition, Stark quotes Stauffer about ongoing evaluation of fighter pilots: “All fighter pilots were systematically examined throughout the entire period that they were on operational duty; as soon as any … anxiety reaction to combat flying was detected, the man was immediately removed from combat duty as a fighter pilot”–although those removed from combat duty could be reassigned to less risky flights. Finally, American flight crews were rewarded in terms of total missions the completed, and medals were typically awarded after the tour of duty was complete, not on whether a particular mission had been especially risky.

The German Air Command during World War II took a different approach. It encouraged rivalry between fighter pilots, and gave decorations and promotions based in part on whether a mission was especially risky. Stark adds: “As noted many times in the Stouffer et al. study, the American army was well aware of and sympathetic to the problem of psychiatric combat breakdown (by 1943 providing treatment for psychiatric casualties, either at forward stations near the front or in dedicated hospitals closer to the rear), whereas the German army was generally hostile to the idea of psychiatric breakdown and those who were considered guilty of malingering or cowardice were not treated well.”

One can easily hypothesize reasons why organizations might take different approaches to stress management. In certain contexts (financial markets?), some kinds of risk-taking might be especially remunerative. In wartime, perhaps an aggressor has reason to encourage risk-taking, while the party fighting back (and expecting a surge of wartime production to arrive) will want to deal with stress differently. Even before the war started, the culture that generated US fighter pilots in World War II might have been quite different from the culture that generated German fighter pilots.

Thus, Stark’s point is not that there is one best approach that organizations should follow for motivating workers in stressful environments. But it can be useful to think explicitly about whether a given work environment seeks to be stress-reducing or stress-increasing–and what tradeoffs can arise.

One of the puzzles of international macroeconomics is that if capital has diminishing returns, then it seems plausible that capital should flow from capital-rich high-income countries to capital-poor low-income countries. After all, the potential returns to capital investment should presumably be high in a capital-poor environment.

However, for some decades now, net capital inflows have been coming into the US economy. Moreover, as the World Bank International Debt Report 2024 points out, since 2022, the interest payments that low- and middle-income countries are making on their external (that is, outside their own country) debts are greater than the amount of new debt capital flowing in. Indermit Gill describes it this way in the “Foreward” of the report:

Since 2022, foreign private creditors have extracted nearly US$141 billion more in debt service payments from public sector borrowers in developing economies than they disbursed in new financing. As this report documents, that withdrawal has upended the financing landscape for development. For two years in a row now, the external creditors of developing economies have been pulling out more than they have been putting in—with one striking exception. The World Bank and other multilateral institutions pumped in nearly US$85 billion more in 2022 and 2023 than they collected in debt service payments.

That has thrust some multilateral institutions into a role they were never designed
to play—as lenders of last resort, deploying scarce long-term development finance
to compensate for the exit of other creditors. Last year, multilateral institutions
accounted for about 20 percent of the long-term external debt stock of developing
economies, five points higher than in 2019. … In 2023, the World Bank accounted for fully a third of the overall net debt inflows going into IDA-eligible countries—US$16.7 billion, more than three times the volume a decade ago.

That reflects a broken financing system. Capital—both public and private—is
essential for development. Long-term progress will depend to an important degree
on restarting the capital flows that most developing countries enjoyed in the first
decade of this century. But the risk-reward balance cannot be allowed to remain
as lopsided as it is today, with multilateral institutions and government creditors
bearing nearly all the risk and private creditors reaping nearly all the rewards.

The report provides a wealth of underlying detail, but here’s one point that stuck with me. If one looks at the ratio of external debt to gross national income, it’s not rising a lot for these low- and middle-income countries in the last couple of years. (IDA refers to International Development Association: it’s the part of the World Bank focused on lending to the lowest-income countries.)

Instead, there was a gradual but substantial run-up of debt for these countries over the last decade or so. A key fact here is that low-income countries typically cannot borrow in their own currency: instead, they commonly borrow in US dollars or sometimes in euros. In addition, they often borrow at adjustable interest rates, while a large developed economy like the United States can typically borrow at fixed nominal interest rates. Thus, when it comes to repaying external debt, low-income countries are vulnerable to higher interest rates and to accompanying shifts in exchange rates. The report notes:

Total debt servicing costs (principal plus interest payments) of LMICs [low-and middle-income countries] reached an all-time high of US$1.4 trillion in 2023. For LMICs excluding China, debt servicing costs climbed to a record of US$971.1 billion in 2023, an increase of 19.7 percent over the previous year and almost double the amounts seen a decade ago. In 2023, LMICs faced historically challenging debt service burdens due to high debt levels, interest rates that hit a two-decade high, and depreciation of local currencies against a strong US dollar. The tightening of monetary policy in the United States in 2022 affected exchange rate movements and drove an increase in the value of the US dollar relative to other currencies, which persisted in 2023 and made repayment of non–local currency debt more costly for LMICs as their local currencies depreciated.

In short, interest payments for the US government are rising as new borrowing at higher interest rates plays a greater role compared to earlier debt accumulated at lower interest rates. But low- and middle-income countries face a double-whammy of their currencies being less valuable on exchange rate markets and also a share of their debts adjusting automatically to higher global interest rates. When those factors hit on top of the greater debt accumulated in the last decade, difficulties can arise. Some countries have already managed to reschedule their debts, but I suspect those are the easy cases–and the hard cases will be coming in the next few years.