Author
Keilan Crawford is not the median American. But millions of Americans are like Keilan Crawford.
Crawford is 36 years old and a father of two from Robbinsdale, Minnesota. On the afternoon we met this spring, he’d woken as usual at 4:00 a.m. for the 90-minute bus ride to the most promising job he’s ever had—earning $25 an hour as a facility maintenance helper for the State of Minnesota.
“I knew there was more out there for me,” Crawford said, over coffee, after a shift at his new job. “I just didn’t know the route for how to get there.” His father was often absent during Crawford’s childhood in Chicago. His mother battled addiction. He left home at 15 and later dropped out of high school. Crawford spent his 20s, in the aftermath of the Great Recession, working fast food and other short-term jobs.
“I think, honestly, I did fall into a little, slight depression—you know, with how life was taking me at the time.” Crawford is grateful today for good choices that avoided a criminal record, which makes things even more difficult for many men he knows. “But it’s still hard to find a job—a decent-paying job, you know—with no background, with no kind of education.” The births of his children inspired him to finally forge a long-term plan, and he committed to a building maintenance training program run by a local nonprofit, Project for Pride in Living.
Crawford embodies the complex ways that monetary policy affects the lives of different people differently. On a relatively low income, he struggles with higher prices. But he says lower interest rates—and a raise in pay—could help him buy a car and transform his exhausting daily routine. “I have a part-time job bringing in extra money,” Crawford said. “But that’s still not nearly enough, I feel like, to get a car loan.”
Monetary policymaking rests fundamentally on aggregates and averages. For most of the Fed’s history, it could hardly have been otherwise, given the data and the economic tools available. In recent years, however, modern computing power, a critical mass of reliable data on diverse households, and state-of-the-art models increasingly empower economists to understand how monetary decisions transmit across our varied economy.
Fed policymakers can increasingly ask and understand how the core act of raising or lowering a policy interest rate will ripple out not just to the median worker or consumer, but to an American like Keilan Crawford. Asking the question, of course, is just the first stage in reaching an answer. Like much in economics, answers will come gradually with data, deliberation, and a healthy debate.
New insights in the age of HANK
The technology and the philosophy of central banking are evolving together. The new tools are coming into their own as officials implement the notion of maximum employment as a “broad-based and inclusive goal,” the language adopted in the 2020 monetary policy framework. At the same time, the recent episode of inflation—the other half of the Fed’s “dual mandate”—has launched a wave of research into how different groups of Americans experience changing prices.
Since at least the 1970s, macroeconomists have concerned themselves with “optimal” monetary policy. Until recent years, that generally meant optimal for a “representative agent.” The limits of mathematical models and computing power typically meant policy was assessed for its impact on a hypothetical amalgam of everyone in the economy. So-called “representative agent new Keynesian” (RANK) models explored how policy changes propagate through an economy with imperfections (that’s the “new Keynesian” part), but where household and worker differences were ignored.
In recent years, economists have worked to technically incorporate an obvious point that proved thorny in practice: People are different—“heterogeneous,” in the language of the social sciences. “It has taken many years to make progress on this,” said Eduardo Dávila, a Yale economist and current Institute visiting scholar. “There was earlier work on heterogeneity and there was earlier new Keynesian work. The last 10 or 15 years have been about trying to put both together.” The result is today’s era of “heterogeneous agent new Keynesian” (HANK) models.
Economists still need to make choices about which and how many forms of variation to include. “These models become complicated because you have this whole cross-section of people—you have to keep track of everybody in the economy,” Dávila said. But “when you look at who is better off, who is worse off, how policy impacts different people, absolutely there is no comparison” to previous methods. “Predictions are completely different once you start to put in heterogeneity.”
Dávila says something notable happens to central banks when economists introduce agents with a range of incomes or wealth into a typical HANK model: Those hypothetical monetary policymakers in the model are naturally and rationally inclined to give more weight to poor people than rich ones.
The reason is rooted in the diminishing “marginal utility of consumption.” In plain English: People value economic improvement (or feel economic pain) more when they have fewer resources to begin with. Relatedly, people with fewer resources are more likely to spend back into the economy a larger share of the economic gains they experience (in economic terms, a higher “marginal propensity to consume”). Under a typical “utilitarian” assumption—that policymakers should seek the greatest good for the most people—Dávila says a central banker in a HANK model “is going to have a desire to redistribute.”
In Dávila’s recent research, however, this desire is the central bank’s undoing. Dávila finds (with Andreas Schaab of the University of California, Berkeley) that as the central bank leans toward lower interest rates to help low-income, higher-debt households, all actors in the economy adjust their expectations of future inflation. In the long run, this undermines price stability.
“Even if we acknowledge that we want to help the poor, the worst thing that can happen in our model is to have this massive inflation,” Dávila said. “That’s what’s going to happen in the end.”
In their model, the best solution is for a central bank to avoid such short-run discretion by tying its hands in some fashion—such as by committing to an economy-wide target. “The current [Federal Reserve] mandates based on aggregates seem to have desirable properties, in the sense that they do not introduce other motives, these distributional motives, which can distort inflation,” Dávila said.
But Dávila cautions that it is early days for HANK models, and his insights could change. “This is really a topic that deserves much, much more exploration,” he said. “This is like a complete frontier area.”
Exploring different targets
Within the Federal Reserve System, Makoto Nakajima is among the HANK-fluent economists working most heavily along that frontier. After joining the Philadelphia Fed in 2008, Nakajima sharpened his modeling skills as he observed the growing interest in inclusive growth.
“I thought that one way I can contribute to the policy discussion at the Fed is to think about how monetary policy affects especially those less privileged, or people facing economic hardship,” said Nakajima, a member of the Institute’s System Affiliates Board and now head of the Philadelphia Fed’s monetary and macroeconomics section. “They might be affected differently than the average household, which is what traditional macroeconomic and monetary theory mainly captures.”
Nakajima has written on how the effects of monetary policy differ by age, income, wealth, and homeownership. And in a 2023 Institute working paper, he uses HANK modeling to explore the outcomes if monetary policymakers targeted the Black unemployment rate in the U.S. rather than the aggregate. “The Black unemployment rate tends to be twice as high as the White unemployment rate, and that ratio kind of stays the same over the business cycle,” said Nakajima. Black-led households face higher risk of job loss and are more likely to live hand-to-mouth. “When they lose a job, they are less likely to be able to rely on their bank account or savings to temporarily use to support expenditures to sustain their standard of living.”
The idea of focusing on the Black unemployment rate was a passionate topic for William Spriggs, the Howard University and AFL-CIO economist and advisor to the Institute who died in 2023. Spriggs emphasized that Black employment recovers more slowly in expansions and suffers more in contractions. In this view, monetary policy reacting to median or aggregate data could consistently leave Black Americans on the wrong side of the cycle.
In Nakajima’s research, he finds that if monetary policymakers wanted to prioritize labor market outcomes of Black workers, those policymakers would respond 80 percent more strongly to changes in the average unemployment rate than what they have been doing. Reducing the Black-White employment gap comes at the potential price of modestly higher inflation, but in Nakajima’s model the trade-off can be worthwhile if the goal is to smooth out fluctuations of Black workers’ earnings.
Recent empirical research by economists at the New York Fed explores the flip side: How do rate increases by the Fed play out in financially distressed communities? Looking first at consumer credit, economists Rajashri Chakrabarti and Maxim Pinkovskiy find that rate hikes bite more severely in U.S. counties with higher debt-to-income ratios. Increases in the federal funds rate since the year 2000 are followed in these places by lower auto loan and mortgage originations, higher credit card balances, and higher shares of delinquency, bankruptcy, and foreclosure.
The economists further find that heavily Black counties experienced even higher rates of these adverse consequences. Black communities especially experienced statistically significant differences in credit delinquencies and lower credit scores for years after the Fed raised its policy rate. In the labor market, the economists find Black, Hispanic, and female workers experienced greater and more persistent declines in employment in the months following an interest rate increase.
These credit and labor market impacts are not purely a result of Fed actions, of course. They are, said Chakrabarti at a recent Fed conference, “an interaction with pre-existing frictions [and] inequities.”
Beware unintended consequences
What if monetary action helps struggling households, but only at the cost of wider inequality? This is the cautionary implication of research from Alina Bartscher, Moritz Kuhn, Moritz Schularick, and Paul Wachtel, published as an Institute working paper and later in final form by Brookings. The economists undertake a statistical analysis to isolate the effects of easing monetary policy (lowering interest rates) on households headed by individuals of different races.
They find that lower rates lead to a positive, although modest, increase in employment outcomes for Black-led households. However, they also find a much larger gain for White-led households through a different channel, as looser policy boosts the value of housing and financial assets (which White households hold more of). While the average Black household sees a $3,300 increase in wealth five years after a 1 percentage point decrease in interest rates, the average White household gains almost $19,000. The conundrum, they write: “Monetary accommodation widens racial wealth inequality as it reduces income inequality.”
This finding invites questions. Duke economist and Institute advisor William “Sandy” Darity Jr., in a formal response to the paper, pointed out that while the researchers study incrementally lowering interest rates in a calm economic environment, the effect might be far different under other circumstances and from starker Fed actions. “When you look at restrictive monetary policy, particularly when you have massive increases in interest rates produced by Fed policy, then I think you are really hard-pressed to say that the employment effects are minor in comparison with the asset effects,” Darity reiterated in an interview for this article.
Another point: The change in value of a stock portfolio cannot necessarily be compared dollar-for-dollar with the effect of gaining, keeping, or losing a job for a low-income family. “If lower-income households lose, let’s say, 10 percent of their income, that’s relatively small from the macroeconomic perspective, because their income is lower from the beginning,” said Nakajima of the Philadelphia Fed. “But if you don’t have any savings, losing 10 percent of your income could have a very significant effect. … The pain that this household is suffering is, I think, in a sense much bigger than a dollar amount that you can easily compute.”
Other research cautions against reducing the effects of monetary policy to a binary trade-off between poor and rich households. Research economists from the central bank of Sweden, for example, propose that effects of monetary policy are “U-shaped”: Monetary easing boosts outcomes for those at the extremes—albeit through different forces—while doing less for households in the middle. Recent research from New York Fed economists finds a similar pattern for monetary contraction, with the hardest effects felt on the extremes.
Evolving models, evolving debate
It is possible that state-of-the-art HANK models lead right back to the same actions policymakers would take without them. “Roughly speaking, low-income households benefit from a tighter labor market, middle-class households benefit from lower mortgage rates, and wealthy households benefit from capital gains on assets” when monetary policy is loosened, write Minneapolis Fed Senior Research Economist Alisdair McKay and MIT economist Christian Wolf in the Journal of Economic Perspectives. They review recent findings and conduct their own analysis of consumption responses to monetary easing. Although the channels of impact are different, they find that the net change in consumption is similar across households of all wealth levels.
Even for a central bank that takes account of income and wealth redistribution, “appropriate policy is unlikely to differ too much from the optimal policy of a central bank that is solely focused on macroeconomic goals,” they write. They estimate the scale of monetary policy actions to meaningfully reduce inequality would need to be extremely large. “Such large changes would likely be costly in terms of other policy goals (notably aggregate output and inflation stabilization).”
Economists Greg Kaplan, Benjamin Moll, and Giovanni Violante are credited with coining the HANK acronym in a 2018 paper. Not surprisingly, they take an ongoing interest in how heterogeneous models might alter the practice of central banking. “Studies of optimal monetary and fiscal policy in HANK models agree that the benefits of aggregate stabilization are dwarfed by the gains from directly alleviating hardship,” they wrote recently for the International Monetary Fund.
“Optimal policies in HANK models almost always favor redistributing toward hand-to-mouth households in downturns,” they continued. “One may be tempted to read this as endorsement of using monetary policy to share prosperity and mitigate adversities. But monetary policy is a blunt tool for redistribution or insurance. HANK models tell us that fiscal policy is likely better suited for this task because it can be targeted more precisely to those in need of support.”
For others, this is a less-than-satisfying place to land. The Fed’s mandate “was set at a time when we didn’t really realize all these collateral effects of monetary policy—that it has all these redistributive effects,” said Stanford economist Adrien Auclert, a HANK modeler whose research has explained the channels through which monetary actions play out across varied households.
Monetary policy needs to act “with the understanding of all these effects that monetary policy has, and all these heterogeneous impacts,” said Auclert, a former Institute visiting scholar. “You cannot claim that those things don’t exist—or you can no longer claim that. Those effects are there. You can have a debate over the extent to which it matters for the aggregate effects of policy.”
Auclert’s perception is that HANK models are just now reaching a stage where policymakers can use them routinely. “It’s amazing just how much those models have been making progress,” Auclert said. “I think over time as the modelers refine the tools, they could become of more direct use in policy decisions.”
Deeper data, challenging choices
With the required five-year lag, we can see the official “Tealbook” briefings that Fed policymakers receive before each meeting of the Federal Open Market Committee (FOMC). The latest available, from December 2018, includes more than 200 figures and tables. Two of them, at the bottom of page 33, show unemployment and labor force participation for Asian, Black, Hispanic, and White workers. Across 168 pages, they were the only evident nod to “heterogeneity.”
Today’s Tealbook briefings will become public in 2029. But like the wider world over the past five years, it seems likely the briefings and the conversation around the FOMC table have changed. For example, the Fed’s latest biannual Monetary Policy Report to Congress now dissects the labor market by race and ethnicity, gender, education, age, and income.
Back in St. Paul, Keilan Crawford is focused on keeping what he’s achieved, affording daily life, and getting ahead. Our coffee break is a short pause from his packed routine: Long, physical days of building maintenance followed by night school, working toward an HVAC degree. When we met, Crawford had just applied for a full-time building engineer position, with a potential raise. It might be enough, he thinks, to consider a cheap used car.
The magazine of the opportunity & Inclusive Growth institute
Since monetary policymakers have been thinking more about workers like him, I invited him to think for a moment about them. How might he weigh things from the policy seat? Maintaining higher interest rates can help keep prices in check, but maybe slow the job market. Or do we drop interest rates—putting a car loan in reach but risking higher inflation?
He thinks hard about what he’d prefer. Ultimately, Crawford laughs and shakes his head. “Honestly, like, the working man is going to suffer no matter which way you go with that.”
Inside and outside of the Fed, the conversation is underway—and the tools are coming into their own—to see if policymakers can find a more satisfying answer.
This article is featured in the fall 2024 issue of For All, the magazine of the Opportunity & Inclusive Growth Institute
Jeff Horwich is the senior economics writer for the Minneapolis Fed. He has been an economic journalist with public radio, commissioned examiner for the Consumer Financial Protection Bureau, and director of policy and communications for the Minneapolis Public Housing Authority. He received his master’s degree in applied economics from the University of Minnesota.
