Americans, Chinese, and Europeans have curtailed their lending and investing abroad—while also selling commensurately fewer financial claims to foreigners. In 2021, the gross value of cross-border financial transactions involving the U.S., China, and the euro area was worth about $7.9 trillion. In 2022, that figure was just $2.8 trillion. U.S. data for the first three months of this year suggest that cross-border transactions volumes have continued to shrink. International financial transactions involving the world’s three largest economies are smaller relative to their combined output than at any point other than the trough of the financial crisis.
All told, institutions that report to China’s central government probably have closer to $6 trillion in foreign assets than the $3.12 trillion SAFE reported in December 2022. The scale of these hidden reserves — foreign current currency assets that aren’t formally counted as “reserves” — also highlights an important fact that is often forgotten amid all the talk of China’s domestic debt problems. The main way China has hid its reserves has been its big state banking system. Globally China is still a massive creditor, and the weight of China’s massive accumulation of foreign exchange is still felt around the globe.
A review of the use of the dollar globally over the last two decades suggests a dominant and relatively stable role. To illustrate this stability, we construct an aggregate index of international currency usage. This index is computed as the weighted average of five measures of currency usage for which time series data are available: Official currency reserves, FX transaction volume, foreign currency debt instruments outstanding, cross-border deposits, and cross-border loans. The dollar index level has remained stable at a value of about 70 since 2010, well ahead of all other currencies. The euro has the next-highest value at about 23, and its value has remained fairly stable as well. While international usage of the Chinese renminbi has increased over the past 20 years, it has only reached an index level of about 3, remaining even behind the Japanese yen and British pound, which are at about 7 and 6, respectively.
Figure 3.1 shows real GDP for China, Japan, South Korea, and Taiwan, in which each economy begins at approximately the same per capita income level. The most striking aspect of these data is that the time path of China’s per capita GDP is remarkably similar to those of the three other East Asian miracle economies. After 25 years of becoming middle-income economies, China, Japan, and Taiwan achieved nearly the same per capita real GDP level. The one exception is Korea, which was somewhat ahead of the other three at that point in time. These data show that there is nothing unusual about China’s real GDP growth rate relative to the experiences of these other economies. We use the model and the fitted TFP catch-up process to predict that China’s relative per capita GDP level will asymptote to about 41% of the U.S. level around 2050, reflecting a substantial slowdown in China’s observed TFP catch-up in recent years. Related: How Soon and At What Height Will China’s Economy Peak?
Artificial intelligence is a familiar-looking monster, say Henry Farrell and Cosma Shalizi [that] began at least two centuries ago with the industrial revolution, when human society was transformed by vast inhuman forces. Markets and bureaucracies seem familiar, but they are actually enormous, impersonal distributed systems of information-processing [with minds of their own]. Economist Friedrich Hayek argued, any complex economy has to somehow make use of a terrifyingly large body of disorganised and informal “tacit knowledge” … [that] no individual brain or government can possibly comprehend. These vast machineries are simply incapable of caring if they crush the powerless or devour the virtuous. The modern world has been built by and within monsters, which crush individuals without remorse or hesitation. We eke out freedom by setting one against another, deploying bureaucracy to limit market excesses, democracy to hold bureaucrats accountable, and markets and bureaucracies to limit democracy’s monstrous tendencies.
We find that it has become easier for a male earner to support a family than it was in 1985. We correct the dollar costs of the Oren Cass' Cost of Thriving Index (COTI) and find that instead of rising by 22 weeks of work, the cost of thriving (for male workers) rose by just over 10 weeks. This improved estimate overstates the increase in COTI, however. Cass’s way of measuring “costs” fails to account for quality improvement in the items he tracks—particularly better health care and nicer cars. Furthermore, COTI does not take into account taxes—a major flaw. It’s especially important in this case because the kind of family Cass is describing has seen a major reduction in its federal tax burden since 1985, going from a net taxpayer to receiving a subsidy, primarily due to the child tax credit (CTC). After including estimates of the federal tax burden or tax subsidy in both years, we find that the 2022 cost of thriving for a family with one male earner is lower than it was in 1985.
Although the High Tech sector still remains more dynamic than the non-High Tech sector of the economy, the gap between the two has gradually faded. Beginning with firm entry rate, we see a long-running decline in both the High Tech sector and the broader economy. The national firm entry rate, hitting a series high of 13.5 percent in 1984, fell below 10% during the Great Recession and has yet to recover. The High Tech sector saw much higher startup rates during the 1980s, peaking at 17.9%, also in 1984. High Tech industries saw another huge surge in new startups in the 1990s, coinciding with the dot-com boom, but their startup rates have yet to recover. In 2018, the gap between the national and High Tech startup rates fell below one percentage point to an all-time low, down from 5 to 6percentage points in the 1990s. The rate of new firm creation in the High Tech sector, once far ahead of other sectors, now looks similar to that of the increasingly sclerotic national economy.
Shortly after graduating, workers with college degrees obtain jobs in professional, non-routine occupations with much greater scope for on-the-job learning and wage growth. Less educated workers hold jobs in similar occupations before and after completing schooling. Occupational sorting causes relative wages to grow especially rapidly for college graduates in the initial phase of their careers. However, after a few years, educated workers are less likely to switch jobs, and nearly 90% of life cycle wage growth occurs within rather than between jobs. Wages increase with job tenure much more in professional, nonroutine occupations, which are disproportionately held by college graduates. I estimate that occupational sorting explains at least half of the life cycle growth in the college wage premium. The college wage premium increases from 22% at age 26 to 49% at age 40 and 56% at age 60. Related: Why Do Wages Grow Faster For Educated Workers?
You will note how this relates to the signaling vs. human capital debates over education. Signaling your quality may put you in a position to learn more over time, as your initial offer likely will be better if you come out of Harvard. But over the longer haul, the wages you earn are the result of what you have learned, not just your initial level of quality. So, most of the return to education is that you learn more over time, and thus most of the return is learning-related rather than initial quality-related. Overall, signaling models behave rather awkwardly in dynamic rather than purely static settings. Related: Why Do Wages Grow Faster For Educated Workers?
The National Center for Education Statistics (NCES) administered the NAEP long-term trend (LTT) reading and mathematics assessments to 13-year-old students from October to December of the 2022–23 school year. The average scores for 13-year-olds declined 4 points in reading and 9 points in mathematics compared to the previous assessment administered during the 2019–20 school year. Compared to a decade ago, the average scores declined 7 points in reading and 14 points in mathematics. The 2023 mathematics scores for age 13 students at all five selected percentile levels declined compared to 2020. The declines ranged from 6 to 8 points for middle- and higher-performing students to 12 to 14 points for lower-performing students, with larger declines for lower performers in comparison to their higher-performing peers. The percentage of students who reported missing 5 or more days doubled from 5 percent in 2020 to 10 percent in 2023.
In 2021 Queen Creek, [a high-end suburb of Phoenix] paid $30 million for 5,000 acre-feet of water [from] Harquahala, a basin west of Phoenix. It’s negotiating for another 8,000 acre-feet there, though the cost will likely be much higher this time around: Buckeye, [another Phoenix suburb,] approved paying $80 million for 6,000 acre feet from the same area in January. That’s more than double what Queen Creek paid in 2021. In 1993 the state created the Central Arizona Groundwater Management Replenishment District (CAGRD). Communities would pay to pump water and in return the agency would replenish what they took out with water from the Colorado or other rivers. Households in the district pay bills for water and separate assessments for groundwater replenishment. Those rates have already gone from $154 per acre-foot in 2002 to $768 per acre-foot in 2021.
Since SVB collapsed, the Fed has been adding liquidity, and the S&P500 is up more than 10%. The high correlation between Fed net QE and the S&P500 seen in the chart below suggests that Fed liquidity is a crucial driver of the stock market. With the Fed turning more hawkish and continuing QT, the downside risks to equities are growing.
Looking at this at a more granular level, we expect the relative share of the US in global equity market cap to decline from 42% in 2022 to 35% in 2030, 27% in 2050, and 22% in 2075. By contrast, we expect China’s share in global market cap to rise from 10% to 15% by 2050, but decline to around 13.5% by 2075. We see the biggest scope for global market cap share increase in India, where our projections imply that India’s global equity market cap share will rise from around 2.5-3% in 2022 to 8% in 2050, and 12% in 2075. Likewise, we project the rest of EMs' share to rise, from 13.5% in 2022 to 17% in 2030, to 24% in 2050, and to 30% in 2075.
Investors are paying a sizable premium for AI exposure. One example: the 32 stocks in the $CHAT Generative AI ETF have returned 77% this year. The 22 $CHAT ETF stocks that have positive earnings projected for the next 12 months trade at an average P/E of 31x. The other 10 $CHAT ETF stocks have infinite P/Es, either due to negative earnings or trivially positive earnings that create nonsense P/E ratios over 100x. The large language model revolution is going to have to live up to the high end of expectations to justify this. I have seen estimates that AI will eventually add ~10% to the fair value of the S&P 500 due to some combination of augmented employee productivity and staff reductions, but I’m not sure how one would measure that.
We document that changes in the perceived cost of capital only modestly affect discount rates, in contrast to the stylized view. Using within-firm variation, we show that, on average, a 1 percentage point increase in the perceived cost of capital leads to a 0.3 percentage point increase in the discount rate. We show that discount rate wedges are associated with investment fluctuations at the firm level. A 1 percentage point increase in the wedge lowers the investment rate by 0.9 points. Many firms rarely change discount rates, so the relation becomes stronger over longer horizons. However, even at the 10-year horizon, 40 percent of firms maintain unchanged discount rates and, even if they change, adjust less than one-to-one with the perceived cost of capital. These results suggest that discount rates have “a life of their own,” beyond the perceived cost of capital. The average US firm in our sample has increased its discount rate wedge by 2.5 percentage points between 2002 and 2021.
Put another way, the increase in nominal pretax business income since 2019Q4 is equivalent to just 17% of the increase in consumer spending excluding imputed rent since the eve of the pandemic. For perspective, the total amount of nominal pretax business income generated in the U.S. in any given quarter has historically been worth about 21% of PCE ex-imputed rents, which means that the growth in profits has been a little low relative to what one might have naively expected.
Core inflation (which excludes food and energy) rose at about a 5% annual rate for a sixth straight month. But other measures of underlying inflation were much lower reflecting the outsized importance of shelter and used cars (again) this month. Some relief should be coming. Look at the BLS's most narrow measure, sometimes called "supercore," which also excludes shelter (which is lagging) and used cars (which are volatile and posted a big increase in May). At an annual rate: 1 month: 1.1% 3 months: 2.3% 6 months: 3.3% 12 months: 4.2%. What does all of this mean? The most standard measure of core inflation is running at a 5% rate. That almost certainly overstates current underlying inflation. But by how much? How much will it fall? I expect some but have no confidence it is falling to 3% let alone 2%. The Fed has strongly telegraphed a pause/skip & nothing about this will change that. But is mildly unfortunate. Over the last 3 months core CPI is 5.0% annual rate. The employment mandate is fine with 283K jobs per month. Both much hotter than expected. Ought to raise now.
Next April, for the first time in more than three centuries, Danes will have to work on the holiday of Great Prayer after the government scrapped the religious day off partly to pay for extra defence spending. The decision, approved in March, was deeply unpopular: in one poll, 70 percent of Danes opposed it. The Stockholm International Peace Research Institute calculates that global defence spending rose by 4 percent to reach a record $2.24tn last year. This year, it is set to continue rising, even as higher rates increase governments’ borrowing costs. During the long decades of the cold war, the west financed its defence spending through higher taxation.
There is no appetite in the U.S. for boosting investment at the expense of current consumption by redirecting spending, workers, and capital. Nor is there much scope for American producers to export even less than they already do. This means that any realistic scenario in which U.S. investment rises substantially would mostly involve some combination of higher production out of existing labor and capital as well as higher levels of imports. In theory, good investments should boost the capital stock and eventually make it easier to further increase both investment and consumption, but that process takes time. Until then, there will be a mismatch between America’s needs and its productive capabilities that can only be bridged by drawing upon foreign production. Foreigners stand to benefit from this even if some of the specific subsidies encouraging U.S. investment have local-content restrictions.
An additional $1 spent auditing taxpayers above the 90th income percentile yields more than $12 in revenue, while audits of below-median income taxpayers yield $5. We begin by estimating the average initial return to all audits of US taxpayers filing in 2010-2014. On average, $1 in audit spending initially raises $2.17 in revenue. Audits of high-income taxpayers are more costly, but the additional revenue raised more than offsets the costs. Audits of the 99-99.9th percentile have a 3.2:1 initial return; audits of the top 0.1% return 6.3:1. Next, we use randomly selected audits to examine the impact of an initial audit on future revenue. This specific deterrence effect produces at least three times more revenue than the initial audit. Deterrence effects are relatively consistent across the income distribution. This results in the 12:1 return above the 90th percentile.
Middle-income earners, not the poorest, appear to have borne the brunt of America’s current bout of inflation. Research by Xavier Jaravel of the London School of Economics has yielded a distinctive shape—an inverted-U curve—that illustrates the distributional effects of inflation in America from mid-2020 to mid-2022. For the lowest earners, inflation was about 13.5%. For those in the middle of the spectrum, inflation was closer to 15.5%. The wealthiest faced inflation of about 14%. Since 2020 nominal wages have increased at an average annual rate of about 4.2% for Americans as a whole. The lowest-earning quartile, however, has seen the biggest gains, with their wages up by 5.3% on average during the same period.
Value-added per employee is a measure of labor productivity. In America’s tradable sector, it has risen steadily over the last two decades in both manufacturing and services, reaching roughly $185,000 (in chained 2012 dollars) in 2021. Over the same period, productivity growth in this sector averaged nearly 3%. The non-tradable economy [net government] is just 0.57% per annum over the last 20 years. This reflects below-average productivity levels and, in most cases, low-to-moderate productivity growth in the large-employment sectors. There was not always a large gap between the tradable and non-tradable sectors. On the contrary, as the chart shows, labor productivity was about $100,000 across the economy in 1998. But by 2021, after more than two decades of steady divergence, per-employee value-added in the tradable sector was nearly double the level in the non-tradable sector.
The percentage of single Americans who are looking for a relationship or casual dates is lower than in 2019, especially among men. Some 42% of single Americans say they’re looking for a committed romantic relationship and/or casual dates, down from 49% in 2019 (but similar to the proportion found in the Center’s February 2022 survey). This drop is largely driven by single men, who are now 11 percentage points less likely than in 2019 to say they are looking for a committed relationship and/or casual dates (50% in July 2022, down from 61% in 2019). During the same time frame, there has been no significant change in the share of single women who are looking for a relationship or casual dates: 35% said this in 2022, compared with 38% in 2019.
The share of U.S. adult children living with their parents has increased since the 1960s. Figure 1 shows that in 2020, approximately one-third of children between ages 18 and 34 lived with their parents, with men and 18-24 year-olds, respectively, more likely to co-reside than women and 25-34 year-olds. We examine the relationship between adult children returning home and parental retirement outcomes using data from the Health and Retirement Study (HRS), a nationally representative panel of individuals over the age of 50 and their spouses; the HRS also tracks children of respondents. Our child-level analysis suggests that a boomerang event is likely associated with negative shocks to a child’s marriage, income, and employment. The event study analysis suggests that many of these shocks are temporary, and correspondingly most boomerang events are transitory. At the parent level, we find no clear, statistically significant association between boomerang children and parental health, wealth, probability of working, hours worked, or well-being. However, we do find an increase in the self-reported probability of working full-time after age 65. That increase is concentrated among men, those under the age of 62, and those in the top half of the initial wealth distribution. Overall, our results provide evidence that parents may delay their anticipated retirement when children return home.
Deficit-financed fiscal transfers generate excess savings. One person’s spending is another person’s income. As we show, taking this fact into account implies that excess savings from debt-financed transfers have much longer-lasting effects than a naive calculation would suggest. In a closed economy, unless the government pays down the debt used to finance the transfers, excess savings do not go away as households spend them down. Instead, the effect of excess savings on aggregate demand slowly dissipates as they “trickle up” the wealth distribution to agents with lower MPCs. Tight monetary policy speeds up this process, but this effect is likely to be quantitatively modest. The partial equilibrium scenario summarizes the conventional wisdom according to which the effect of excess savings will dissipate in a few quarters. By contrast, our benchmark scenario suggests that these effects will stick around for roughly 5 years.
Why have US aggregate profit rates increased while financial market rates decreased since 1980? We propose a mismatch hypothesis: Profit rates in the national accounts track the return on capital for all firms, while financial market rates track the cost of capital for public firms only. We show public-firm profit rates halved since 1980, matching trends in financial markets and suggesting low market power. Mechanically, this residual private capital return series shows private capital returns had to have increased substantially to account for this secular break between aggregate and public firm profitability. The degree of this shift is significant: Private firms’ profit rates are on average 10% higher than public firms’ in the post-2000 period. Nonfinancial domestic private-firm profit rates doubled, suggesting high market power or risk. Size and sector differences cannot explain the divergence, though intangible-intensity might. Our results indicate substantial biases in extrapolating public-firm trends to the aggregate economy.
Almost all recent breakthroughs in artificial intelligence globally have come from large companies, in large part because they have the computing power. Amazon, whose AI powers its Alexa voice assistant, and Meta, which made waves recently when one of its models beat human players at “Diplomacy,” a strategy board game, respectively produce two-thirds and four-fifths as much AI research as Stanford University, a bastion of computer-science eggheads. Alphabet and Microsoft churn out considerably more, and that is not including DeepMind, Google Research’s sister lab which the parent company acquired in 2014, and the Microsoft-affiliated OpenAI.
Moving to the USA raises citations to math publications four-fold, or 2.5-fold if you restrict attention to people who become a math academic at home or abroad. Comparing New Zealander migrants who return or stay abroad - the ones who stay abroad get up to four times as many citations as those who return. PhD students who stay in the USA get 4-6 times as many citations to their work as their peers in the same program who end up moving back to a country with GDP per capita outside the top 25%. That’s a pretty consistently large effect. And we get similar kinds of results when we look at other proxies for scientific achievement, whether it’s counting publications, patents, or becoming an invited speaker to the International Congress of Mathematicians. A model of the innovation/immigration/trade economy between the US and EU...implies if the USA doubled the H-1B visa cap from 65,000 to 130,000, it would raise the real GDP per capita growth rate in each region by 9% in the long run.
Overall GDP is basically at CBO's pre-pandemic forecast. The big story for economic growth was the consistent strength of consumer spending, which is more than two-thirds of the economy. It has been consistently running well ahead of CBO's pre-pandemic forecast. Likely continued impact of huge fiscal support in 2020 and 2021. The countervailing story is the collapse of residential investment. It has fallen for 7 straight quarters, fell 19.3% over the last four quarters, with a decline of 26.7% (annual rate) in Q4. Biggest since the financial crisis. And the third story is that over the pandemic recovery the huge increase in US demand was partly accommodated by rising imports and production shifted from exports to domestic consumption. But the large increase in the trade deficit has been narrowing.
Summers has argued that the increase in public debt due to the fiscal response to COVID will lead, other things equal, to an increase in r [real return on capital]. He is right about the sign of the increase in public debt’s impact on r, but the effect is likely to be quite small. The debt-to-GDP ratio in advanced economies has increased only from 75 percent in 2019 to 82 percent in 2022; under standard assumptions, this implies an increase in r of no more than 15–30 basis points. That would be insufficient to offset the pre-COVID downward trend in safe rates, let alone to close the gap between r and g [growth rate of output.]
Starting in 2019, tighter immigration policies under the Trump Administration followed by pandemic disruptions depressed the US foreign-born working-age population. The foreign-born labor force fell 2.8 million below its long-term trend level in April 2020 and remained 1.2 million below trend in March 2022. The foreign-born population has grown 137k per month in the past 18 months, compared to 68k per month from Jan. 2010 to Jan. 2019. The chart above shows the foreign-born labor force has grown 110k per month in the past 18 months, compared to 42k per month during the prior period.
An ageing population, and the dependency it creates, will hamper supply and stoke inflation. Mikael Juselius and Elod Takats of BIS’s core insight: “The young and the old are inflationary, while the working-age cohort is disinflationary.” That is, prime-age workers create more supply than demand, while their elders and juniors do just the opposite.
The above chart plots the history of US tax increases since 1950 (as % of GDP and vs Federal receipts as a % of GDP). While there have been tax increases of 2% of GDP or more, they occurred when overall tax receipts were much lower. The red square shows the required increases in taxes, which if spent entirely on increasing discretionary spending, would reduce the ratio of entitlements to non-defense discretionary spending back to 2.2x (its 2006 level). The Sanders high net worth income and capital gains tax plan and the Warren wealth tax plan appear as well.
Median weekly earnings for all workers were 7.4% higher, year-over-year, at the end of 2022, according to an analysis of newly released Labor Department data. The bottom 10th of wage earners—those that make about $570 a week—saw their pay increase by nearly 10%. Better pay increases late last year went to workers who attended college, a reversal from earlier in the pandemic when those who hadn’t completed high school saw outsize gains. The annual rate of wage growth for workers with less than a high school diploma touched a recent peak in the second quarter of 2022, when it was up 11.1% over the prior year, higher than the 7.6% wage growth during that period for workers with a bachelor’s degree or higher. The median raise for Black Americans employed full-time was 11.3%, compared with the prior year.
The labor-force participation (LFP) rate of prime-age workers (aged 25-54) and the foreign-born workforce have almost fully recovered. Neither explains the current squeeze. The biggest shortfall comes from Americans getting older and leaving work behind. Since 2019 those aged at least 65 have gone from less than 16% of the population to nearly 17%. Moreover, unlike prime-age workers, many people who retired early as covid-19 struck have not come back to work. LFP among older Americans, which rose from 12.5% in 2000 to 20.7% in early 2020, has dipped to 19.3%, the same as in 2016. The aging of the population accounts for the loss of 1.9m workers (0.7% of people aged at least 16), while the overall drop in LFP, mainly among the old, is responsible for a further 0.5m (0.2%).
Over two-thirds of the structural fiscal imbalance derives from the unsustainable growth rates of federal health programs, most especially Medicare and Medicaid. Irrespective of future policy decisions in other areas such as tax policy, income security, and annually appropriated domestic and defense spending, federal finances will not be stabilized until Medicare and Medicaid’s growth rates are moderated. A survey of fiscal stewardship records produces the unsurprising result that more recent officeholders have tended to run far higher federal deficits than those countenanced by previously elected officials. The largest average federal deficits were operated during the Trump administration, followed, in turn, by the Obama, Ronald W. Reagan, and George H. W. Bush administrations.
Fiscal conservatives should aim to permanently stabilize the debt at 95 percent of GDP. This goal would mean keeping deficits near 3 percent of GDP, compared to the baseline deficits rising past 6 percent of GDP over the next decade and 11 percent of GDP in three decades. In the short run, this means: Freezing annual discretionary appropriations. Building momentum for mandatory spending reforms with a modest package of savings (perhaps $400 billion over the decade) that address lower-hanging fruit such as leftover pandemic spending, program overpayments, and federal spending benefits for upper-income families. Begin working toward Social Security and Medicare reform—which drive nearly 100 percent of long-term deficits—by building bipartisan working groups behind the scenes.
In 2016, Ron Johnson rode Trump’s coattails and the Republican trail blazed by the former governor Scott Walker to a 3.4 point (50.2 to 46.8) victory and swept into office, in large part by running up huge margins in Milwaukee’s predominantly white suburbs. That changed in 2022. Craig Gilbert, a fellow at Marquette Law School conducted a detailed analysis of Wisconsin voting patterns and found that Johnson, "performed much worse in the red and blue suburbs of Milwaukee than he did six years earlier in 2016. So again, how did Johnson win? The simple answer: white rural Wisconsin. As recently as 17 years ago, rural Wisconsin was a battleground. In 2006, Jim Doyle, the Democratic candidate for governor, won rural Wisconsin, about 30 percent of the electorate, by 5.5 points. “Then came the rural red wave,” Gilbert writes. “Walker carried Wisconsin’s towns by 23 points in 2010 and by 25 points in 2014.” In 2016, Johnson won the rural vote by 25 points, but in 2022, he pushed his margin there to 29 points.
President Trump’s record on fiscal responsibility does not compare favorably to his immediate predecessors. Surely, it would not be fair to judge President Trump simply by the total budget deficits under his watch, however, as the $10 trillion 10-year baseline deficit that he inherited dwarfed the $4 trillion projected baseline deficit inherited by President Obama and the $6 trillion projected baseline surplus inherited by President Bush. That is far from a level playing field. On the other hand, President Trump also received the largest automatic deficit reductions from his inherited baseline. During his presidency, economic and technical factors that fall mostly outside of political control produced $3.9 trillion in 10-year deficit reduction, mainly through falling interest rates on the federal debt.
The US is missing about a fifth of its pre-pandemic low-income workforce. At least some of those workers moved to higher-paying jobs, but, after adjusting for wage growth, researchers found employment for the poorest quarter of the workforce was still 13.5% below pre-pandemic levels at the end of 2021. Analyzing local trends, researchers found an important clue to where those missing workers went: Low-wage workers are scarcest where 2020’s devastation was worst. “It is clear there are large swaths of the population who are still not employed, and these are low-wage workers who lost their jobs in precisely the places where high-income people cut back on spending so sharply a couple of years ago,” Chetty said.