Macro Roundup

Over the past 30 years the age of peak earnings has been almost constant in the US (40-50) but had a major decline in the PRC from ~ 55 to 35. China has seen lower returns to human capital then the US. Three core findings, “….In this paper, we first document stark differences in the cross-sectional age-earnings profiles between the U.S. and China, the two largest economies in the world, during the past thirty years. We find that, first, the peak age in cross-sectional age-earnings profiles, which we refer to as the “golden age,” stayed almost constant at around45-50years old in the U.S., but decreased sharply from55to around35years old in China; second,the age-specific real earnings grew drastically in China, but stayed almost stagnant in the U.S.; andthird, the cross-sectional and life-cycle age-earnings profiles looked remarkably similar in the U.S., but differed substantially in China….”

Hanming Fang and Xincheng Qiu, "“Golden Ages”: A Tale of the Labor Markets in China and the United States," National Bureau Of Economic Research, November 2021, https://www.nber.org/papers/w29523

Korea’s path“…Is this a realistic prediction? Only history will tell for sure, but interestingly, Figure11shows that such a pattern of increasing “golden ages” actually happened in Korea during the past ten years, using data from the Korean Labor and Income Panel Study (KLIPS). Korea experienced its fastest growth during the1960s to1990s. After that, it began to slowdown. Appendix Figure A.6depicts the decomposition for Korea, together with the decomposition for U.S. and China. It is worth noting that the cohort effects are particularly large from cohort1945to cohort1960, but starts to decelerate afterwards. This is consistent with our explanation of the race between inter-cohort productivity growth and returns to experience. As inter-cohort productivity growth starts to give its way to experience in Korea, the “golden age” comes back to older ages, as in our hypothetical scenario in Figure10…”

Potential implications, “…We also use the inferred components, particularly the series of the quantities of human capital, to revisit several important and classical applications in macroeconomics and labor economics, including the growth accounting and the estimation of the TFP growth, and the college wage premium and the skill-biased technical change. We find that once we adjust for the changes in the quantities of human capital, the estimated contribution of the TFP to GDP per capita growth is smaller than the previous estimates in the literature. We also find that the skill-biased technical change played an important role in the rising college premium to ensure that the relative price of college human capital does not drop as much as it would otherwise do when there is a large increase in the quantity of college human capital. A simple simulation exercise using our framework also suggests that, as the Chinese economy slows down to a “new normal” growth rate similar to that in the U.S., the golden ages of the cross-sectional age-earnings profile in China will start to increase to older ages, similar to what has happened in Korea in the last ten years….”

Theory, “…To explain these striking differences, we propose and empirically implement a unified decomposition framework to infer from the repeated cross-sectional earnings data the life-cycle human capital accumulation (the experience effect), the inter-cohort productivity growth (the cohort effect), and the human capital price changes over time (the time effect), under an identifying assumption that the growth of the experience effect stops at the end of one’s working career. The decomposition suggests that China has experienced a much larger inter-cohort productivity growth and higher increase in the rental price to human capital compared to the U.S.; but the return to experience is higher in the U.S….”

Datapoint suggesting underlying talent might be constrained, “…The contributions of relative labor supply, relative human capital per worker, and skill-biased technical change to the evolution of relative human capital prices are depicted in Figure9. Itshows clearly that in both U.S. and China, the relative quantity of college human capital grows rapidly, which would have led to sharp declines in the price of college human capital relative to non-college human capital. Due to skill-biased technical changes, the relative price of collegehuman capital did not decline even more in the U.S. and actually increased in China in the last thirty years….”

“…For example, college and high school graduates may possess different types of skills that are not perfect substitutes. To do so, we perform the decomposition as discussed in Section4separately for college workers and high school workers. College and high school workers are allowed to have different paths of life-cycle human capital growth, different inter-cohort human capital growth, and different time series of human capital price changes. The only restriction is that for both college workers and high school workers, there is no additional skill accumulation from experience in the last two experience bins towards the end of working life. Since our imputation of potential experience assumes that college graduates start to gain experience from22years old and high school graduates start to gain experience from18years old, effectively it is assumed that college graduates do not have additional returns to experience in52-61years old and high school graduates in48-57years old. This is largely overlapped with the “flat spot” proposed by Bowlus and Robinson (2012). After detailed investigation of the U.S. data, they conclude that a reasonable choice for the flat spot of the experience effect is around50-59for college graduates and46-55for high school graduates The results are presented in Figure7. First, within an education group, the returns to experience are still higher in the U.S. than in China. Within a country, the experience effects are larger for college workers than high school workers. This is consistent with findings documented by the previous literature that life-cycle wage growth tends to be faster for workers with more education (see Bagger et al., 2014, for example). The difference between the two education groups in their experience effect profiles, however, is much smaller compared to the difference in the cohort effects that we are turning to….”

Controlling For Education The Returns To Experience Are Still Higher In US Relative TO PRC

“…We present the contribution of each source — physical capital per worker, human capital per worker, and the residual — to the growth of GDP per worker in Figure5. We find that all three sources contribute almost equally to the U.S. growth, with the contribution of human capital slightly exceeding the other two sources. The picture is quite different in China. Although the absolute level of the growth in human capital is larger in China than in the U.S., the relative contribution of human capital turns out be the least important to China’s growth. But this is merely a result of an even faster speed at which the physical capital and TFP grow in China. In fact, physical capital is responsible for almost60% of the growth in GDP per worker, and TFP for almost another30% in China….”

Chinese GDP Per Worker Driven By Physical Investment, Not Human Capital Or TFP

“…Conceptually, across-sectionalage-earnings profile, which summarizes earnings of workers of different ages at a given point of time, is a different notion to thelife-cycleearnings profile, which tracks the earnings of a typical person over his life course. Thus one should not expect the crosssectional age-earnings profiles to coincide with the life-cycle ones. In Figure3, we reproduce the cross-sectional profiles from Figure1on the left, and plot the life-cycle earnings path of various birth cohorts on the right, with each curve representing a10-year cohort bin. The top panel is for the U.S., and the bottom panel for China. In the U.S. (Figure3a), cohorts with year of birth expanding half a century share remarkably similar life-cycle earnings paths. Furthermore, life-cycle profiles on the right of Figure3a closely resemble the cross-sectional profiles on the left (which is reproduced from the right panel of Figure1a), in both its shape and level. In a stationary environment where the life-cycle profile does not vary across cohorts, the cross-sectional profiles and the life-cycle profiles essentially coincide with each other. In such an economy, a30-year-old worker who wants to predict his (real) earnings10years later can simply take a look at the contemporary earnings of a40-year-old worker. This provides a justification for the voluminous prior literature that use cross-sectional profiles as approximations to life-cycle patterns. Although conceptually it is not correct to interpret crosssectional age-earnings profiles as life-cycle patterns, in practice they are close to each other for the U.S. case. In other words, stationarity is an reasonable assumption when studying the U.S. earnings profiles. However, as shown in Figure3b, the life-cycle patterns of different cohorts differ drasticallyfor China. More recent cohorts enjoy both much higher earnings and steeper life-cycle earnings growth. These life-cycle profiles also demonstrate no resemblance at all to the cross-sectional profiles, although they are actually linked to each other. Note that both the left and the right panels are just different ways to visualize the same underlying data.13It is perhaps not surprising that in a fast-growing economy such as China, stationarity is not a valid approximation…”

Cross Section versus Life Earnings Age-Earnings Profiles

“…To sum up, Figure2plots the evolution of the cross-sectional “golden ages” in the U.S. and China during1986-2012. For each country and each year, we run a kernel regression of log earnings on age to predict age-specific earnings, and obtain an estimated golden age in that year as the age achieving the maximal predicted earnings. Furthermore, we fit a linear time trend of the estimated golden age for each country. Figure2shows clearly that in the U.S., the golden age has stayed constant at around48years old in the past thirty years, while in China there exhibits a strong downward trend in the golden ages from1986to2012, decreasing from more than55years old to around35years old….”

PRC, “…In Figure1b, we plot the cross-sectional age-earnings profile for Chinese male workers, using the same procedure as discussed before. There are several striking contrasts between Figure1a and Figure1b. First, Chinese workers have experienced a dramatic increase in real earnings in the past30years for all age groups. It is reflected in the large vertical upward shifts of the ageearnings profiles for later cross sections. The earnings of Chinese urban male workers increased by nearly six folds. This is in marked contrast to the earnings stagnation in the U.S. Second, while the shape of the cross-sectional age-earnings profiles and hence the corresponding “golden ages” have stayed more or less constant in the U.S., the “golden age” in China is continuously evolving to younger ages. Prior to2000, the age-earnings profiles of China had a familiar hump-shape with the “golden age” at around55, although there already were some signs of a declining “golden age” in1996-2000. Between2001and2004, the age-earnings profile is almost flat and humps at around age40-45. After2005, the “golden age” is35years old…”

US, “…Figure1a depicts the cross-sectional age-earnings profiles for male workers in the U.S. Each curve represents a cross section that pools five or four adjacent years. In the construction of each curve, we first perform a nonparametric kernel regression of annual labor earnings on age separately for each cross section, where the Epanechnikov kernel function and rule-of-thumb bandwidth estimator are applied, and then display the smoothed values with the95% confidence intervals. To avoid potential impacts of extreme values, we drop outliers defined as earnings in the top2.5% and bottom2.5% in each year. We normalize all earnings to the2015year using CPI. Individuals are weighted by the person-level ASEC weight. Figure1a reveals that, first, the “golden age” in the U.S. is relatively stable at around50years old during the past three decades;second, the U.S. has witnessed little growth in age-specific mean real earnings. That is, both the shape and the level of the age-earnings profiles are largely unchanged…”

The Evidence

Scott Winship, "Economic mobility in America: A state-of-the-art primer," Archbridge Institute, December 1, 2021, https://www.aei.org/research-products/report/economic-mobility-in-america-a-state-of-the-art-primer/

"...The PSID and NLS yield conflicting results. In the PSID, the intergenerational elasticity rises over time, sometimes substantially, indicating falling mobility. In the NLS, the intergenerational elasticity is flat over time. For example, in the PSID, comparing parent family income to daughters’ earnings, the IGE rose from 0.25 for cohorts born 1952-59 to 0.43 for those born 1976-83. In the NLS, the elasticity was 0.26 for those born 1951-53, 0.23 for those born 1962-64, and 0.26 for those born 1981-83...."...Outpacing one’s father’s earnings has increased among daughters, but fewer daughters exceed their mothers’ earnings or, perhaps, their parents’ family income...."

Absolute Mobility Among Women

“…The red line in Figure 16, below, displays estimates of absolute family income mobility for the four sets of PSID cohorts. For cohorts born 1952-59, 47 percent of sons had higher earnings in their early 30s than their fathers had when the sons were between the ages of 8 and 21. The figure for sons born 1976-83 was 50 percent—no differentin the sense of being statistically significant. In contrast, the blue line in the chart comes from the Chetty et al. data and shows the trend for 1949-84 birth cohorts comparing the individual incomes of fathers and sons. (Individual income includes non-earnings income, so the trend is not strictly comparable to the PSID trend.) The four blue dots plot the eight-year averages for the same sets of birth cohorts that are combined in the PSID. The Chetty estimates indicate that absolute mobility fell dramatically over time. Across the cohorts covered in the PSID, the decline was from 61 percent to 42 percent…. If absolute mobility has, in fact, been flat in recent decades, that would suggest that two-thirds to 80 percent of today’s forty-somethings are better off than their parents were at the same age—much higher than the 50 percent suggested by Chetty et al. Similarly, Part One of this primer reported estimates from the PSID, using size-adjusted family income and pooling sons and daughters. The range of those estimates ran from 63 percent to 77 percent for recent cohorts of adults…”

Relative Mobility Among Women

“…Daughters’ relative mobility also likely declined, whether comparing father earnings, mother earnings, or parental income to daughters’ earnings or family income. Whether fathers’ or mothers’ earnings are the baseline, the decline in mobility was modest, except comparing mothers’ and daughters’ earnings when women without any earnings are included. The declines using parents’ family income as a baseline are modest in the NLS but sizable in the PSID. In the PSID the adulthood gap between the richest and poorest children rose by 14 to 19 percentiles over the long run. Lower mobility for daughters primarily reflects diminished upward mobility from the bottom….”

Absolute Mobility Among Men

“…Absolute mobility is concerned with changes in real (inflation-adjusted) income between generations. If poor children stay poor as adults but see large income gains over what their parents received, they experience upward absolute mobility. Using a summary measure often described incorrectly as an indicator of relative mobility (the “intergenerational elasticity”), mobility comparing fathers’ and sons’ earnings changed minimally, possibly increasing between the 1952-59 and 1976-83 birth cohorts. Using parents’ family income as a benchmark, the change in mobility over 30 years is minimal, though the shortterm trends in the PSID and NLS look very different. Mobility was the same in 2014 (among men born 1982-84) as in 1981 (among men born 1949-51) in the sense that the absolute gaps between sons in childhood were similarly narrowed by adulthood in both years.For example, a 10 percentage-point gap between sons in childhood typically narrowed to a 2.0-point gap in adulthood in 1981. In 2014, it only narrowed to a 2.5-point gap. (The elasticity rose from 0.21 to 0.26.) This estimated change in the NLS samples was small enough that we cannot discount the possibility that no change occurred in the real world represented by these samples. Turning to the likelihood that sons will do better than their parents, I find that there has been little change in the share of men who have higher earnings than their fathers.It appears that there was little change in the share of sons whose family income exceeds that of their parents, though that share falls in the PSID when incomes are adjusted for family size. In the NLS, for instance, the share was 51 percent among sons born 1949-51 and 53 percent among sons born 1982-84. However, the change in the PSID (using size-adjusted incomes) was from 66 percent to 56 percent, comparing cohorts born 1952-59 to those born 1976-83….”

Contrast With Chetty,"...The finding of relatively flat absolute mobility is in contrast to widely publicized results from Raj Chetty and his colleagues showing a large decline. I argue that an ideal measure, capturing family income over all of childhood rather than in a single year, would show a decline in absolute mobility, though a smaller one than is reported by Chetty’s study and others. The difference between my results and those other studies is that I measure parent income in adolescence rather than near birth. Because income growth was so fast during the 1950s and early 1960s, it is more difficult for sons born mid-century to exceed parental income when the bar is set later in childhood rather than earlier in childhood...."

that compare fathers’ and sons’ earnings…”

among sons born 1976-83. Due to imprecision in these estimates (which come from samples intended to represent the broader population of men), we cannot conclude that every cohort difference in Figure 2 would be found if we could observe the mobility of the entire population of men in both cohorts. Given some simplifying assumptions, statistical techniques allow us to say that if there were no true difference in mobility between two cohorts of men, we would be very unlikely to find a difference of a given magnitude in the PSID data. These techniques suggest that only one of the trends in Figure 2 is likely to be “real” in this “statistically significant” sense.13 Sons raised in the top fourth of father earnings became much less likely to end up in the top fourth themselves (right-most set of bars).14 Despite this evidence of increasing downward mobility, the IRA in Figure 1 does not reflect an increase, because it appears that upward mobility from the bottom fourth fell (though none of the changes in the left-most set of bars are statistically significant). No previous research has examined long-term trends in relative mobility in the US looking at transition probabilities

“…Figure 2 displays four sets of bars, each of them depicting outcomes for men who grew up with fathers in a different part of the male earnings distribution. The left-most bars show results for the 25 percent of sons with the lowest-earning fathers, the next set does the same for the sons of the next-poorest quarter of fathers, the next bars apply to sons with fathers in the third quartile of earnings, and the right-most bars depict the 25 percent of sons raised by the highest-earning fathers. Within each set are two bars, one for the earliest and most recent set of PSID birth cohorts. Each bar, in turn, is divided into four portions, showing where in the distribution of male earnings sons from a given cohort and quartile of the father earning distribution end up. In a world where the ranking of fathers on the earnings ladder had no relationship to the ranking of sons, each of the bars in Figure 2 would be divided into four equal sections. For instance, 25 percent of men starting in the bottom fourth would end up in the bottom fourth, 25 percent would end up in the second fourth, 25 percent in the third fourth, and 25 percent at the top. The analyses here are primarily concerned with the trend in mobility, so the important comparisons are within each set of bars in Figure 2. For instance, the left-most set of bars indicates that 40 percent of sons born 1952-59 and raised in the bottom fourth of father earnings were in the bottom fourth of male earnings themselves when observed as adults. In the subsequent set of birth cohorts, upward mobility out of the bottom fourth became less common over time: 49 percent remained in the bottom fourth. Men who started out in the bottom fourth were also slightly less likely to rise to the top fourth of male earnings over time; that share was 5 percent among sons born 1952-59 and 4 percent

“….In Figure 1, the PSID estimates for four cohorts and NLS estimates for three are displayed as lighter lines, whilethe dark lines display the linear trends through the four (or three) data points. The red lines display the IRA trend for male earnings, from the PSID. The trend shown by the light red line compares sons born 1952-59, 1960-67, 1968-75, and 1976-83, displaying the data points at 1955.5, 1963.5, 1971.5, and 1979.5. Relative mobility was unchanged over this period. The IRA for the earliest set of cohorts was 0.36, and for the most recent cohorts it was 0.32. (Appendix 2 provides estimates for all cohorts for all trends.) The change is far too imprecise to have confidence an actual decline in the IRA (an increase in mobility) occurred. Given the magnitude of the change and the number of men in the PSID samples, statistical methods allow us to assess how likely it would be to find a change in the data even if there were no true change in the American population. The decline in the IRA fails to achieve “statistical significance” in this sense.11 The linear trend line closely resembles the trend between the earliest and most recent cohorts….”

Relative Mobility Among Men

“….Relative mobility is concerned with the extent to which adults transcend their parents’ rankings by moving up o down in ranks. If everyone gets richer, but the poorest children still end up as the poorest adults, there is limited relative mobility. Relative earnings mobility, comparing fathers and sons, may have increased slightly over time. To the extent that it did, it likely reflected increased downward mobility from the top.Relative mobility trends are similar whether parental family income in adolescence is compared with men’s subsequent earnings or with their own family income in adulthood. Relative mobility may have declined, but any change was modest. According to the NLS data, for example, rather than the richest and poorest adolescent sons being separated by just under 25 percentiles in adulthood, as in 1981 (for cohorts born around 1950), they were separated by around 30 percentiles in adulthood in 2014 (for cohorts born in the early 1980s). There is evidence that both upward mobility from the bottom and downward mobility from the top declined. When sons’ earnings are the outcome, both the PSID and NLS indicate reduced downward mobility, and both indicate reduced upward mobility when considering sons’ family income….”

Evidence, “…The red line in Figure 16, below, displays estimates of absolute family income mobility for the four sets of PSID cohorts. For cohorts born 1952-59, 47 percent of sons had higher earnings in their early 30s than their fathers had when the sons were between the ages of 8 and 21. The figure for sons born 1976-83 was 50 percent—no different in the sense of being statistically significant. In contrast, the blue line in the chart comes from the Chetty et al. data and shows the trend for 1949-84 birth cohorts comparing the individual incomes of fathers and sons. (Individual income includes non-earnings income, so the trend is not strictly comparable to the PSID trend.) The four blue dots plot the eight-year averages for the same sets of birth cohorts that are combined in the PSID. The Chetty estimates indicate that absolute mobility fell dramatically over time. Across the cohorts covered in the PSID, the decline was from 61 percent to 42 percent…. If absolute mobility has, in fact, been flat in recent decades, that would suggest that two-thirds to 80 percent oftoday’s forty-somethings are better off than their parents were at the same age—much higher than the 50 percentsuggested by Chetty et al. Similarly, Part One of this primer reported estimates from the PSID, using size-adjustedfamily income and pooling sons and daughters. The range of those estimates ran from 63 percent to 77 percent forrecent cohorts of adults…”

Note Contrast With Chetty,"...The finding of relatively flat absolute mobility is in contrast to widely publicized results from Raj Chetty and his colleagues showing a large decline. I argue that an ideal measure, capturing family income over all of childhood rather than in a single year, would show a decline in absolute mobility, though a smaller one than is reported by Chetty’s study and others. The difference between my results and those other studies is that I measure parent income in adolescence rather than near birth. Because income growth was so fast during the 1950s and early 1960s, it is more difficult for sons born mid-century to exceed parental income when the bar is set later in childhood rather than earlier in childhood...."

"...Europe has overtaken China in creating billion-dollar tech startups, according to a Analysis that points to Europe’s rising status in the field and the effect of Beijing’s crackdown on capitalist entrepreneurialism. Both economies still lag behind the U.S. in creating so-called unicorns....While China still has far more large tech companies than Europe, and a larger internal market for them, trends have shifted over the past year according to research by Atomico, a London-based venture-capital firm that produces an annual analysis, the State of European Tech.... This year, China has added 26 tech companies with valuations above $1 billion, known as unicorns, bringing its total to 300, according to Atomico and data from analysts Dealroom.co. Europe has 98 new unicorns this year, bringing its total to 321, according to Atomico. It defines Europe as all 27 European Union countries and 18 others, including the U.K., Norway and Switzerland... The U.S. still leads Europe and China with 1,178 unicorns, of which 384 were added this year, according to Atomico. Europe has also leapfrogged China on venture capital investment this year, according to Atomico. In the first nine months of 2021, China drew $45 billion in venture funding compared with $52 billion for all of 2020, while Europe has already surpassed last year’s total, Atomico said. Through September, venture capital funds invested $77 billion in Europe, up from $48 billion for the whole of last year....European startups also generally carry lower valuations than their U.S. counterparts, attracting deal-hungry American investors. Europe this year has accounted for 33% of early-stage investments of less than $5 million, compared with 35% in the U.S. and 17% in Asia, according to Atomico. Europe’s share falls and the U.S. share grows with the size of funding rounds, Atomico says, offering a reminder that more money is chasing bigger deals in the U.S. than Europe...."

Daniel Michaels, “Europe Tops China in Spawning $1 Billion Tech Startups,” Wall Street Journal, December 6, 2021,
https://www.wsj.com/articles/europe-tops-china-in-spawning-1-billion-tech-startups-11638869014

Europe Tops China in Spawning $1 Billion Tech Startups

BRUSSELS—Europe has overtaken China in creating billion-dollar tech startups, according to a Analysis that points to Europe’s rising status in the field and the effect of Beijing’s crackdown on capitalist entrepreneurialism. Both economies still lag behind the U.S. in creating so-called unicorns.

Over the past year, China has reined in its most successful and freewheeling tech giants, including Alibaba Group Holding, Tencent Holdings and Didi Global Inc., leading to a loss of market value that has cast a chill across the whole sector. Europe, meanwhile, has attracted further investment in its expanding tech industry, particularly from U.S. funds.

While China still has far more large tech companies than Europe, and a larger internal market for them, trends have shifted over the past year according to research by Atomico, a London-based venture-capital firm that produces an annual analysis, the State of European Tech.

This year, China has added 26 tech companies with valuations above $1 billion, known as unicorns, bringing its total to 300, according to Atomico and data from analysts Dealroom.co. Europe has 98 new unicorns this year, bringing its total to 321, according to Atomico. It defines Europe as all 27 European Union countries and 18 others, including the U.K., Norway and Switzerland.

The U.S. still leads Europe and China with 1,178 unicorns, of which 384 were added this year, according to Atomico.

Europe has also leapfrogged China on venture capital investment this year, according to Atomico. In the first nine months of 2021, China drew $45 billion in venture funding compared with $52 billion for all of 2020, while Europe has already surpassed last year’s total, Atomico said. Through September, venture capital funds invested $77 billion in Europe, up from $48 billion for the whole of last year.

Both Europe’s number of unicorns and venture capital inflows are records, according to Atomico partner Tom Wehmeier, who leads the preparation of the report.

Europe’s venture funding total is near that of Asia overall, which saw venture capital investments of $110 billion through September, versus $87 billion in all of last year, Atomico said.

Investors, particularly from the U.S., have been drawn to Europe by an expanding number of tech startups launched by a growing population of entrepreneurs, many of whom have already worked in other tech companies. Over recent years, European governments and the EU have tried to encourage people to start businesses and embrace a degree of risk, particularly as the continent’s traditional social safety nets have strained under economic pressure and demographic shifts.

European universities’ strong science and engineering programs have long fed into large industrial enterprises but increasingly send graduates to the tech industry.

Tech-sector ties across the Atlantic have deepened recently, particularly as China and other parts of the world have become less welcoming. U.S. tech companies including Amazon.com Inc., Alphabet Inc. and Facebook parent Meta Platforms Inc. have established research centers in Europe or partnerships with universities to tap into the continent’s talent pool. Specialists can be less expensive to employ in Europe than in U.S. high-tech centers such as around San Francisco or Seattle.

European startups also generally carry lower valuations than their U.S. counterparts, attracting deal-hungry American investors. Europe this year has accounted for 33% of early-stage investments of less than $5 million, compared with 35% in the U.S. and 17% in Asia, according to Atomico. Europe’s share falls and the U.S. share grows with the size of funding rounds, Atomico says, offering a reminder that more money is chasing bigger deals in the U.S. than Europe.

China, meanwhile, has pushed to bring greater state control over its private sector. President Xi Jinping has said companies should do more to serve the Communist Party’s economic, social and national-security concerns. Along the way, the government has launched probes, derailed public offerings and clipped the wings of recently highflying tech entrepreneurs, erasing hundreds of billions of dollars in market value in the process.

Atomico’s Mr. Wehmeier said Europe’s relative predictability—which many people criticize as inertia or stodginess—has become an advantage relative to the situation in China now and to the unpredictability in the U.S. in recent years.

“Investors struggle with uncertainty,” Mr. Wehmeier said. “Europe’s stability is a great foundation for building a healthy tech ecosystem, and that has become more apparent recently.”

Cathy Rampell argues that the slowdown in immigration, relative to the pre-Trump trend, and a backlog of employment authorization applications are likely having a significant impact on labor supply, "...But there’s one underappreciated factor contributing to labor shortfalls that the Biden administration could alleviate almost immediately: the “missing” immigrant workers. Immigration inflows slowed sharply during the Trump administration and then collapsed under the combination of Trump-era policies and pandemic-driven closures. The number of visas issued by the State Department’s Foreign Service posts, for example, fell by more than 60 percent between fiscal years 2016 and 2020. There are millions fewer immigrants here today than would have been the case if pre-Trump trends in immigration had continued. But the labor force is also losing immigrants already here legally, whose work permits are expiring because the Biden administration hasn’t gotten its act together....But the queue to get any specific application reviewed is long — and growing. As of June 30, nearly 1.4 million employment authorization applications were pending. That’s double the size of the backlog that existed right before covid-19 reached the United States. It’s triple the number from the quarter just before Donald Trump took office...."

Cathy Rampell, "The missing immigrant workers,"Washington Post, November 22, 2021, https://www.washingtonpost.com/opinions/2021/11/22/legal-immigrant-workers-paperwork-renewal-backlog/

The missing immigrant workers

Helen Muradyan, a second-year resident physician, stopped working last month.

Not because her skills aren’t needed. To the contrary: The Southern California community hospital and health clinic that employed Muradyan struggle to find staff even during normal times. The pandemic worsened their staffing shortages.

“At one point we were operating at 150 percent of capacity,” Muradyan told me. “We worked day, night. We worked without breaks or anything, without seeing our loved ones, without seeing our family.”

But Muradyan, an immigrant from Armenia, had to stop working — because the U.S. government couldn’t be bothered to process her application to renew her work permit. Eventually, her existing work permit expired, and her employers had to terminate her.

Many factors contribute to our nationwide labor shortages, which are, in turn, driving supply-chain problems and inflation. Most of those issues — lack of child care, early retirements, fear of getting ill, burnout — would be difficult for employers or policymakers to resolve even if wages rise. But there’s one underappreciated factor contributing to labor shortfalls that the Biden administration could alleviate almost immediately: the “missing” immigrant workers.

Immigration inflows slowed sharply during the Trump administration and then collapsed under the combination of Trump-era policies and pandemic-driven closures. The number of visas issued by the State Department’s Foreign Service posts, for example, fell by more than 60 percent between fiscal years 2016 and 2020. There are millions fewer immigrants here today than would have been the case if pre-Trump trends in immigration had continued.

But the labor force is also losing immigrants already here legally, whose work permits are expiring because the Biden administration hasn’t gotten its act together.

For immigrants in the country lawfully, renewing an existing work permit was once relatively straightforward. When Muradyan applied for a renewal in 2019, she recalls the process took two to three months. That’s close to the average processing time such applications required in recent years.

Her experience this year has been very different.

U.S. Citizenship and Immigration Services (USCIS) received her renewal application in April; then, for months, Muradyan heard almost nothing. Requests to expedite processing were rejected twice, then ignored. On Oct. 13, her existing permit expired. It might be months more before USCIS reviews her pending application: The agency field office handling her case estimates that similar applications are taking 9.5 to 11 months to process.

Terminations that result from these delays affect more than workers. Muradyan’s employers, for instance, have had difficulty covering her scheduled shifts.

“Sometimes, my colleagues call me asking like, ‘Okay, your patient came — what’s supposed to happen for her exactly?’” Muradyan says. “My hospital is struggling, my colleagues are struggling, my patients are struggling.”

Muradyan is struggling, too.

An asylum seeker, Muradyan lives with her elderly parents and her 10-year-old U.S.-born son, who has started asking whether they will be able pay their rent. She lost her health insurance, which was tied to her employment. The longer it takes USCIS to process her paperwork, the worse the possible consequences get: Under her residency program’s rules, if she misses 12 weeks of work, she has to redo an entire year of training (and would have to wait until July 2022 to restart). Her license to prescribe medication will expire if she is out of work too long.

She is hardly the only person hurt by USCIS’s backlog.

Muradyan is a plaintiff in one of several national class-action lawsuits against USCIS over processing delays. A nonprofit representing the co-plaintiffs in her suit, the Asylum Seeker Advocacy Project, says at least 2,000 of its members have either recently lost their jobs because of work-permit processing delays or are on the verge of having their documents expire.

These asylum seekers have been employed in industries experiencing high-profile labor shortages: truck driving, food services, nursing, tech. And some firms that employ these unlucky immigrants are losing multiple workers.

Abelardo Rios is a telecommunications field technician who has been installing 5G equipment in Florida. His employer sometimes struggled to recruit workers even before covid-19, he says, “because not everybody is willing to go up a tower that’s 400 feet.”

Nonetheless, he and his direct supervisor are being let go this month because their work-authorization renewal applications are stuck in the USCIS backlog.

“My employer basically has told us that if they keep us working, they would be breaking the law, and they can’t do that,” Rios said. Because he, too, wants to stay on the right side of the law, he’s not sure how he’ll make his next mortgage payment. Like Muradyan, Rios is his family’s sole breadwinner.

Cases such as Rios’s or Muradyan’s shouldn’t take long to adjudicate: On average, USCIS employees spend only 12 minutes before rendering a decision on each employment authorization application, according to a 2019 agency estimate.

But the queue to get any specific application reviewed is long — and growing. As of June 30, nearly 1.4 million employment authorization applications were pending. That’s double the size of the backlog that existed right before covid-19 reached the United States. It’s triple the number from the quarter just before Donald Trump took office.

Multiple factors have expanded the backlog: Pre-pandemic, Trump officials had deliberately slowed down processing of nearly all immigration-related applications. The red tape and other obstacles they added increased costs not just for applicants but also USCIS. The agency had a budget crisis last year and ultimately froze the hiring of new employees and laid off contractors.

Covid safety measures then further slowed processing.

The Biden administration has reversed many Trump-era policies but the agency remains understaffed. And shrinking the work-permit backlog does not appear to be a Biden administration priority, even as it touts a “whole-of-government” approach to unblocking supply chains and reducing inflationary pressures.

When asked what role the constricted immigrant labor force might play in today’s economic bottlenecks, White House officials often change the subject. When I asked USCIS what it is doing to shorten wait times for employment documents, the agency said it was “committed to using all available policy and operational improvements to reduce both the number of pending cases and overall processing times,” and cited several changes it has made to expedite processing. (Some of those were the result of a recent settlement in a different class-action case.)

The problem, says Cato Institute research fellow David Bier, is that the administration has "done a lot of the reversing of the bad stuff that Trump did to slow things down even more, but not enough to go above and beyond that necessary to get rid of the backlogs and added delays that the Trump policies originally created.”

So what could USCIS do, while it rebuilds its staff?

Eliminating some redundancies from the process is a good start. The agency could also automatically extend more existing work permits while applicants wait for their renewals to be processed. Some categories of noncitizens already get this courtesy, but even those grace periods are too short. Asylum seekers, for instance, have their existing work permits auto-extended for up to 180 days while the renewal application awaits processing. But as Muradyan’s case and others show, that’s not enough time to avoid a lapse in employment.

“It doesn’t make sense, that if someone really wants to work, they can’t go back to work because of these delays,” she said. “It’s like they just don’t care.”

The FRBNY's Dean Parker and Moritz Schularick find that term spread is a useful indicator of financial crisis both domestically and overseas The effect is grounded in an uptick at the short end of the yield curve likely driven by higher risk taking by financial intermediaries in the run up to a crisis, "...It is clear that theterm spread is significantly lower in the years before a financial crisis, both in the international sample and in the U.S. case. When all countries are considered, the term spread is approximately one percentage point lower on average in the two years before the onset of a crisis. For the United States specifically, this effect is almost doubled in the year before an onset, reaching 2 percentage points lower than usual on average.For a concrete example, in the run-up to 2007-08 financial crisis in the United States, long-term rates declined despite increasingly tight monetary policy, giving rise to Alan Greenspan’s famous description of the development as a “conundrum.” As a result, the term spread fell by 3 percentage points at the time. Similarly, though not as widely discussed contemporarily, the term spread declined a total of 3.3 percentage points in the United States between 1924 and 1928 in the lead-up to the 1929 crash and Great Depression. In the run-up to normal recessions, however, we observe only limited movement.... There are various possible explanations for the predictive ability that depend on whether the decrease of the term spread is driven by the short-term rate rising or the long-term rate falling. As seen in the panel chart below, when taken in isolation, rising short-term rates seem to play the dominant role for the narrowing of the term spread. This is because the long-term rate does not seem to behave differently in the run-up to a crisis. Such an increase in short-term funding rates relative to the yield on long-term investments is indicative of declining interest margins and higher risk taking by financial intermediaries. Our findings hence echo the evidence for rising financial risk taking in financial markets before financial crises that is reflected both in quantities of credit growth and a depressed price of risk..."

Dean Parker and Moritz Schularick, "The Term Spread as a Predictor of Financial Instability," Liberty Street Economics, November 24, 2021,
https://libertystreeteconomics.newyorkfed.org/2021/11/the-term-spread-as-a-predictor-of-financial-instability/

The Term Spread as a Predictor of Financial Instability

The term spread is the difference between interest rates on short- and long-dated government securities. It is often referred to as a predictor of the business cycle. In particular, inversions of the yield curve—a negative term spread—are considered an early warning sign. Such inversions typically receive a lot of attention in policy debates when they occur. In this post, we point to another property of the term spread, namely its predictive ability for financial crisis events, both internationally and in historical U.S. data. We study the predictive power of the term spread for financial instability events in the United States and internationally over the past 150 years.

The Term Spread as a Crisis Predictor

The behavior and slope of the yield curve are longstanding tools for predicting the onset of economic recessions within the next year, or future employment growth. Consideration of the yield curve as a predictor for recessions rose to prominence with Estrella and Mishkin’s (1996) paper. Rudebusch and Williams (2008) also highlighted the power of the term spread in predicting aggregate economic downturns. While there are many academic studies and other articles on the efficacy of the term spread and its inversion for predicting the onset of normal recessions, little is known about its forecasting performance for financial crises. Our study aims to close this gap.

In normal times, a bond with a longer maturity fetches a higher rate on the market due to increased duration risk. However, in the lead up to an economic downturn, various factors can alter the shape of the yield curve. Changes in risk attitudes and search for yield, central bank policy actions, and international demand for safe assets can all influence the term spread and drive the yield for short-term debt closer to or even above the yield for long-term debt. Are such changes in the term spread predictive of a higher risk of financial instability?

The Behavior of the Term Spread around Financial Crises

For our analysis, we use the Macrohistory Database compiled by Jordà, Schularick, and Taylor (2017). The data cover eighteen advanced economies around the world from 1870 to 2017 and contains country specific short- and long-term rates. For the United States specifically, these are the three-month Treasury bill and ten-year Treasury bond rates, respectively. The main benefit of the Macrohistory Database is that it enables examining financial crises in a variety of economic and temporal contexts. Our international sample consists of eight-three crises out of 2,304 total observations, or approximately four percent of the sample. The United States has six years marked as crisis years over the past 128 years: 1873, 1893, 1907, 1930, 1984, and 2007. A financial crisis is defined by Jordà, Schularick, and Taylor as a period of banking distress characterized by major bank failures, exceptional losses in the banking sector, and/or significant government intervention. Briefly, the 1873 and 1893 crises were bank runs precipitated by railroad failures, 1907 was brought on by the collapse of Knickerbocker Trust, 1984 marks the beginning of the Savings & Loan crisis, and 1930 and 2007 saw the Great Depression and Global Financial Crisis, respectively. We also corroborate our results using a different definition of financial instability, namely large declines in the market value of bank equity (Baron, Verner, and Xiong 2020).

The chart below plots the term spreads in the five years before and after a financial crisis or recession. We test, using regression analysis, whether the level of the term spread is above or below normal levels in these windows. The blue dots and bars in the chart represent point estimates and confidence intervals, respectively. They can be read as the average levels of the term spread in the lead-up to and aftermath of an event compared to periods where no event occurred. Positive values imply a higher level of the term spread than usual and negative values imply a lower level than usual.

Using the Term Spread as a Forecasting Tool

With the knowledge that something abnormal occurs in the term spread in the years preceding a financial crisis, the natural next step is to examine its efficacy as a predictive tool and how it functions when used as part of a larger model including other known predictors of crises. Schularick and Taylor (2012) show that real bank-loan growth is a strong predictor of impending crisis and that model serves as the base for our later models. As for definitions of financial crises, we use both the Jordà-Schularick-Taylor narrative and the bank equity declines from Baron, Verner, and Xiong (2020).

We run logistic regressions to predict the two types of crises and compare models using real bank-loan growth and term spread as predictors. We also test the inclusion of an inverted yield curve (a negative term spread). The results are reported in the table below with each column reporting a separate regression.

Our main finding is that, for both crisis definitions, including the term spread alongside standard predictors yields a significant improvement in forecasting power. A smaller term spread increases the risk of a financial crisis. Using the inverted yield curve as a predictor results in a slightly worse predictive ability of the model.

There are various possible explanations for the predictive ability that depend on whether the decrease of the term spread is driven by the short-term rate rising or the long-term rate falling. As seen in the panel chart below, when taken in isolation, rising short-term rates seem to play the dominant role for the narrowing of the term spread. This is because the long-term rate does not seem to behave differently in the run-up to a crisis. Such an increase in short-term funding rates relative to the yield on long-term investments is indicative of declining interest margins and higher risk taking by financial intermediaries. Our findings hence echo the evidence for rising financial risk taking in financial markets before financial crises that is reflected both in quantities of credit growth and a depressed price of risk (Krishnamurty and Muir 2017; Baron and Xiong 2017).

The term spread is often used as an early warning indicator for recessions. In this post, we showed that it performs well as a predictor of financial crises, both internationally and in the United States alone. There is a significant benefit to including the term spread as a predictor for two separate crisis definitions. We identified that this effect is driven by the short end of the yield curve rising and offered a potential explanation based in higher risk taking by financial intermediaries in the time before a crisis.

“…As companies across the U.S. fight to find workers, Amazon is emerging as a de facto wage-and-benefit setter for a large pool of low-skilled workers. Business experts have long researched what is known as the Amazon effect in disrupting traditional retailers. Now Amazon’s every move is causing ripple effects well beyond the retail space in local markets throughout America, including on inflation, regional job markets and labor standards, according to an examination of federal labor data and interviews with economists, researchers, local employment officials and current and former Amazon employees….“If they are not leading, they are reinforcing it,” said Lynn Reaser, a professor at Point Loma Nazarene University and longtime economist at financial institutions that include Bank of America Corp. “Everyone is comparing job offers, and they always have Amazon as a benchmark.” Even Amazon’s own internal employee challenges ripple through the market. The company’s turnover rate has exceeded more than 100% across many of its facilities, according to an analysis by The Wall Street Journal….A 10% increase in Amazon’s advertised hourly wages in 2018 led to an average increase of about 2.6% among other employers in labor markets where Amazon is located, according to a paper this year by researchers from the University of California, Berkeley and Brandeis University. Amazon’s influence on wage increases had a big effect because a large fraction of similar jobs then were below $15 an hour, researchers concluded. In comparison, when Walmart and Target Corp. announced $9 starting pay in 2015, the effect was smaller because there was a larger fraction of employers that were already at or above that pay level. Walmart in September raised its minimum wage to $12 an hour. At the same time, the researchers found Amazon’s increase in pay failed to raise overall employment levels and actually led to a small decline. While some employers that raised wages hired additional workers, others cut back on employment or hours. In Amazon’s case, the company’s increase to $15 an hour led to an average decrease in “probability of employment” by 0.8 percentage point, the paper says. Other research showed similar results….”

Sebastian Herrera, "Amazon Emerges as the Wage-and-Benefits Setter for Low-Skilled Workers Across Industries,"Wall Street Journal, December 7, 2021, https://www.wsj.com/articles/amazon-emerges-as-the-wage-and-benefits-setter-for-low-skilled-workers-across-industries-11638910694

Amazon Emerges as the Wage-and-Benefits Setter for Low-Skilled Workers Across Industries

Max Ramirez knew he had a problem when he found out about the truck.

About two years ago, Amazon.com Inc. employees rigged a vehicle to carry a makeshift billboard advertising starting pay of roughly $16 an hour. They drove the truck all over the small Texas city where Mr. Ramirez helps run a rival warehouse operation.

Within a few months, a handful of the employees at his company, mattress manufacturer Serta Inc., had decamped to Amazon. “We had no choice but to compete,” he said. The company raised its starting pay by roughly $2 to about $15 an hour and has since raised it about another dollar, he said.

As companies across the U.S. fight to find workers, Amazon is emerging as a de facto wage-and-benefit setter for a large pool of low-skilled workers. Business experts have long researched what is known as the Amazon effect in disrupting traditional retailers. Now Amazon’s every move is causing ripple effects well beyond the retail space in local markets throughout America, including on inflation, regional job markets and labor standards, according to an examination of federal labor data and interviews with economists, researchers, local employment officials and current and former Amazon employees.

The nation’s second-largest private employer is planning mock fulfillment centers in high schools to plant the seeds of future careers, sending recruiters to local fairgrounds and bombarding job boards with promises of large sign-on bonuses and pay—in some cases nearly triple the federal minimum wage.

The effect is magnified because Amazon churns through hundreds of thousands of employees each year, creating an even more voracious appetite for labor that often compels the company to push up compensation or improve recruitment in other ways—especially during peak times such as the holiday shopping period now under way.

“Amazon has the economies of scale,” said Jesse McCree, a workforce development official in Harrisburg, Pa., an area of the country where Amazon is competing heavily with other large logistics and warehouse companies. “They are influencing the market because of scale and name recognition and can afford to pay more than the smaller guys. As they go, even the big companies are going to pay attention.”

Similar effects are evident in areas near Austin, Los Angeles, Cincinnati and Louisville, Ky., according to a Journal review of local data and interviews with employers and workforce officials.

At produce distributor Castellini near Cincinnati, Chief Executive Brian Kocher can’t get away from Amazon. The company has wrapped job advertisements around cars, buses and billboards. And recently, when Mr. Kocher went to play his favorite Solitaire game on his iPhone, Amazon ads popped up there, too.

Castellini in the past year has raised wages three times, with its pay now starting around $16 an hour. Since many of its employees in the area are Spanish speakers, Castellini has recently focused on hiring and promoting managers who speak the language to better connect with workers. The company also implemented $750 bonuses for any employee who refers family and friends to work there.

Amazon has had a significant impact on the area since 2017, when the company struck a deal with the Cincinnati/Northern Kentucky International Airport to open a $1.5 billion air hub. Paul Verst, chief executive of Verst Logistics, which provides warehousing, transportation and packaging services for clients such as manufacturers and consumer-goods companies, said construction costs have risen by at least $30 a square foot to a range of about $90 to $100 due to increased demand for building space.

Mr. Verst recently gave employees a $3-an-hour raise to compete with Amazon. Starting pay now ranges from $16 to $19 an hour. He said his family-owned company aims to retain employees by connecting with them personally. He signs a birthday card for each worker. Tenures for many workers have averaged 10 to 15 years, he said. Still, the company has lost a handful of employees to Amazon, which has advertised pay of $20 or more an hour and $1,000 sign-on bonuses in the area.

“It was for economic reasons that they left,” he said.

There have been about two job openings for every unemployed person in Cumberland County, the Harrisburg-area county where several Amazon facilities are located. Warehouse competitors include pet food retailer Chewy Inc., United Parcel Service Inc. and food and agriculture giant Cargill Inc.

Wage wars in the area have been fierce ever since Amazon raised its starting pay nationally by several dollars to $15 an hour in 2018, local officials said. Much of the battle for hourly employees has played out near a stretch of the area’s Interstate 81 highway, where companies have erected billboard after billboard advertising sign-on bonuses and “immediate openings.”

On occasion, Chewy workers have left the company to work at Amazon almost immediately after receiving new training, according to a former area manager. During the Covid-19 pandemic, a period when Amazon hired workers as aggressively as any company in modern history, Cargill was at times so short of employees that it flew workers in from other locations, according to the company. Both Chewy and Cargill now advertise pay near $20 an hour in the area.

Employee turnover in Cumberland County rose after Amazon’s arrival spurred competition among local firms for workers. Three years ago, around the time when Amazon bumped its starting pay to $15 an hour, wages for warehouse employees in Cumberland averaged from $10.50 an hour to $12.50 an hour. Now, they range between $15 to $21 an hour, according to the Cumberland Area Economic Development Corporation.

“Amazon is the standard-bearer,” said Zach Pasquariello, a former Chewy area manager in the region. “Chewy was always following in Amazon’s footsteps and trying to do what Amazon does, but we were always a little bit behind.” Chewy declined to comment.

Job openings across the U.S. outnumber the people who are unemployed, Labor Department figures have shown, demonstrating an unusual tightness in the labor market that has seen a sharp rise in wages. How much of this is due to Amazon’s influence is difficult to pinpoint, due to limited data and the unique influence of the pandemic.

But as Amazon’s footprint has grown rapidly across the country, the potential for the company to influence wages or other market dynamics has increased, economists say. Amazon, which had around 1.4 million total employees at the end of September, hires hundreds of thousands of people every year, putting it on pace to surpass Walmart Inc. as the nation’s largest employer in a matter of years.

“If they are not leading, they are reinforcing it,” said Lynn Reaser, a professor at Point Loma Nazarene University and longtime economist at financial institutions that include Bank of America Corp. “Everyone is comparing job offers, and they always have Amazon as a benchmark.”

Even Amazon’s own internal employee challenges ripple through the market. The company’s turnover rate has exceeded more than 100% across many of its facilities, according to an analysis by The Wall Street Journal.

Amazon has recorded higher injury rates than the national average, and its speedy delivery requirements can quickly burn workers out. The company has faced lawsuits, union challenges and government intervention related to the treatment of its workforce, which has pushed it to introduce new safety measures such as body mechanics training for employees and vows from its top leaders to better listen to workers.

Amazon has said it is working to better understand the needs of its employees and has opposed unions because it prefers to negotiate with workers directly. The company also has said that many of the people it adds are re-hires, demonstrating that many workers return to the company after having earlier left.

Amazon’s wage increases pass through to workers outside the company, researchers have found, as many employers raise their own pay to combat churn. In September, Amazon announced that its starting wage now averages $18.32 an hour, an amount that’s nearly triple the federal minimum wage of $7.25 an hour.

To fight off Amazon, competitors have tried to offer lighter workloads, more flexible schedules, bonus pay and other perks. But Amazon is rolling out new plans to compete in those areas as well.

Amazon wants to use its size and scale to make its jobs as flexible as possible, J. Ofori Agboka, an Amazon senior human-resources executive, said in an interview. Having hundreds of thousands of employees makes it easier to offer workers different work hours, a perk many have requested, he said. Amazon recently broadened a program that allows some employees to switch schedules and pick their own work hours, or cancel a shift at the last minute. The company also offers a child-care network to employees and flexible hours for a few weeks for employees transitioning back to work.

“What does flexibility mean for each employee, and how can we meet that?” Mr. Agboka said.

Candidates are now essentially being hired on the spot, Mr. Agboka said, with many workers able to see their start date less than half an hour after beginning an application online. Amazon is also working to fix common retention issues, he said, such as employees who are dismissed after a minor incident like missing work due to an emergency.

It is too early to know how well some of Amazon’s new initiatives to increase flexibility will be felt in the labor force, but its pay increases are already having a widespread impact.

LaShay Moran, who lives near Louisville, joined Amazon during its current hiring spree. Ms. Moran, 42, left her job at auto parts distributor Premier Performance LLC in part because Amazon offered more money—$17.50 an hour versus $16 at Premier—and a $3,000 sign-on bonus.

“Everybody knows what it is,” she said of Amazon’s name recognition.

A 10% increase in Amazon’s advertised hourly wages in 2018 led to an average increase of about 2.6% among other employers in labor markets where Amazon is located, according to a paper this year by researchers from the University of California, Berkeley and Brandeis University.

Amazon’s influence on wage increases had a big effect because a large fraction of similar jobs then were below $15 an hour, researchers concluded. In comparison, when Walmart and Target Corp. announced $9 starting pay in 2015, the effect was smaller because there was a larger fraction of employers that were already at or above that pay level. Walmart in September raised its minimum wage to $12 an hour.

At the same time, the researchers found Amazon’s increase in pay failed to raise overall employment levels and actually led to a small decline. While some employers that raised wages hired additional workers, others cut back on employment or hours. In Amazon’s case, the company’s increase to $15 an hour led to an average decrease in “probability of employment” by 0.8 percentage point, the paper says. Other research showed similar results.

In response to its overall effect on employment, Amazon has previously pointed to the billions it has invested in infrastructure and the amount of jobs it creates.

In San Marcos, Texas, Amazon established its first fulfillment center in 2016 and soon approached the nearby Texas State University as the area’s largest employer.

Other warehouse operators in the area quickly felt the company’s presence. Amazon advertised on radio stations hundreds of miles away and attracted workers from competitors in short order.

Amazon staffs more than 4,500 workers in three facilities throughout the surrounding Hays County, and a fourth site recently opened that will employ at least hundreds more. The company has added more than 2,000 employees during the Covid-19 pandemic, according to local data.

Worker churn has been high at the facilities. Turnover in Hays County, which is located southwest of Austin, swelled to 101% in 2017, the first full year Amazon operated there. The rate plunged to 68% by 2019, according to a Journal analysis. That decrease happened after Amazon reduced the number of employees at the San Marcos warehouse, according to the Greater San Marcos Partnership, a local business group.

Since Amazon raised its starting pay in 2018, wages in the San Marcos area are also up substantially. Pay for employees in the warehouse and storage industry in Hays County grew by 12.7% in the two years after Amazon’s move to $15, compared with 4.3% in the two years preceding it, according to labor market research firm Emsi Burning Glass.

Walmart and others matched and sometimes exceeded Amazon’s pay at their facilities to stay competitive, said Jason Giulietti, president of the Greater San Marcos Partnership. A large sign recently hung from a Walmart distribution center south of Amazon’s San Marcos facility advertising pay of up to $20 an hour.

At lighting manufacturing giant Signify, which has a warehouse in Hays County, plant manager Haiver Montenegro said he competes with Amazon using flexible schedules, including usually not requiring employees to work on weekends.

“If we were just to compete with hourly rates with Amazon, we would never get there,” he said. “The biggest discussion around resources in this area has to be around what an employer has to offer as a whole package.”

Still, he is short about 15 production associates, with roughly 120 manufacturing employees at the warehouse, according to the company. Signify recently raised starting wages there to around $15 an hour in part because of Amazon’s wage hikes, and he is considering putting up job advertisements on billboards for the first time.

"...The number of U.S. homes with a married couple and kids fell to a record low, according to new government data, as the pandemic further delayed weddings and more adults don’t plan to have kids at all. The share of the U.S.’s 130 million households headed by married parents with children under age 18 fell to 17.8% in 2021 from 18.6% last year, according to the Census Bureau.That’s down from more than 40% in 1970. By absolute numbers, there are just 23.1 million homes with nuclear families, the fewest since 1959, the data show. The pandemic delayed many marriages over the past two years, adding six months to a woman’s age at first marriage -- the most since 1987 -- to now 28.6 years. In the 1950s and ‘60s, women typically married at 20.4 years of age and 22.8 years for men...."

Alexandre Tanzi, "U.S. Married Couple Households With Children Fall to Record Low," Bloomberg, December 3, 2021, https://www.bloomberg.com/news/articles/2021-12-03/u-s-married-couple-households-with-children-fall-to-record-low

U.S. Married Couple Households With Children Fall to Record Low

The number of U.S. homes with a married couple and kids fell to a record low, according to new government data, as the pandemic further delayed weddings and more adults don’t plan to have kids at all.

The share of the U.S.’s 130 million households headed by married parents with children under age 18 fell to 17.8% in 2021 from 18.6% last year, according to the Census Bureau. That’s down from more than 40% in 1970.

By absolute numbers, there are just 23.1 million homes with nuclear families, the fewest since 1959, the data show.

Casey Mulligan on the natural experiment Chicago is running, “… Figure 1 below shows points-in-time numbers of people in Illinois prisons who were convicted in Cook County and admitted to prison within the past half year. The stock of new convicts had been between 3000 and 4000, until the pandemic when it dropped to 1000. Figure 2 limits the sample to convicted murderers. They had been entering Illinois prisons at a rate of about 130 per year (65 per half year), until the first half of 2020 when the rate dropped to about 40 per year (20 per half year).I hope that Chicago-area activists can work to help maintain this progress. Perhaps they could even lend some expertise to the rest of Illinois (not shown in the charts above), which has missed a social-justice opportunity by continuing to admit convicted murderers to prison at pre-pandemic rates….”

I hope that Chicago-area activists can work to help maintain this progress. Perhaps they could even lend some expertise to the rest of Illinois (not shown in the charts above), which has missed a social-justice opportunity by continuing to admit convicted murderers to prison at pre-pandemic rates.

Formerly experts thought that incarceration enhanced public safety. But that was based on old data, examined by researchers lacking sufficient social-justice training. Especially during a pandemic, there is no reason to worry about adverse consequences until somebody has conducted A study using state-of-the-art empirical methods.

A troublesome skeptic might insist on metrics to indicate when the tradeoff-holiday has ended, when the path to safety might again involve more incarceration rather than less. Could it be dangerous to barrel forward on the social-justice highway, when at any moment we could find ourselves in high-speed reverse?

Such a reactionary critic forgets that social justice is crowd sourced. The overwhelming odds are that one of us will notice the situational shift and promptly alert the polity. In the worst-case scenario, remember: if we want to make an omelette, we must be prepared to break a few eggs.

Social Justice Thrives in Cook County during Tradeoff Holiday

During this terrible pandemic, a silver lining has been that many inconvenient tradeoffs no longer apply. A Yale study showed that paying people not to work does not, for the time being, prevent anyone from working. Perhaps also emptying the prisons can enhance public safety. We can have social justice and safety at the same time.

Figure 1 below shows points-in-time numbers of people in Illinois prisons who were convicted in Cook County and admitted to prison within the past half year. The stock of new convicts had been between 3000 and 4000, until the pandemic when it dropped to 1000. Figure 2 limits the sample to convicted murderers. They had been entering Illinois prisons at a rate of about 130 per year (65 per half year), until the first half of 2020 when the rate dropped to about 40 per year (20 per half year).

Casey Mulligan, “Social Justice Thrives in Cook County during Tradeoff Holiday,” supply and demand (in that order), November 13, 2021, http://caseymulligan.blogspot.com/2021/11/social-justice-thrives-in-cook-county.html

“…the time-series of investment and payouts to shareholders reported inFigure 10reveal an interesting pattern. increase in payouts to shareholders outweighs the increase in investment a.er the recent TCJA-17, but not a.er the Kennedy’s tax cuts..is can be observed both at the aggregate and at the c-corporations level. In fact, payouts does not appear to deviate much from the pre-reform trend, unlike investment which exhibits a clear acceleration. The increase in capital formation is gigantic: aggregate business investment is 50% higher in 1966 than in 1963, and c-corporations’ investment is 80% higher….”

Core finding, "... the results are reported in Figure 11. In response to the Kennedy’s tax cuts, the model predicts a large increase in GDP and investment, and a small effect on payouts to shareholders: the opposite of Trump’s Tax Cuts and Jobs Act. Similarly, the corporate tax multiplier for Kennedy’s tax cuts is around 2:5 for GDP, 1:85 for investment, and close to zero for payouts to shareholders. For the TCJA-17, the multiplier is around 0:6 for each variable. For every dollar of lost corporate tax revenues, the Kennedy’s corporate tax cuts stimulated GDP four times more than the TCJA-17. The intuition behind these results is the following. In the early 1960s, the corporate tax rate was high and tax depreciation policy was not accelerated as it was mimicking economic depreciation. As a result, the corporate tax wedge was well below one (around 0:72) before the reform. The Kennedy’s tax cuts increased the wedge significantly (to around 0:84), thus providing strong stimulus to the investment of c-corporations. Moreover, since around 75% of economic activity was taking place in the c-corporate sector, the aggregate effect was less diluted than in 2017….”

Evidence from the Trump vs Kennedy Tax Cuts

Mike Pettis Comment, "In this very interesting paper @Furno_Francesco compares the Kennedy corporate tax cuts in the early 1960s with the recent Trump corporate tax cuts, and finds that the former stimulated output roughly four times more than the latter. Put differently, the Kennedy tax cuts seem to have increased business investment while the Trump tax cuts were mostly passed on to shareholders which, as Atif Mian, Ludwig Straub and Amir Sufi have explained elsewhere, was likely in turn to lead to higher household debt. Furno explains that “A large part of this difference can be attributed to differences in pre-reform tax depreciation policy.” As I argued in my book and elsewhere, I think the reasons the Kennedy tax cuts flowed into investment and Trump tax cuts into savings had to do with the very different relationship between savings and investment at the time of the tax cuts. From a macro point of view corporate tax cuts increase business profits, which raise corporate savings. Whether these are used to fund investment or are passed on to shareholders depends on whether the main constraint on business investment is scarce savings or weak demand. In the former case, higher business profits can lead fairly automatically to higher business investment. If savings are abundant and the cost of capital low, however, the constraint is likely to be weak demand. In that case businesses will have little incentive to increase investment, and a cut in their taxes simply increases the after-tax profits they pass on to shareholders or use to acquire other companies. In the early 1960s, much of the world was still in the process of rebuilding itself after the ravages of two world wars. With high global investment needs and low global savings (the war caused income to plummet and, with it, savings), the main economic constraint was a scarcity of savings needed to fund investment. This was a time when the US was exporting so much of its domestic savings to the rest of the world that capital controls were imposed to restrict the outflow of dollars. Under these conditions it is not surprising that any policy that generated an increase in savings was likely - in classic supply-side fashion - to lead to an increase in business investment. In recent decades, however, conditions were dramatically different. The world has become awash in ex ante savings, and the US has become the dumping ground for nearly half of the world’s excess savings. What is more, US businesses are sitting on huge piles of cash and can only use them to buy back shares or purchase other companies. It is weak demand, in other words, and not scarce savings, that has limited business investment. In that case policies that increase in savings have little to no impact on business investment, and in fact could actually reduce business investment if the increase in business savings were balanced by a reduction in household income (and, with it, consumption). To me this is the key difference between the Kennedy and Trump tax cuts. The former provided more savings at a time when American businesses wanted to invest more, but were unable to do so in part because of the fierce global competition for US savings. The latter provided more savings at a time when the world was awash in excess savings and American businesses wouldn’t invest until they saw a rise in demand. As an aside, I am re-reading Charles Arthur Conant’s 1900 book, and it is amazing how similar the problem of excess savings the global economy faced in the 1880s-90s. At one point he proposes establishing a state pension system to help generate the demand needed to absorb excess global savings. This is because “experience has shown that when saved capital accumulates rapidly, the groping after new uses for it causes waste and disaster”.

Core of paper“…In the model, corporate tax changes affect the economy primarily through the investment decision of c-corporations, which is affected not only by the tax rate but also by tax depreciation policy. Specifically, the possibility to deduct investment from the tax base (partially) counteracts the distortion introduced by the tax rate: the faster investment is deducted from the tax base, the smaller the distortion to the rate of return on investment. As a result, when tax depreciation policy is very accelerated - like it was in 2017 - the rate of return on investment is almost unaffected by corporate tax policy, and a reduction in the corporate tax rate is not particularly expansionary……However, irrespective of how much stimulus is provided to investment, a corporate tax cut always entails a transfer of resources from the government to c-corporations. When pre-reform tax depreciation policy is very accelerated, the tax-savings from a rate cut are not used for investment and are distributed to the shareholders. When pre-reform tax depreciation policy is not very accelerated, instead, the extra cash is used for investment. Moreover, since a change to the corporate tax rate affects only c-corporations, the aggregate effect is diluted by the presence of pass-through businesses. After a rate reduction, pass-through entities are not only excluded from the tax cut, but they are also put at a competitive disadvantage. ‘is happens because they compete with c-corporations in the production of (imperfectly) substitutable goods, which further amplifies the shift of economic activity from pass-through businesses to c-corporations and reduces the aggregate effect even more….”

"…We estimate the impact of district-level schooling mode (in-person versus hybrid or distance learning) on test scores. We combine Spring 2021 state standardized test score data with comprehensive data on schooling in the 2020-21 school year across 12 states. We find that pass rates declined compared to prior years and that these declines were larger in districts with less in-person instruction. Passing rates in math declined by 14.2 percentage points on average; we estimate this decline was 10.1 percentage points smaller for districts fully in-person. Changes in English language arts scores were smaller, but were significantly larger in districts with larger populations of students who are Black, Hispanic or eligible for free and reduced price lunch…”

“…Panel A of Figure 2 shows the year-on-year changes in math test scores by group for prepandemic years (left figure) and for the pandemic year (right figure). In most cases, the test score declines in the pandemic year are larger in the districts with less in-person instruction. This difference is largest in Virginia. There is no evidence that this pattern aligns with the pre-pandemic years. Prior year-on-year changes do not seem to systematically vary with 2020-21 schooling mode. Panel B shows the results for ELA. While the overall changes are smaller in ELA as compared to math, the pattern is similar….”

“…Figure 1 illustrates the changes in pass rates across states during the pandemic period (the 2020-21 school year) and in several previous periods (the 2015-16 school year through 2018-19). Panel A shows the scores in math and Panel B presents the scores in ELA. The solid triangles represent the change in pass rates between the 2018-19 school year (the last pre-pandemic year) and the 2020-21 school year. The open circles show prior year-on-year changes. Our analysis finds a decline in pass rates from the 2018-19 school year to the 2020-21 school year in all states. The decline ranges from -31.9 percentage points (Virginia, math) to -2.3 percentage points (Wyoming, ELA). Based on the prior year changes, typical year-on-year variation is very small. The pandemic year is a significant outlier. The changes in pass rates for math (Panel A) are considerably larger than the changes in scores for ELA (Panel B). This is broadly consistent with a larger literature which shows math scores are more responsive to schooling differences(Betts and Tang, 2011; Angrist et al., 2013)….”

The Evidence

Core findings, “…In this paper, we use data on district schooling modes alongside Spring 2021 test score data from 12 states to analyze the relationship between schooling mode and test scores. We estimate the relationship between the share of the year that in-person learning was offered (as opposed to hybrid or virtual learning) and student standardized test scores, both overall and interacted with district demographics. Our main regression specification uses a standard panel data approach, with data at the district-year level, to estimate the impact of the time spent in-person on 2021 test scores. These regressions include district-specific trends and a full set of demographics and year interactions. We demonstrate first that, across all 12 states in our sample, there were large reductions in test scores between 2020-21 and previous years. The average decline in math is 14.2 percentage points, versus 6.3 percentage points in English language Arts (ELA). These changes are well outside typical year-on-year variation. Following this, we show our main finding: the decline in students' 2021 test scores as compared to prior years was significantly larger in districts which offered less access to in-person schooling. For example, we estimate that offering full in-person instruction rather than fully hybrid or virtual instruction reduces test score losses in math by 10.1 percentage points (on the base of 14.2 percentage points). In ELA, the loss is reduced in fully in-person settings by 3.2 percentage points.These interactions between test score losses and schooling mode are highly significant. We demonstrate, as well, that access to in-person learning varied by demographics and baseline test scores. Districts with lower baseline test scores and a higher share of Black students offered less in-person learning. Within states, districts with a higher share of Hispanic students also offered less in-person learning (although across states this effect is reversed). Finally, we observe that the impact of access to in person learning on 2021 test scores also varies by subject and demographics. Specifically, among districts with a larger share of Black and Hispanic students, districts with less in-person schooling saw a greater decline in ELA test scores than those with more in-person schooling. Although the impact of schooling mode on ELA is fairly small for districts which are majority white, it is large for those districts with a majority of students of color. Meanwhile, the impact of access to in-person learning had a similar effect on math scores for all districts, regardless of their racial composition….”

Clare Halloran, Rebecca Jack, James Okun and Emily Oster, “Pandemic Schooling Mode and Student Test Scores: Evidence from US States,” National Bureau Of Economic Research, November 2021, https://www.nber.org/papers/w29497

New NBER takes a look at a world where ideas are the limiting constraint, "...High productivity entrants gradually replace obsolescing incumbents by poaching their workers, a process that is intermediated via a frictional labor market. When the likelihood of entrants imitating technologies in the tail of the distribution falls (ideas are harder to find), so does growth. Consistent with US data over the past 30 years, firm entry, incumbents’ employment response to productivity shocks, and job-to-job transitions decline, while the share of old firms increases. With lower imitation, however, there is less misallocation, because the slower rate of obsolescence induces productive growing firms to invest more in costly hiring...."

The dynamic, "... First, growth slows down. Second, the implications of a growth slowdown via this mechanism is consistent with facts related to firm dynamics. Employment reallocation falls, as entrants are relatively lower productivity and poach less from incumbents, while incumbents obsolesce at a slower rate. The firm entry rate declines, and the share of employment in old firms increases. Following productivity shocks, firms also hire fewer workers, consistent with Decker, Haltiwanger, Jarmin, and Miranda (2020). As we discuss below, a lower rate of obsolescence reduces misallocation, with workers more efficiently sorted across the productivity distribution of firms. Poaching a worker from another firm to fill a new job is more difficult in a less misallocated labor market so firms expand more slowly following a shock to their productivity...."

Impact on labor market, "....Third, we document the implications for worker dynamics. We find that on all dimensions labor market flows slow down: employment-to-unemployment (EU), unemployment-to-employment (UE) and employment-to-employment (EE) transitions rates decrease. Ceteris parabus, the decline in the UE rate would lead to an increase in unemployment, but the decline in the rate of obsolescence reduces the rate of layoffs. On net, the second force is stronger such that unemployment declines. Overall, we find weaker creative-destruction effects of a faster growth rate on unemployment than in canonical matching models without on-the-job search. One important reason is that in our environment, as in the real economy, half of job separations are EE transitions which do not entail any unemployment spell. Fourth, these observations, combined with a decline in job-to-job mobility may suggest that misallocation is worse when growth is slower due to a lower intensity of imitation. Interestingly, however, we find the opposite. As growth slows, the misallocation that exists in the economy improves. The dispersion of the marginal product of labor across firms declines and the correlation between size and productivity increases. This leads to a higher level of output, despite slower growth. We find that this level difference is quantitatively significant..."

The evidence,"...In summary, we have calibrated a model of endogenous growth at rate g that matches key facts regarding firm and worker dynamics in the US economy over the period 2011-2016.... Figure 1 verifies that the model matches the empirical distribution of firms and employment by firm age and firm size. Most firms are small, but most employment is at large firms. Most firms are old and most employment is at old firms. Second, Figure 2 reveals the dynamics of the distribution of firm size and productivity over the lifecycle that deliver the empirical marginal distributions in Figure 1. The figure shows how misallocation is partially resolved over the lifecycle, with employment and productivity becoming increasingly correlated as firms age. A key feature of the model, which we highlight in BEMV, is that the model can generate small, productive, young firms. Young firms with high productivity are small because they are young and yet to accumulate workers through the frictional labor market. This is a key source of misallocation in the economy, due to search frictions, on which we will return when we discuss our results...."

Adrien Bilal, Niklas Engbom, Simon Mongey and Giovanni Violante, “Labor Market Dynamics When Ideas are Harder to Find,” National Bureau Of Economic Research, November 2021, https://www.nber.org/papers/w29479

Business dynamism

"..Figure 4 shows firm dynamics outcomes as we vary the underlying imitation parameter y. Since the parameter is not of direct interest, we plot these key outcomes as a function of the resulting endogenous growth rate. Not surprisingly, a fall in y (i.e. moving to the right in these figures) is associated with an increase in firm creation (panel A). As imitation from the incumbents is easier, entrants are relatively more productive and firm creation is encouraged. As more entrants attempt to enter and there is more competition for inputs, the price of inputs rises faster. Consequently, incumbent firms exit at a faster pace and the rate of obsolescence increases.This is our counterpart of the creative-destruction effect of Aghion and Howitt (1994). Quantitatively, a fall in the aggregate growth rate of one percentage point annually—broadly consistent with the US experience over the past decades—is associated with a decline in firm entry of about 30 percent. For comparison, the US has experienced a reduction in firm entry of over 40 percent since the early 1980s to now...Also consistent with the US data over the past 40 years, a larger share of firms are old when growth is slower (panel B). This outcome arises naturally in our model in response to a fall in the rate of obsolescence—firms remain in the market for longer, extending the firm life-cycle. The overall job reallocation rate falls with the aggregate growth rate (panel C). The reason is that as the rate of obsolescence declines, firms do not fall behind the market as quickly. Consequently, there is less need to reallocate employment across production units. Quantitatively, as the aggregate growth rate declines by 1 percentage points annually, the job reallocation rate decreases by around 10 percent. For comparison, the US has experienced a decline in job reallocation of roughly 30 percent since the early 1980s to now...."

"...Figure 6 illustrates a surprising result. As growth slows, the misallocation that exists in the economy improves, despite there being less firm and worker turnover.The dispersion of the marginal product of labor across firms falls (panel A) and the correlation between size and productivity increases (panel B). This is consistent with the well documented reallocation of sales in the economy toward larger firms (Autor, Dorn, Katz, Patterson, and Van Reenen, 2017), and leads to a higher level of output, despite slower long-run growth. What explains this finding? Our model is unique in the class of growth models in that firms operate a single product, start small, and then grow via costly hiring of workers. When the rate of economic growth is high, firms’ productivity is quickly slipping behind the productivity of the economy as a whole. High productivity firms are therefore less ambitious with respect to these expansion plans. Hiring is costly and takes time, and has a lower benefit when the firm will be swiftly replaced by better firms. High productivity firms therefore no longer grow as large, which shows up as increased misallocation of labor in the economy. We find that this output loss from misallocation can be quantitatively important. Panel (C) computes a measure of ’lost years’ caused by this level effect. It asks how many years, at the baseline annual growth rate of 1.75 percent, it would take to compensate the downward shift in the level of output caused by misallocation. For example, the additional misallocation that occurs at a growth rate of 4 percent, would be offset by 22 years of economic growth at 1.75 percent. The relation is highly nonlinear: when the economy grows fast, the growth-induced misallocation is weak. Starting from low growth rates, however, a rise in the growth rate of 1 percentage point per year would require over 15 years of baseline-growth to make up for the productivity loss from misallocation. Panel (D) shows that in terms of the present discounted value of output, an improvement in imitation that increases the growth rate to at least 2.9 percent is required to offset the misallocation effects from higher growth. Admittedly, here we do not compute the transition between one balanced growth path and the other, and y is an exogenous parameter. Nonetheless, these calculations suggest that computing the misallocation costs of higher growth may be an important consideration in the formulation of progrowth economic policies that could be implemented to change objects related to diffusion of ideas in the economy...."

Evidence that NAFTA drove working class voters out of democratic party to GOP, “…We begin by showing thatNAFTA had a major, negative effect on employment in areas exposed to Mexican import competition. By 2000, counties in the top quartile of our measure of NAFTA exposure saw a 5-8 log point decline in total employment, relative to the bottom quartile. These losses were concentrated in manufacturing and, importantly, exhibit no pre-trends from the mid 1980s to 1994. While we begin all of our analysis by showing trends with county year raw data, the basic shape of our event-study coefficients are unchanged as we add a large number of controls: pre-period county-level measures (e.g., 1990 manufacturing share of employment, 1990 share with a college degree) interacted with year fixed effects, to control flexibly for other secular changes (e.g., automation, skill biased technological change) that may affect communities differentially across time; the “China shock” measure from Autor et al. (2013) interacted with year fixed effects, to ensure we isolate the NAFTA effect from the rise of Chinese imports; and fixed effects at the state×year level, to pick up any policy or other unobserved variation within states across time... Having documented large, negative local employment effects in communities exposed to Mexican import competition, we then ask how voters in these communities reacted. NAFTA was a major issue in the 1992 U.S. presidential campaign, with Ross Perot making opposition to it a major motivation for his third-party campaign, a third-party campaign that picked up 19 percent of the popular vote, making it the most successful such campaign since Teddy Roosevelt’s run as the Bull Moose Party candidate in 1912....We focus on House election votes in most of our political analysis, as every House seat is up for election every two years, allowing for a balanced panel of years. While NAFTAexposed counties (many in the upper South) begin the sample period more Democratic as measured by House election votes than the rest of the country, they exhibit a sharp change in trend and become increasingly Republican. In contrast to our employment effects, which show no pre-1994 pre-trend regardless of specification, this political turning point occurs in either the 1992 or the 1994 election, depending on the controls we include in the event study analysis. We find this ambiguity unsurprising, given the political salience of NAFTA in the 1992 election, even though its provisions did not go into effect until January 1994. Beyond the ambiguity between 1992 and 1994, there is no political pre-trend in NAFTA exposed counties from 1980 to 1990. The shift we document is large. While these counties are in 1990 the most Democratic of our four quartile groups, by 2000 they are as or more Republican than any of the quartiles. We present a variety of microdata-based evidence that the political shift was indeed due to NAFTA. First, we show that in the areas we define as most vulnerable to NAFTA, survey respondents significantly oppose NAFTA and this opposition continues to the present day. Second, in repeated cross-section data from the American National Election Surveys (ANES) we show that, in each year of survey data from 1986 to 1992, Democrats enjoy a significant and steady advantage among those with protectionist views, but between 1992 and 1996 a significant number of protectionist voters move toward the GOP and remain there. Finally, in an ANES panel dataset from 1992 to 1994, we can look at the same voters over time during this key moment. We indeed find that a significant share of those who in 1992 express protectionist views have moved their party-identification toward the GOP by 1994. We show these effects are robust to flexibly controlling for a variety of demographic variables as well as views on other political and policy questions...This shift among protectionist respondents is larger for whites (especially men and those without a college degree) and those with conservative social views, suggesting an interactive effect whereby racial identity and social-issue positions mediate reactions to economic policies…"

Jiwon Choi, Ilyana Kuziemko, Ebonya Washington and Gavin Wright, “Local Economic and Political Effects of Trade Deals: Evidence from NAFTA,” National Bureau Of Economic Research, November 2021, https://www.nber.org/papers/w29525

“…We show the results of estimating the event-study equation (5) in Figure 13. The first series shows the coefficient estimates from an equation with no controls (so the vector Xi is omitted), so the plotted points are just raw differences. We see that pre-NAFTA, those with protectionist views were less supportive of the GOP. Sometime between 1992 and 1996 (the protectionist question is not asked in 1994) a significant number of protectionist voters moved toward the GOP, so that the raw difference disappears. In Appendix Figure?? we show this result as two separate series instead of as a difference, and we also include 2016, which shows that in that year there was yet again another movement of protectionist voters toward the GOP, consistent with Trump's anti-trade message. In the second series, we add standard demographic and socioeconomic controls: race, sex, age, education, and income. The same 1992-1996 shift toward the GOP among protectionist voters remains, even after allowing these characteristics to have their own effect each year. In the final series, we add controls for trust in the federal government, views (a thermometer" going from cold to warm) toward African Americans (given the importance of race in U.S. politics), views on abortion, and weekly religious attendance. We are limited somewhat by wanting to use questions that are asked each or most years in our sample. For example, questions on gay rights and regulations on _rearms purchases are missing in our pre-period (we are able to add a richer set of controls in our panel-data analysis in the next subsection). As the final series of the figure shows, the large shift between 1992 and 1996 remains even after flexibility controlling for views toward the other salient issues available over our sample period….”

“…Figure 12 makes clear that, throughout the 1980s, NAFTA-vulnerable counties exhibited a steady Democratic lean in House elections, but beginning in 1994 shift in the Republican direction….”

Evidence, “…Our first set of political results takes the same county-level event-study approach we have used for our local-economic analysis, except that our dependent variable occurs every other year, as we focus on House election outcomes. Because of the important role of NAFTA in the 2016 election, we extend our post-period to 2016, though readers are free to discard evidence this far from the original 1994 event. Except for the extension of the sample period and our observing the dependent variable biennially instead of annually, the analysis remains identical to those in the previous sections. Figure 12 shows the House election outcomes using our standard event-study specifications. Unlike the employment and other economic outcomes, which had a clear break beginning in 1994 (the year of NAFTA implementation), our baseline specification shows a break beginning in the 1992 election. Given the large role that the debate over whether to ratify NAFTA played in the 1992 election, it is not clear that 1994 as opposed to 1992 is the most legitimate start of the post-period" for the political analysis. Adding state-year fixed effects generally does not change the event-study coefficients, though now 1994 appears more like the break in trend than does 1992. It appears that within states, NAFTA-exposed counties break from their pre-period trend in 1994, even if the break occurs in 1992 when both within- and between-state variation in NAFTA exposure is considered. As we have found throughout the previous analysis, adding the ADH CZ-level measure interactions with year fixed effects (the second series in Figure 12) makes no appreciable difference to the event-study coefficients. Nor does 1990 county-level manufacturing interacted with each year. Again, as with previous outcomes, the set of covariates that most attenuates the implied effects of NAFTA are the 1990 college share interacted with year fixed effects. But even with this specification, we find a visually compelling shift in the two-party Democratic share of vote in House races. In this specification, there is a at pre-trend with no coefficient’s confidence interval exclude zero, but in the post-period an average coefficient value of roughly -0.75 (compared to the other specifications, where the post-period coefficient values are more like -1.5). What do these magnitudes suggest in terms of actual political outcomes? Taking our preferred specification (the second series, with state_year _fixed effects) the average of the coefficients in the late 1990s and early 2000s is roughly 0.60, suggesting that the most exposed quartile increase its share of GOP House votes by roughly 4.8 percentage points relative to the least exposed. As Table?? shows, the baseline for this group was only 0.38, so the 4.8 percentage point swing is very large. Appendix Figure A.12, which presents the data in completely raw form, shows that by 1998, these once-solidly Democratic counties voted as or more Republican in House elections as did any other county quartile…”

The political response in areas vulnerable to NAFTA

Evidence, “…Figure 4 shows our four groups of counties, with data unadjusted except for each group's (log) employment level being normalized to zero in 1993. While in the event-study analysis we weight by 1990 county population (both to match past work as well as to better match to the political results, where counties are essentially weighted by population of voters) in this raw analysis we do not. The four series are nearly on top of one another in the years leading up to NAFTA. But they begin to diverge thereafter. By 2000, the most-exposed quartile has roughly five to ten log-points less than other counties relative to the 1994 baseline. Note also that the second-most-exposed quartile also shows a slowing of employment growth relative to the least-exposed half of counties, suggesting that the effect is not entirely driven by a few counties in the most-exposed quartile… As shown in Table A, counties that would be more exposed to NAFTA were already different on important dimensions in 1990: for example, they had higher reliance on manufacturing employment and lower rates of college-degree completion. The fourth series in Figure 5 adds 1990 manufacturing share of county employment interacted with year fixed effects, which barely moved the coefficients. The final series is the same, but instead adds 1990 share of adults with a college degree interacted with year fixed effects. Of all the controls we add, this one has the most appreciable effect. Nonetheless, the effects are large, negative and highly significant..."

China). While growing before NAFTA, Mexican imports enjoy more rapid growth beginning in 1994. Interestingly, despite the larger focus on China in the empirical labor economics literature, it is not until 2004 that China supplants Mexico as the most important low income source of imports (though its rise since 2004 is indeed more rapid than any period for Mexico)….”

Good factoid, “…Figure 1 shows the value of imports to the US from Mexico (and, for context, also includes