Macro Roundup

FRBNY uses a natural experiment of a UI duration cut in Missouri (2011) and finds the cut significantly increased job finding rates through two mechanism: 1) an uptick in the search effort of unemployed workers and 2) availability of jobs (labor market tightness).

Workers reacted to the cut in UI durations by raising their search effort. FRBNY found the numbers imply a sizable increase in the job finding rate of 11.1%; the elasticity of the job finding rate with respect to UI durations was around -0.45. Under their most conservative estimate of the increase in the job finding rate they concluded that at least 1/3 is due to market-level effects through tightness and ~ 2/3 part is due to higher search effort.

These two effects accounts for 55% of the full percentage point decline in the unemployment rate caused by the policy change in Missouri.

They also note that"...over this episode, labor force participation fell in Missouri and applications for disability insurance spiked..." reinforcing Casey Mulligan point that disability had become a substantiate for UI as workers aged out of it though they don’t explore this angle in the paper.

....Following the policy change, the vacancy filling rate dropped immediately by 5% relative to the synthetic control—signifying a tightening of the labor market in Missouri, as the vacancy-filling rate is inversely related to labor market tightness....The tightening of the labor market and the increase in search effort led to a higher job finding rate in Missouri. We find that this increase induced by the policy change was 11.1%. Under our most conservative estimate of the increase in the job finding rate, we conclude that at least one third is due to market-level effects through tightness and the remaining part is due to higher search effort...We find that about 50% of the decline in unemployment following this policy change is attributable to higher exits from unemployment. Importantly, we estimate market level effects to be sizable—about 36% of the total effect according to our most conservative estimate and up to 100%...."

Fatih Karahan, Kurt Mitman and Brendan Moore, "Individual and Market-Level Effects of UI Policies: Evidence from Missouri," Federal Reserve Bank of New York, December 2019, https://www.newyorkfed.org/medialibrary/media/research/staff_reports/sr905.pdf

New NBER suggests that the opioid epidemic was not driven by the economic fallout from the decline in mining but driven by health shocks from the work in the mines themselves.

Results imply the decline in mining will have a positive spillover in terms of a reduction in mortality from opioid use. A result so intuitive it’s almost counterintuitive.

"...The results suggest that ahigher share of coal miners among the local total workforce is associated with a higher opioid death rate. This conclusion is the opposite of what mainstream media would claim......Our results indicate thata lower dependency of the local economy on coal mining is associated with lower opioid death rates.Specifically, a one percent increase among the coal-producing counties in the share of coal miners in the workforce increases the opioid mortality by 0.192 percent.To better understand the magnitude of the impact,if coal mining activity in 2016 were at its 2011 peak level, the opioid death rate would have been fifteen percent higher.The impact is stronger among underground than surface coal mines. … Based on these results, we argue that the shift from coal to natural gas not only has produced environmental benefits but also has helped to blunt damages from the opioid epidemic.... Results are strongest for underground coal mining activity and for miners in the 20 to 39 age group ….”
Gilbert E. Metcalf and Qitong Wang, "Abandoned by Coal, Swallowed by Opioids?" National Bureau of Economic Research, December 2019, https://www.nber.org/papers/w26551

Marcia,

This crossed my desk this morning. Given our conversation on Friday, I thought you might find it interesting.

Summary:

College graduates … earn 70% more over their lifetime than their peers who never entered college. However, this earnings gap grossly exaggerates financial gains from a college degree. This is because student bodies can differ vastly across colleges with regard to learning ability—a term that refers collectively to all student characteristics at the time of high school graduation that matter for both their academic success and labor market success (e.g., college preparedness, work ethic, grit, ambition).

If we make an adjustment for student composition effects using test scores, we find that Type 4 graduates earned only 5% more per year relative to their peers graduating from Type 2 colleges. The younger cohort, however, experienced a 10% gain—a full doubling of the return to college quality.

A large part of the rising importance of college quality can be attributed to rising sorting on learning ability. A choice of a high-quality college has become a better indicator of higher student ability and would manifest as rising returns to college quality.

Consistent with the explanations above, we found evidence of increased student sorting. With less room to increase operating capacity, four-year schools increased their admission standards, and lower-ability students became more likely to enroll in community colleges.

Oksana Leukhina and Joseph McGillicuddy, "What’s behind Rising Returns to High-Quality College Education?," Federal Reserve Bank of St. Louis, October 25, 2019, https://www.stlouisfed.org/publications/regional-economist/third-quarter-2019/rising-returns-high-quality-college-education

1. Seewww.bloomberg.com/news/articles/2019-03-12/college-arms-race-goes-too-far-with-wealthy-gaming-the-system.
2. This research was conducted with restricted access to Bureau of Labor Statistics (BLS) data. The views expressed here do not necessarily reflect the views of the BLS.
3. This standardized test was not a college entrance exam like the SAT or ACT.
4. Seehttps://nces.ed.gov/ipedsfor more information.
5. Based on authors’ calculations.
6. See Bowen et al.
7. See, for example, https://nces.ed.gov/ipeds/use-the-datafor college characteristics orwww.payscale.com/college-salary-reportfor salary information.
8. This ability should not be seen as merely the result of innate intelligence or academic potential. Learning ability can be shaped by many factors, including disparities in economic and educational opportunities, all of which are outside the scope of our analysis.
9. See Hendricks and Leukhina.
10. See also Dillon and Smith.
11. Based on authors’ calculations and consistent with Philippe Belley and Lance Lochner, “The Changing Role of Family Income and Ability in Determining Educational Achievement.” Journal of Human Capital, Winter 2007, Vol. 1, No. 1,pp. 37-89.
12. In 2000 dollars. Because tuition is not reported in NLSY79, we obtained tuition estimates for the older cohort from the High School and Beyond survey, conducted by the National Center for Education Statistics.

Oksana Leukhina is a senior economist at the Federal Reserve Bank of St. Louis. Her research interests include growth, labor and demographic economics. She joined the St. Louis Fed in 2017. Read more about the author and her research.

Joseph McGillicuddy is a senior research associate at the Federal Reserve Bank of St. Louis.

References

American Universities and Colleges. Westport, Conn.: Praeger Publishers, 1983.

Barron’s Profiles of American Colleges: Descriptions of the Colleges. 1980-1992. Barron’s Educational Series Inc.

Bowen, William G.; Chingos, Matthew M.; and McPherson, Michael S. Crossing the Finish Line: Completing College at America’s Public Universities. Princeton, N.J.: Princeton University Press, 2009.

Dillon, Eleanor; and Smith, Jeffrey. “The Consequences of Academic Match between Students and Colleges.” Journal of Human Resources, (forthcoming).

Hendricks, Lutz; and Leukhina, Oksana. “The Return to College: Selection and Dropout Risk.” International Economic Review, August 2018, Vol. 59, No. 3,pp. 1077-1102.

ABOUT THE AUTHORS

What’s behind Rising Returns to High-Quality College Education?

While financial returns to a college degree have been widely studied, less is known about how much it matters where one goes to college. Yet, more than ever, college quality is perceived to be a decisive factor in determining a child’s future success. This was the motivation behind the recent “Varsity Blues” scandal, in which affluent parents were accused of bribing college officials to guarantee their children’s acceptance into top schools.

But many more parents pour significant resources into improving their children’s college prospects without resorting to illegal means. For example, private SAT tutors can cost hundreds of dollars an hour, and college application consultants charge tens of thousands of dollars for their services.1 One may wonder whether financial gains associated with entry into highly selective schools justify the higher tuition and college prep costs.

What is the financial gain from choosing a top-tier academic schoolover a lower-ranking college closer to home or even a community college? How has it evolved over time? These are important questions for policymakers and for high school students facing their most important financial investment decision. We investigate these questions.

Data Sources on Student Outcomes and College Quality

As our data sources, we employed the 1997 and the 1979 National Longitudinal Surveys of Youth (henceforth, NLSY97 and NLSY79).2 NLSY97 is an ongoing survey that tracks the lives of 8,984 millennials, many of whom entered college around 2000. The NLSY79 follows an older cohort that comprises 12,686 baby boomers, many of whom entered college around 1980.

In each survey round, the individuals answer questions on a variety of topics, including education and income. These two are the only surveys that contain complete earnings histories for at least 15 years following college graduation and allow us to identify colleges that students attended and degrees received. All surveyparticipants were also administered an aptitude test that covered numerical operations, vocabulary, paragraph comprehension and logical reasoning.3

To rank these colleges on “quality,” we compiled a comprehensive data set of over 3,000 colleges and universities in the U.S. and collected information on their average SAT scores and freshmen enrollment in 2000. The main source for this information is the Integrated Postsecondary Education Data System.4 For colleges with missing reports, we used average SAT scores published in Barron’s Profiles of American Colleges and American Universities and Colleges.

We categorized all colleges into four types. The lowest type (Type 1) comprises community colleges offering a transferable associate degree. Four-year institutions are ranked in terms of their freshmen’s average SAT score, from lowest to highest, and they are split into three groups based on freshman enrollment. Type 2 comprises the lowest-ranked colleges that account for a third of all freshmen; Type 3 comprises the middle-ranked colleges and Type 4 represents the top-ranked colleges, each with a third of enrolled freshmen.

We will refer to higher-type colleges as higher-quality colleges because better SAT averages not only indicate better learning (and networking) opportunities from one’s peers but also strongly correlate with measures of instructional quality (e.g., faculty-student ratios and faculty salaries).5 We chose to include community colleges in our analysis because over a third of college entrants start in a community college, with 95% of them stating their ultimate goal is a bachelor’s degree.6

According to our classification, higher-type colleges host a more strictly selected group of students, provide higher-quality instruction and cost more. We can now identify the quality type of each college attended by each survey participant. We use responses from both surveys to analyze how postsecondary education decisions and their effects on future earnings have changed over time.

Pitfalls in Assessing Financial Gains from a Specific College

A great deal of college information is available to today’s high school students, such as college graduation rates and average post-graduation salaries.7 Students should take caution when using graduation rates and salary data to compare financial gains or losses from enrolling in different colleges. This is becausestudent bodies can differ vastly across colleges with regard to learning ability—a term that refers collectively to all student characteristics at the time of high school graduation that matter for both their academic success and labor market success (e.g., college preparedness, work ethic, grit, ambition).8

Such student composition differences partly account for graduation outcomes and salary differences across colleges and must be adjusted for when inferring personal gains from selecting one school over another.

If the highly selected Harvey Mudd College students, nearly all of whom major in science, technology, engineering and mathematics (STEM) fields, report earning $89,000 shortly upon graduation, while the graduates of the less selective University of California, Irvine earn $60,000, a student admitted to both schools who plans to study engineering should not conclude that choosing UC Irvine over Harvey Mudd would entail giving up a third of post-graduation salary.

Likewise, she should not interpret the average graduation rates (92% for Harvey Mudd and 88% for UC Irvine) to mean that her graduation probability would decline by 4 percentage points. The higher graduation rates and earnings of the Harvey Mudd students partly reflect the student composition effect—they are (on average) higher achievers who do not shy away from the more lucrative STEM fields.

To more accurately assess the expected salary associated with choosing UC Irvine over Harvey Mudd, the student in our example should compare graduation rates and average post-graduation salary for a very specific set of students—engineering majors with similar to her own high school performance, family income and parental education. Doing so would largely reduce the expected salary loss suggested by the naive calculation.

Effects of College Quality: The NLSY79 Cohort

Let’s start with a few facts about this cohort. About half of the NLSY79 high school graduates enrolled in college and about half of those who enrolled earned a bachelor’s degree. College graduates went on to earn 70% more over their lifetime than their peers who never entered college. However, this earnings gap grossly exaggerates financial gains from a college degree. After appropriate adjustments are made for student composition differences, the average gain from graduation measures at 35%.9

For the purpose of our analysis of college quality effects, we approximated student learning ability (as defined earlier) by aptitude test score percentile. This measure is practical but imprecise, and we will discuss the implications of this imprecision later. We adjusted for student composition effects by comparing outcomes across colleges for students with similar test scores.

Consistent with the intuition from the simple example discussed above, we found thatthe effects of high-quality colleges are exaggerated if one fails to adjust for student composition differences.

At first glance, school quality appears to have a massive influence on the probability that a student graduates from college: Only 20% of the older cohort starting out at a community college (Type 1) obtained a bachelor’s degree within six years, while the top (Type 4) schools saw 86% of their freshmen graduate, suggesting that a choice of a Type 4 college over a community college path increases graduation probability by 66 percentage points.

However, this simple analysis exaggerates the causal effect that college quality has on graduation rates. Part of the reason that the graduation rate is higher at better schools is because their students are, on average, of higher learning ability (as defined earlier)—Type 1 students scored in the 54th percentile of all high school graduates, while Type 4 students scored in the 83rd percentile.

If we compare graduation rates for students with similar aptitude test performance—thereby making an adjustment for student composition effects—we find that, on average, a student’s probability of graduating increases by 50 percentage points when he or she chooses a Type 4 college path over a community college path to a bachelor’s degree. Even though the effect is significantly reduced, it remains very large.

A similar narrative emerges when we look at how college quality affects earnings after graduation. After adjusting for work experience, we found that students who graduated from a Type 4 school earned 8% more per year than those who graduated from a Type 2 school.

However,if we make an adjustment for student composition effects using test scores, we find that Type 4 graduates earned only 5% more per year relative to their peers graduating from Type 2 colleges. As in the case of graduation outcomes, we find that college quality effects on post-graduation earnings are significantly reduced after making this adjustment, and these effects are fairly small. Note that college quality mattered greatly for graduation outcomes. Where the degree came from mattered less for earnings outcomes.

Aptitude test scores also held a strong predictive power on both graduation rates and post-graduation earnings. When comparing students entering the same type of college, we found that students with test scores in the top quartile (fourth) among high school graduates attained a bachelor’s degree at a rate that was 47 percentage points higher and earned 25% more when compared with students scoring in the bottom quartile (first).

How Did the Effects of College Quality Change over Time?

What can we say about the younger cohort who entered college in 2000—the cohort that faced greater returns to a college degree and showed greater enrollment rates? We found that the importance of college quality as a determinant of graduation rates and post-college earnings increased, while the importance of aptitude test scores declined.10

Figure 1 helps visualize the increasing importance of college quality for graduation outcomes. The left-hand panel of Figure 1 plots graduation rates by college quality for freshmen scoring in the lowest quartile of the test score distribution. The positive relationship between college quality and graduation rates is steeper for the younger cohort—mainly because students with low test scores performed better at higher-quality schools.

The right-hand panel tabulates graduation rates for students in the top quartile of the test score distribution. Again, the positive relationship between college quality and graduation rates is steeper for the younger of the two cohorts—mainly because high-scoring students performed worse at lower-quality schools.

The same trends appear in post-graduation earnings. As discussed earlier, the older-cohort students graduating from a Type 4 school went on to earn 5% more than their peers who graduated from Type 2 schools. The younger cohort, however, experienced a 10% gain—a full doubling of the return to college quality.

Figure 2 helps visualize the declining importance of aptitude test scores for graduation outcomes. It plots graduation rates against the test score quartile, for students attending the same college type. The left-hand panel shows graduation rates for students who entered community colleges, and the right-hand panel shows graduation rates for students who entered the highest type colleges.

Thus, the younger cohort saw a decline in the importance of test score quartile as a determinant of graduation and earnings.

Let’s consider the average results for all students in a given cohort. Whereas the high-scoring students in the older cohort attained a bachelor’s degree at a rate that was 47 percentage points higher than that of their low-scoring peers, this gap dropped to 26 percentage points for the younger cohort—a decline by nearly a factor of 2. Similarly, the younger cohort saw the importance of test score quartile as a determinant of earnings reduced by a factor of 2.

What Explains the Rising Importance of College Quality?

We argue that it is highly likely thata large part of the rising importance of college quality can be attributed to rising sorting on learning ability. If the relative learning ability of students enrolled in top-quality colleges increased, this change would manifest itself in our analysis as increasing returns to college quality, in both graduation outcomes and earnings.

Let’s try to unpack this statement. In a 2018 paper, Lutz Hendricks and Oksana Leukhina found that aptitude test scores are fairly imprecise measures of student abilities. This means that by comparing outcomes across college types for students with similar test scores, as we did here, we are not comparing students with the same learning ability.

It is likely the case that low-scoring students who were accepted and enrolled in higher-quality schools are of higher learning ability than what their test scores suggest—they may have simply had a bad day when they took their tests. Thus, even though we made adjustments for student composition effects in terms of test scores, our measure of returns to college quality partly reflects the remaining student composition effects in terms of learning ability.

We will argue that this part has become more important over time as the ability gap between top- and bottom-quality colleges has widened. Let’s refer to the size of this gap as the strength of student sorting across college types. We will first explain what factors could potentially drive the increase in student sorting on ability and conclude by showing some evidence of this phenomenon.

Potential Explanations for the Rise in Student Sorting

First, the younger cohort faced a significantly greater college graduation premium, which resulted in the rise of college enrollment (from 46% of high school graduates to 57%) characterized by an influx of students with lower test scores from families of medium to higher incomes.11 As a result,the variation in learning ability among college entrants increased. With less room to increase operating capacity, four-year schools increased their admission standards, and lower-ability students became more likely to enroll in community collegeswith open-door admission policies.

Second, tightening financial constraints may have also played a role in increasing sorting. We documented that the average tuition paid by freshmen increased by a factor of 3 for the younger cohort, from $4,000 to $13,552.12 Even though the amount of grants and scholarships also increased,the direct cost of college more than doubled, from $3,628 to $8,624. (See Endnote 12 for details.)

In the meantime, the increase in the borrowing limits for subsidized Stafford loans did not keep up with the rising college costs. The borrowing limit for the first year of college barely changed at all, increasing from $2,500 in 1982 to $2,625 in 2000, although the annual loan limit increased more for subsequent years in college.

As a result, the younger cohort had to rely more heavily on family transfers, work more hours while in college and enroll in less-expensive colleges. Such downgrading of college choice would disproportionately affect the less-college-ready students as they tend to come from lower-income families. In addition, low-ability students with limited financial resources would be more likely to downgrade their college choice when faced with less time left to study, as it would be more difficult for them to keep up with school work in top colleges.

Finally,it is possible that the true return to a degree from higher-quality schools increased, perhaps as a result of increased expenditures on faculty, advising and career services. As a result, we would expect to see upgrading of college choice predominantly among the high-ability students—as these students can more easily keep up with stricter learning standards.

All of these factors have likely contributed to the widening in the learning ability gap between top- and bottom-quality institutions. As a result, the performance of students attending top-quality schools rises in terms of both graduation rates and earnings relative to the performance of students at bottom-quality schools. Put differently, a choice of a high-quality college has become a better indicator of higher student ability and would manifest as rising returns to college quality.

Evidence of Rising Student Sorting

Consistent with the explanations above, we found evidence of increased student sorting.Figure 3 compares college sorting between our two cohorts. The left-hand panel shows the fraction of lowest-scoring students who began college at each type of school. In both groups, a majority of students started at community colleges, but that share increased from just over a half of the older cohort to nearly three-quarters of the younger cohort.

While most of the action seems to concentrate in the rise of the community college track among lowest-scoring students, we also saw a substantial increase in sorting among four-year schools. Figure 4 plots the average freshmen SAT score percentile by college quality (defined by the decile of their rank, using the methodology in the section titled “Data Sources on Student Outcomes and College Quality”) for all four-year schools. We clearly see the relative decline of student ability in lower-quality schools.

We showed that the effects of high-quality colleges are exaggerated if one fails to adjust for student composition differences. When making the adjustment, the effects appear smaller and work mainly through increasing one’s graduation chances—where one’s degree comes from matters less for post-college earnings.

However, we also demonstrated that returns to college quality increased over time and argued that a large part of this increase is explained by the widening learning ability gap between students attending low-quality schools and students attending high-quality schools. We put forward several plausible explanations for this phenomenon, such as tightening financial constraints and the rising graduation premium. More work is needed to disentangle the role of these factors.

Endnotes

Mene Ukueberuwa, "American Innovation, a Murder Mystery,"Wall Street Journal, December 13, 2019, https://www.wsj.com/articles/american-innovation-a-murder-mystery-11576277193American Innovation, a Murder Mystery Addressing 20,000 rallygoers at Dallas’s American Airlines Center recently, President Trump praised a “beautiful new Louis Vuitton plant” he’d toured that morning. The plant will bring about 1,000 jobs to a small town near Fort Worth, and Mr. Trump was displaying his good habit of celebrating work wherever Americans find it. But despite cheers of approval, one imagines he and his supporters would have preferred to christen an advanced “machine and manpower” plant—the kind that might assemble airplanes for the arena’s namesake. A craft workshop for a French luxury brand doesn’t quite capture the MAGA ideal. The president’s pledge to make America great again has always borne the implicit claim that America no longer makes anything great, whether Fords or farm equipment. There’s some truth to that, and both airplanes and handbags are examples of how U.S. manufacturing has fared since the mid-20th century. The key measure is productivity—how much value each worker produces—which bears heavily on job creation. Highly productive industries like chemicals, weapons systems and aerospace still draw strong domestic investment. But employment has shrunk because better machines and logistics have diminished the need for manpower. Productivity has risen more slowly for simpler types of manufacturing such as textiles and home goods, leading many companies to seek cheaper production overseas. There has also been a notable slowdown in improvements to the quality of goods produced. For about a century, beginning with the Second Industrial Revolution in the 1870s, American growth was driven by tinkerers, trusts and corporate labs, which eked out regular advances in products based on engines, electrical circuits and signals, and synthetic materials. Today, another half-century later, a coast-to-coast flight still takes you as long as it took your father in the 1970s. And with the major exception of computers, nothing in your luggage is likely to be much more useful or valuable than dad’s equivalent. Who killed American innovation? Answering that question could be a vital step toward re-energizing growth. Unfortunately, in their search for a culprit, Mr. Trump and many policy makers and commentators have implicated causes that aren’t responsible for the slowdown and tried solutions that hinder growth: • The free traders. Since the beginning of his 2016 campaign, Mr. Trump has often blamed the decline of heavy industry on his predecessors’ trade policies. “I did disagree with Ronald Reagan very strongly on trade,” he said in his final debate with Hillary Clinton. As president he has blasted the North American Free Trade Agreement—a signature achievement of Mrs. Clinton’s other half—as often as any federal policy. Describing the effect Mexico and other trade partners have had on domestic production, Mr. Trump said last year that “they’ve destroyed the steel industry, they’ve destroyed the aluminum industry and other industries, frankly.” In this view, politicians opened U.S. manufacturing to an assault from foreign competitors that produce and sell more cheaply. That reduces American manufacturers’ incentive to invest in better plant and products, benefiting rivals in protected economies. Yet open trade can’t explain the overall slowdown of U.S. innovation. To be sure, competition from Mexico reduces the incentive for U.S. Steel to invest in new fabrication techniques. But progress has also slowed since the mid-20th century for American companies that are heavily supported and protected, like Boeing. And there are countless industries in which U.S. companies retain a large lead in innovation despite trailing in global market share—most notably, cutting-edge electronics like smartphones and semiconductors. U.S. protectionism couldn’t have prevented other nations from eventually catching up. Georgia Tech economic historian Steven Usselman notes that American firms throughout most of the 20th century expected this “convergence” of international capability to occur decades earlier than it did. Yet even as manufacturing executives lobbied for support in their sectors, they backed general liberalization of trade as a net economic good. • The investors. It’s common to blame stagnant innovation on large companies that reduce productive investment in favor of shorter paths toward profit. A growing number of Republicans have adopted this critique, notably Florida Sen. Marco Rubio. In May Mr. Rubio issued a 40-page report lamenting a drop in research and innovation and blaming “shareholder primacy theory” that focuses on increasing returns to shareholders. Instead of investing in plant and products, he argues, executives focus on cutting costs and optimizing debt to boost quarterly earnings. Critics of shareholder capitalism point to charts showing that plant and equipment have shrunk as a share of corporate assets since the 1950s, and that R&D in manufacturing have declined. But it’s unlikely these trends were caused by a sudden aversion to long-term commitments among investors and managers. Consider that in place of hard goods, companies are investing more in intangible assets, like the patents gained by acquiring a startup, or enterprise software that simplifies communication and accounting. These expenses may not boost employment in the same way as, say, new delivery trucks. But they reflect as much willingness to spend in ways that don’t deliver returns immediately, and in some cases never do. More important, there are entire business models based on risky long-term strategies—and those companies don’t have trouble finding backers. American financiers overinvested in rare-earth mining and hydraulic oil and gas production a decade ago based on rosy projections of global growth. And no one could accuse WeWork’s owners of valuing quarterly earnings over long-term potential. The rise of venture capital means more investors than ever are seeking the riskiest bets, and not only in software. Lavish support is available in biotechnology, robotics and other types of engineering in which backers have faith in an eventual payoff. • The regulators. This is the favorite suspect of innovators themselves, who understandably bristle under the increasing burden of federal rules. A company that makes or moves physical things faces limits on land use, passenger safety, pollution, chemical content and more, nearly all of which have gotten stricter since the 1970s. Peter Thiel, the billionaire tech investor, condemns federal regulators as the greatest obstacle to American innovation. “I would say that we’ve lived in a world in which bits were unregulated and atoms were regulated,” he said in a 2015 interview. His point is that restrictive rules are responsible for the slowdown in the engineering of mechanics (atoms) relative to software (bits). Whatever the benefits, every mandatory trial, emissions cap and maintenance rule reduces manufacturers’ output, decreases profit and restrains the ability to experiment. Yet the international comparison again casts doubt on this explanation. Countries that have developed quickly since the innovation slowdown, such as China, Japan and South Korea, placed less emphasis on consumer and worker protections than the U.S. did. Their bureaucrats attempt to boost industrial output, rather than hinder it. Despite decades of infrastructure marvels and greater shares of the world manufacturing market, none of these countries have caught up to the U.S. in achieving breakthroughs. Unlike 20th-century America, most of their growth has come from technologies discovered elsewhere. If regulation were decisive in hampering technological progress, one would expect zealously permissive regimes to surpass the fussy West. With legislators, investors and regulators largely off the hook, what’s left? The best evidence suggests that the seemingly boundless American progress of last century died a natural death. After 100 years of success, engineers around the 1970s found themselves less able to develop ever more productive designs and applications for the core technologies of the Second Industrial Revolution: combustion, electricity, signals, synthetics. As with previous revolutionary technologies—the steam engine, steel, gunpowder—the returns from each discovery gradually diminished. Productivity growth has slowed as a result. No other explanation accounts for why the slowdown has affected nearly every field of manufacturing. The internationally competitive garment industry and the protected sneaker industry both use similar production methods to those of 50 years ago. Neither high-tech combine harvesters nor simple winter jackets fulfill their functions much better now than two generations ago. No other country, regardless of trade or industrial policy, has demonstrated any more ability to break through. This realization is vital because it reframes every development in the U.S. economy since engineering progress diminished. For manufacturers since the 1970s, with few opportunities to grow by investing in better plant and equipment, offshoring parts of the production process may have become their only way to compete, not merely a way to grow faster. When companies lost much of their ability to improve product quality, lower production costs let them cut prices instead, raising living standards by other means. Critics like Mr. Rubio argue that the “financialization” of the economy has been counterproductive. In truth, the financial sector’s keenness for viable bets has helped sustain the overall growth of U.S. manufacturing. The low rate of nonperforming loans suggests the U.S. isn’t overleveraged compared with peers that boast a higher manufacturing share of output. Most important, the rise of software and high-end electronics looks less like a distraction from productive growth and more like a savior that has kept America moving when no other industry could. From IBM to Snapchat, tech companies have produced countless exportable goods and services and brought massive sums of foreign dollars into the U.S. economy. The software revolution has also propped up manufacturing by improving design capability, supply-chain logistics and marketing. A 2016 Deloitte survey of industry executives world-wide predicts that the U.S. will overtake China as the most competitive destination for manufacturing, as an ever-increasing share of growth occurs at the high end. Recognizing that the 20th century’s innovation streak died naturally should encourage policy makers to resist the allure of a Frankenstein industrial policy: slapping tariffs, subsidies and incentives on industries like electrodes, trying to revive the circumstances of a bygone economy. Protectionism can help companies maintain or increase market share. But manufacturing is shrinking as a share of employment in developed and developing economies alike. There’s little evidence that state support fosters commercially viable breakthroughs better than the market, so it’s an inefficient strategy for an advanced economy. Incentives to redirect capital, like Mr. Rubio’s proposal to punish stock buybacks through the tax code, would cause waste by discouraging companies from saving for better investments later. There’s no telling how innovation in engineering may eventually spring back to life, whether a new style of software next year or a mind-bending new field of physics next decade. Meanwhile, policy makers should take care to preserve the economic freedom that helps businesses find ways to prosper absent revolutionary progress. And Mr. Trump should heed his better angels by continuing to support every type of great thing America makes.

WSJ reports how Germany’s elevated savings rate compared with her peers has left her households poorer due to anemic investment returns/negative interest rates (and flooded the world with risk adverse savings):

“….In a world of negative interest rates, stockpiling cash has left German households poorer than their international peers in terms of wealth, hindered German companies from investing in their future and restrained economic growth across Europe. Germany’s median household wealth is €61,000 ($67,000), lower than Greece’s, according to European Central Bank data. Germans are the least likely of all Europeans to own their own homes, an important source of wealth for many families, in part because they consider real estate risky…. households here continue to sock away 11% of their disposable income—compared with less than 7% in the U.S. and up from 9.3% in 2013, according to the Organization for Economic Cooperation and Development….. Crucially, Germans have been shunning those assets that have appreciated in the low-rate environment, like stocks and real estate, which are considered too risky. Only one in 10 households own stocks, while 70% have savings accounts, according to Bundesbank data. Some €140 billion in cash is being squirreled away in German mattresses or the like….”

Michael Pettis, “Excess German Savings, Not Thrift, Caused the European Crisis,” China Financial Markets, May 21, 2013, https://carnegieendowment.org/chinafinancialmarkets/51899

Pirmin Fessler, Peter Lindner and Esther Segalla, “Net Wealth Across The Euro Area: Why Household Structure Matters and How to Control For It,” European Central Bank, April 2014, https://www.ecb.europa.eu/pub/pdf/scpwps/ecbwp1663.pdf

“Household Net Worth,” Organisation for Economic Co-operation and Development, Accessed December 16, 2019, https://data.oecd.org/hha/household-net-worth.htm

“Wealth, Asset Ownership, & Debt of Household Detailed Tables: 2016,” United States Census Bureau, Accessed December 16, 2019, https://www.census.gov/data/tables/2016/demo/wealth/wealth-asset-ownership.html

Tom Fairless and Patricia Kowsmann, "Germans Keep on Saving Their Money—Even When It Hurts,"Wall Street Journal, December 3, 2019, https://www.wsj.com/articles/germans-keep-on-saving-their-moneyeven-when-it-hurts-11575388824

Our Census’s also has a “Median Value of Assets for Households, by Type of Asset Owned and Selected Characteristics: 2016” For US Whites that number is 114,700, overall it’s 94,670so the OECD comparison in perhaps more apt.

In answer to your question on what American Net Household Income looks like. There are few comparable number for the ECB’s “Median net wealth of households” Perhaps the best apples to apples of household net worth is theOECD series which has the US @ 183,648 and Germany at 58,455(Note that is close to the ECB #)

"....Germany is world champion in exporting capital ("Exportweltmeister"). No other country invests larger amounts of savings outside its borders. However, Germany plays in the third division when it comes to investment performance, as we show in this paper. We study the returns on German foreign investments from 1950 to 2017 and find that: (1) Germany's annual returns on foreign assets were 2 to 5 percentage points lower than those of comparable countries. Germany ranks last among the G7 countries, also in the last decade; (2) Domestic returns on German assets have outperformed foreign returns abroad by about 3 percentage points per year; (3) Germany's external wealth provides very little consumption insurance as foreign returns are highly correlated with domestic activity; (4) The capital exports do little to diversify demographic risks as Germany mainly invests in countries with similar demographics. Taken together, these facts raise substantial doubts whether German households, firms, and banks allocate their savings in a beneficial way...."

The WSJ also cites in passing research we’ve seen that shows German’s international investment returns are anemic:

“…This has happened not just in China but also in Germany. In the 1990s Germany could be described as saving too little. It often ran current account deficits during the decade, which means that the country imported capital to fund domestic investment. A country’s current account deficit is simply the difference between how much it invests and how much it saves, and Germans in the 1990s did not always save enough to fund local investment. But this changed in the first years of the last decade. An agreement among labor unions, businesses and the government to restrain wage growth in Germany (which dropped from 3.2 percent in the decade before 2000 to 1.1 percent in the decade after) caused the household income share of GDP to drop and, with it, the household consumption share. Because the relative decline in German household consumption powered a relative decline in overall German consumption, German saving rates automatically rose. Notice that German savings rate did not rise because German households decided that they should prepare for a difficult future in the eurozone by saving more. German household preferences had almost nothing to do with it. The German savings rate rose because policies aimed at restraining wage growth and generating employment at home reduced household consumption as a share of GDP…”

The WSJ seemingly attributes the high German savings rate to national character. However the article does not addressPettis’s point that Germany ran current account deficits during 1990’s until the Hartz reforms drove down the household share, driving up the savings rate which seems worth considering in the context of the article. The German savings rate is a policy choice: