What Were the Results of the Washing Machine Tariffs?

As reported by The Washington Post,

When economists at the University of Chicago and the Federal Reserve studied the 2018 duty on washing machines, they found the expected rise in retail prices from foreign manufacturers such as Samsung and LG. Surprisingly, though, these brands also increased dryer prices. Then domestic manufacturers followed suit, simply because they could.

All told, the research shows, U.S. consumers are spending an additional $1.5 billion a year on washers and dryers as a result of the tariffs. That’s an extra $86 for each washing machine and $92 for each dryer, the authors estimate. And less than 10 percent of that goes to the U.S. treasury — about $82.2 million — the study showed…Foreign manufacturers are passing some costs on to consumers, while domestic ones are simply pocketing extra profits, according to the study.

…Manufacturers also capitalized on buyer habits when they bumped up the price of dryers, which were not subject to the tariffs. “Many consumers buy these goods in a bundle,” Tintelnot said. “Part of the price increase for washers was hidden by increasing the price of dryers.”

In sum, “U.S. consumers shouldered 125 to 225 percent of the costs of the washing-machine tariffs. And the duty was mostly a dud on the job-creation front,” costing consumers about $815,000 for every one of the 1,800 jobs created.

That’s exciting. Looks like tariffs are exactly what they are cracked up to be.

Do Minimum Wage Hikes Drive Some Restaurants Out of Business?

Image result for restaurants minimum wage

From a recent NBER paper (quoting from an earlier draft): 

As theory would suggest, we find robust evidence that the impact of the minimum wage depends on how close a restaurant is to the margin of exit, proxied by its rating. Looking at city-level minimum wage changes in the San Francisco Bay Area (the “Bay Area”), we present two main findings. First, at all observed minimum wage levels, restaurants with lower ratings are more likely to exit, suggesting that they are less efficient in the economic sense. Moreover, lower rated restaurants are disproportionately affected by minimum wage increases. In other words, the impact of the minimum wage on exit is most pronounced among restaurants that are closer to the margin of exit. 

…Our results suggest that a $1 increase in the minimum wage leads to an 14 percent increase in the likelihood of exit for the median 3.5-star restaurant, but no impact for five-star restaurants (the point estimate is in fact negative, suggesting that the likelihood of exit might even decrease for five-star restaurants, but the estimate is not statistically different from zero). These effects are robust to a number of different specifications, including controlling for time-varying county characteristics that may influence both minimum wage policies and restaurant demand, city-specific time trends to account for preexisting trends, as well as county-year fixed effects to control for spatial heterogeneity in exit trends.

…Overall, our findings shed on the economic impact of the minimum wage. Basic theory predicts that the minimum wage will cause firms that cannot adjust in other ways to cover their increased costs to exit the market. We find that lower rated firms (which are already closer to the margin of exit) are disproportionately impacted by the minimum wage. After a minimum wage increase, they are more likely to exit the market altogether and more likely to raise their prices (pg. 2-5).

This matches previous research, which finds that labor-intensive restaurants tend to exit and make room for capital-intensive restaurants. 

Who Bears the Cost of the Minimum Wage?

Image result for minimum wage

From a forthcoming article in the American Economic Review (quoting from the draft version) on Hungarian minimum wage hikes:

Most firms responded to the minimum wage by raising wages instead of destroying jobs. Our estimates imply that out of 290 thousand minimum wage workers in Hungary, around 30 thousand (0.076% of aggregate employment) lost their job, while the remaining 260 thousand workers experienced a 60% increase in their wages. As a result, firms employing minimum wage workers experienced a large increase in their total labor cost that was mainly absorbed by higher output prices and higher total revenue. We also estimated that firms substituted labor with capital and their profits fell slightly. These results suggest that the incidence of the minimum wage fell mainly on consumers. Given the relatively small effect on employment, our results also suggest that minimum wages can redistribute income from consumers to low-wage workers without large efficiency losses. Our findings also indicate that the optimal level of the minimum wage is likely to vary across industries,cities and countries. In countries where low-wage jobs are concentrated in the local service sector (such as Germany or the U.S.) raising the minimum wage is likely to cause limited disemployment effects or efficiency losses. Moreover, in cities where mainly rich consumers enjoy the services provided by low wage workers this redistribution will be from rich to poor. The heterogenous responses across industries also underline the advantages of sector-specific minimum wage polices used in some European countries such as Italy or Austria. For instance, setting a higher minimum wage in the non-tradable sector than in the tradable sector can push up wages relatively more where it will generate more modest disemployment effects (pg. 23-24).

Passing the costs on to consumers fits with previous evidence. This also makes evident that the kind of industry (e.g., tradable vs. non-tradable) also matters when it comes to positive/negative effects of the minimum wage.

Stuff I Say at School – Part VIII: The Impact of Openness

This is part of the Stuff I Say at School series.

The Assignment

1. Do you feel that a country can thrive in an insular or isolated capacity? Is exchanges needed for a country to be successful? Do you see any examples of countries who have been reluctant to adopt new ideologies or integration?

2. What did we learn from the Columbian exchange that would be applicable to modern day society?

The Stuff I Said

1. While I think a country can thrive to some extent in isolation depending on a number of factors, it will not thrive as much as it could have had it been integrated into a larger exchange network. An extreme historical case is Tasmania: when the island was cut off from the mainland by rising sea levels, the population not only failed to progress, but actually regressed. Anthropologist Joseph Henrich surveyed the archaeological evidence and found that the isolation caused Tasmanians to lose a number of skills and technologies they had once possessed, including bone tools, cold-weather clothing, nets, fishing spears, barbed spears, etc. Even their canoeing skills and technologies worsened. Beyond comparative advantage, trade leads to innovation (what author Matt Ridley calls “ideas having sex”). And it is innovation–technological innovation in particular–that truly transforms standards of living. 

Protectionism and isolationism have had a bit of a global resurgence lately, but these positions fly in the face of the expert consensus as far as economic welfare is concerned (check out the survey data on tariffs at the bottom of the post). This populist backlash to globalization led to a string of recent academic books empirically and philosophically defending economic openness:

2. I’ll rely on Nobel laureate Angus Deaton for the next question:

The historian Ian Morris has described how increased trade around the second century CE merged previously separate disease pools that, since the beginning of agriculture, had evolved in the West, South Asia, and East Asia, “as if they were on different planets.” Catastrophic plagues broke out in China and in the eastern outposts of the Roman Empire. The Columbian exchange after 1492 is an even better-known example. Many historical epidemics started from new trade routes or new conquests.

…Yet globalization also opens its routes to the enemies of disease. We have already seen how the germ theory of disease–a set of ideas and practices developed in the North–spread rapidly to the rest of the world after 1945. Knowledge about drugs to control high blood pressure spread rapidly across the world after 1970, producing…synchronized declines in mortality…That cigarette smoking caused cancer did not have to be rediscovered country by country. While the origins of HIV/AIDS are in dispute, there is no dispute about its rapid spread from one continent to another. The scientific response–the discovery of the virus, the deduction of its means of transmission, and the development of chemotherapy that is transforming the disease from a fatal to a chronic condition–was extraordinarily rapid by historical standards, although hardly rapid enough for the millions who died as they waited. Today’s understanding of the disease, although still incomplete, has underpinned the response–not just in the rich world–and in the worst affected African countries rates of new infection have fallen in the past few years, and life expectancy is beginning to rise again (The Great Escape, pg. 150-151).

From Gregory Mankiw’s Principles of Economics, 7th ed. (pg. 32).

From the IGM Economic Experts Panel, University of Chicago

Stuff I Say at School – Part VII: The Importance of Institutions

This is part of the Stuff I Say at School series.

Summary & Commentary on Week’s Readings

Acemoglu et al argue that inefficient institutions persist for a number of major reasons. First, the lack of third-party enforcement of commitments prevents elites from relinquishing their monopoly on political power. Furthermore, the beneficiaries of the economic status quo are usually unwilling to risk their economic welfare through competition. This leads them to promote protectionism and further engage in rent-seeking activities. Institutions that encourage these kinds of activities fail to grow. We see this kind of conflict manifest in various areas of the economy, from labor and financial markets to regulations in pricing. The more institutions concentrate political power in the hands of the few, the more incentives are warped and distort paths to economic growth.

Image result for why nations fail

In their book Why Nations Fail: The Origins of Power, Prosperity, and Poverty, Daron Acemoglu and James Robinson distinguish between inclusive and extractive institutions, with the former creating the conditions for prosperity. “Inclusive economic institutions,” they write,

…are those that allow and encourage participants by the great mass of people in economic activities that make best use of their talents and skills and that enable individuals to make the choices they wish. To be inclusive, economic institutions must feature secure private property, an unbiased system of law, and a provision of public services that provides a level playing field in which people can exchange and contract; it also must permit the entry of new business and allow people to choose their careers…Inclusive economic institutions foster economic activity, productivity growth, and economic prosperity (pg. 74-75).

On the other hand, extractive economic institutions lack these properties and instead “extract incomes and wealth from one subset of society to benefit a different subset,” empowering the few at the expense of the many (pg. 76).

The importance of getting institutions right is highlighted by Rodrik and Subramanian’s study. Three theoretical culprits have been blamed for the vast income inequality between countries: (1) geography, (2) integration (globalization, international trade), and (3) institutions. Regression analyses indicate that institutions trump all other explanations. This is also shown from the outset of Acemoglu and Robinson’s Why Nations Fail, in their story of Nogales, Arizona (United States of America) and Nogales, Sonora, (Mexico). Acemoglu and Robinson lay out their archetype story of two towns with the same essential culture, geography, and relative free trade (NAFTA), in most ways they are the same place. The only reason they are two towns is an institutional barrier between two separate countries. Yet one is rich and one is poor because of institutions. The direct effects of geography are weak at best, while there were no direct effects from integration. However, there were indirect effects of integration: institutions have significant, positive effects on integration, while integration has a positive impact on institutions. This, in some sense, creates a virtuous, growth-enhancing cycle. Rodrik and Subramanian point out that the institutional factors emphasized the most have largely been market-oriented (e.g., property rights, enforceable contracts). Yet, factors such as regulation, financial stabilization, and social insurance also matter in getting institutions right.

The interaction between political and economic institutions is an important insight. For example, even though most research finds that seemingly liberal political institutions like democracy have no direct impact on economic growth, more recent evidence from Acemoglu and colleagues suggests that they may in fact contribute to growth. What’s more, the evidence strongly suggests that economic openness—particularly international trade—contributes to growth. A 2010 study used data from 131 developed and developing countries and found that reductions in trade protections led to higher levels of income per capita. A World Bank study found that between 1950 and 1998, “countries that liberalized their trade regimes experienced average annual growth rates that were about 1.5 percentage points higher than before liberalization. Postliberalization investment rates rose 1.5-2.0 percentage points, confirming past findings that liberalization fosters growth in part through its effect on physical capital accumulation…Trade-centered reforms thus have significant effects on economic growth within countries” (pg. 212). A 2016 IMF paper found that trade liberalization boosts productivity through increased competition and greater variety and quality of inputs. All this suggests that Sachs and Warner were correct when they found “that open policies together with other correlated policies were sufficient for growth in excess of 2 percent during 1970-89” (pg. 45; fn. 61). Their findings also suggest “that property rights, freedom, and safety from violence are additional determinants of growth” (pg. 50). Acemoglu and Robinson in a 2005 paper found “robust evidence that property rights institutions have a major influence on long-run economic growth, investment, and financial development, while contracting institutions appear to affect the form of financial intermediation but have a more limited impact on growth, investment, and the total amount of credit in the economy” (pg. 988).

In short, inclusive institutions are necessary to fully reap the benefits of an open economy.

Is Student Loan Forgiveness for the Marginalized?

Image may contain: text

I saw this floating around Facebook recently with the news of Elizabeth Warren’s student loan plan. For those unfamiliar with what Mayfield is referencing, here’s the entry from the HarperCollins Bible Dictionary:

As another Bible dictionary clarifies, “Though Leviticus 25 does not explicitly discuss debt cancellation, the return of an Israelite to his land plus the release of slaves implies the cancellation of debts that led to slavery or the loss of land.”

So does Warren’s plan benefit “the marginalized”?

According to Adam Looney at the Brookings Institution, Warren’s proposal is “regressive, expensive, and full of uncertainties…[T]he top 20 percent of households receive about 27 percent of all annual savings, and the top 40 percent about 66 percent. The bottom 20 percent of borrowers by income get only 4 percent of the savings. Borrowers with advanced degrees represent 27 percent of borrowers, but would claim 37 percent of the annual benefit.”

E Warren Distribution of benefit

He continues,

Debt relief for student loan borrowers, of course, only benefits those who have gone to college, and those who have gone to college generally fare much better in our economy than those who don’t. So any student-loan debt relief proposal needs first to confront a simple question: Why are those who went to college more deserving of aid than those who didn’t? More than 90 percent of children from the highest-income families have attended college by age 22 versus 35 percent from the lowest-income families. Workers with bachelor’s degrees earn about $500,000 more over the course of their careers than individuals with high school diplomas. That’s why about 50 percent of all student debt is owed by borrowers in the top quartile of the income distribution and only 10 percent owed by the bottom 25 percent. Indeed, the majority of all student debt is owed by borrowers with graduate degrees.

Drawing on 2016 data from the Federal Reserve’s Survey of Consumer Finances, Looney’s final analysis

shows that low-income borrowers save about $569 in annual payments under the proposal, compared to $900 in the top 10 percent and $2,653 in the 80th to 90th percentiles. Examining the distribution of benefits, top-quintile households receive about 27 percent of all annual savings, and the top 40 percent about 66 percent. The bottom 20 percent of borrowers by income get 4 percent of the savings…[W]hile households headed by individuals with advanced degrees represent only 27 percent of student borrowers, they would claim 37 percent of the annual savings. White-collar workers claim roughly half of all savings from the proposal. While the Survey of Consumer Finances does not publish detailed occupational classification data, the occupational group receiving the largest average (and total) amount of loan forgiveness is the category that includes lawyers, doctors, engineers, architects, managers, and executives.  Non-working borrowers are, by and large, already insured against having to make payments through income-based repayment or forbearances; most have already suspended their loan payments. While debt relief may improve their future finances or provide peace of mind, it doesn’t offer these borrowers much more relief than that available today.  

The Urban Institute’s analysis has similar findings (though their tone is more optimistic):

figure 2
figure 3

I’m not sure whether or not Warren’s plan is a good one (I’m skeptical, especially given some of the results abroad). But I’m not big on acting like college graduates in a rich country are the marginalized of society.

Does Good Management Produce More Equal Pay?

Nicholas Bloom–whose research on the economics of management I’ve relied on in my own work–and colleagues have an interesting article in Harvard Business Review:

For 2010 and 2015, the U.S. Census Bureau fielded the Management and Organizational Practices Survey (MOPS) in partnership with a research team of subject matter experts, including one of us (Nick), as well as Erik Brynjolfsson and John Van Reenen. The MOPS collects information on the use of management practices related to monitoring (collecting and analyzing data on how the business is performing), targets (setting tough, but achievable, short- and long-term goals), and incentives (rewarding high performers while training, reassigning, or dismissing low performers) at a representative sample of approximately 50,000 U.S. manufacturing plants per survey wave. We refer to practices that are more explicit, formal, frequent, or specific as “more structured practices.” From the MOPS and related data, researchers have demonstrated just how important the use of these structured management practices is for companies and even entire economies, since firms that implement more of these practices tend to perform better.  We wanted to know what effect these management practices have on workers.

We found that companies that reported more structured management practices according to the MOPS paid their employees more equally, as measured by the difference between pay for workers at the 90th (top) and 10th (bottom) percentiles within each firm.

The authors fully admit, “To be honest, it surprised us…If anything, we expected the opposite…We hypothesized that more structured management would lead to rewarding high-performers over others, therefore leading to a rise in inequality inside of the firm. As the chart above shows, the reality is exactly the reverse – and that remains true even after controlling for employment, capital usage, firm age, industry, state, and how educated the employees are.” They continue,

Our research finds that the negative correlation between structured management and inequality is driven by a strong negative correlation between the use of structured monitoring practices and inequality. By contrast, higher usage of structured incentives practices was positively correlated with inequality, albeit weakly. In other words, our finding seems to suggest that companies that collect and analyze specific and high-frequency data about their businesses tend to have a smaller gap between the earnings of workers at the top of the income distribution and the earnings of workers at the bottom of the distribution.

The authors offer several possible explanations:

Previous research shows that firms with more structured management practices are more profitable on average, and there’s long been evidence that when companies make extra profits they share some of them with workers. Perhaps companies with more structured practices allocate these profits such that less well-paid workers get more of the pie.

The relationship could also result from increased efficiency. Maybe firms with more structured practices have more efficient low-paid workers, as a response to training or monitoring practices, and their pay reflects that extra efficiency.

Finally, it could be that firms with more structured practices are more focused on specific tasks and rely more on outsourcing. More and more companies are outsourcing tasks like cleaning, catering, security, and transport. If outsourcing is more common for firms that use more structured practices, workers performing tasks outside of the companies’ core tasks would no longer be on those companies’ direct payrolls. If the jobs that are outsourced are lower-paying than the jobs that are held by employees, the companies’ pay data will become more equal.

Other research finds that paying employees higher wages

  • Motivates employees to work harder.
  • Attracts more capable and productive workers.
  • Lead to lower turnover
  • Enhance quality and customer service
  • Reduce disciplinary problems and absenteeism
  • Require fewer resources for monitoring
  • Reduces poor performance caused by financial anxiety

Looking forward to Bloom et al.’s published work.

Is Contract Enforcement Important for Firm Productivity?

Contract enforcement is a major player in measuring the ease of doing business in a country. A new working paper demonstrates the importance of enforceable contracts to firm productivity:

In Boehm and Oberfield (2018) we study the use of intermediate inputs (materials) by manufacturing plants in India and link the patterns we find to a major institutional failure: the long delays that petitioners face when trying to enforce contracts in a court of justice. India has long struggled with the sluggishness of its judicial system. Since the 1950’s, the Law Commission of India has repeatedly highlighted the enormous backlogs and suggested policies to alleviate the problem, but with little success. Some of these delays make international headlines, such as in 2010, when eight executives were convicted in the first instance for culpability in the 1984 Bhopal gas leak disaster. One of them had already passed away, and the other seven appealed the conviction (Financial Times 2010)

These delays are not only a social problem, but also an economic problem. When enforcement is weak, firms may choose to purchase from suppliers that they trust (relatives, or long-standing business partners), or avoid purchasing the inputs altogether such as by vertically integrating and making the components themselves, or by switching to a different production process. These decisions can be costly. Components that are tailored specifically to the buyer (‘relationship-specific’ intermediate inputs) are more prone to hold-up problems, and are therefore more dependent on formal court enforcement.

…Our results suggest that courts may be important in shaping aggregate productivity. For each state we ask how much aggregate productivity of the manufacturing sector would rise if court congestion were reduced to be in line with the least congested state. On average across states, the boost to productivity is roughly 5%, and the gains for the states with the most congested courts are roughly 10% (Figure 3).

Are Immigrants a Threat?

Image result for immigration

From a new working paper:

The empirical evidence comes down decidedly on the side of immigrants being less likely to commit crimes. A large body of empirical research concludes that immigrants are less likely than similar US natives to commit crimes, and the incarceration rate is lower among the foreign-born than among the native-born (see, for example, Butcher and Piehl 1998a, 1998b, 2007; Hagan and Palloni 1999; Rumbaut et al. 2006). Among men ages 18 to 39—prime ages for engaging in criminal behavior—the incarceration rate among immigrants is one-fourth the rate among US natives (National Academies of Sciences, Engineering, and Medicine 2015).

…There is some evidence that the lower propensity of immigrants to commit crimes does not carry over to immigrants’ children. The US-born children of immigrants—often called the “second generation”— appear to engage in criminal behavior at rates similar to other US natives (Bersani 2014a, 2014b). This 4 “downward assimilation” may be surprising, since the second generation tends to considerably outperform their immigrant parents in terms of education and labor-market outcomes and therefore might be expected to have even lower rates of criminal behavior (National Academies of Sciences, Engineering, and Medicine 2015). Instead, immigrants’ children are much like their peers in terms of criminal behavior. This evidence mirrors findings that the immigrant advantage over US natives in terms of health tends to not carry over to the second generation (e.g., Acevedo-Garcia et al. 2010).

Although immigrants are less likely to commit crimes than similar US natives, they are disproportionately male and relatively young—characteristics associated with crime. Does this difference in demographic composition mean that the average immigrant is more likely than the average US native to commit crimes? Studies comparing immigrants’ and US natives’ criminal behavior and incarceration rates tend to focus on relatively young men, leaving the broader question unanswered. However, indirect evidence is available from looking at the relationship between immigration and crime rates. If the average immigrant is more likely than the average US native to commit crimes, areas with more immigrants should have higher crime rates than areas with fewer immigrants. The evidence here is clear: crime rates are no higher, and are perhaps lower, in areas with more immigrants. An extensive body of research examines how changes in the foreign-born share of the population affect changes in crime rates. Focusing on changes allows researchers to control for unobservable differences across areas. The finding of either a null relationship or a small negative relationship holds in raw comparisons, in studies that control for other variables that could underlie the results from raw comparisons, and in studies that use instrumental variables to identify immigrant inflows that are independent of factors that also affect crime rates, such as underlying economic conditions (see, for example, Butcher and Piehl 1998b; Lee, Martinez, and Rosenfeld 2001; Reid et al. 2005; Graif and Sampson 2009; Ousey and Kubrin 2009; Stowell et al. 2009; Wadsworth 2010; MacDonald, Hipp, and Gill 2013; Adelman et al. 2017). The lack of a positive relationship is generally robust to using different measures of immigration, looking at different types of crimes, and examining different geographic levels.2 Further, the lack of a positive relationship suggests that immigration does not cause US natives to commit more crimes. This might occur if immigration worsens natives’ labor market opportunities, for example.

The few studies that examine crime among unauthorized immigrants have findings that are consistent with the broader pattern among immigrants—namely, unauthorized immigrants are less likely to commit crimes than similar US natives (apart from immigration-related offenses).4 Likewise, studies that examine the link between the estimated number of unauthorized immigrants as a share of an area’s population and crime rates in that area typically find evidence of null or negative effects (pg. 3-5).

Comparatively, the effects of border control on crime is mixed. The authors conclude,

A crucial fact seems to have been forgotten by some policy makers as they have ramped up immigration enforcement over the last two decades: immigrants are less likely to commit crimes than similar US natives. This is not to say that immigrants never commit crimes. But the evidence is clear that they are not more likely to do so than US natives. The comprehensive 2015 National Academies of Sciences, Engineering, and Medicine report on immigration integration concludes that the finding that immigrants are less likely to commit crimes than US natives “seems to apply to all racial and ethnic groups of immigrants, as well as applying over different decades and across varying historical contexts” (328). Unauthorized immigrants may be slightly more likely than legal immigrants to commit crimes, but they are still less likely than their US-born peers to do so. Further, areas with more immigrants tend to have lower rates of violent and property crimes. In the face of such evidence, policies aimed at reducing the number of immigrants, including unauthorized immigrants, seem unlikely to reduce crime and increase public safety (pg. 11).

Does Female Autonomy Lead to Long-Term Economic Growth?

From a new study:

A number of development economists have found higher gender inequality to be associated with slower development. Amartya Sen (1990) estimated a large number of ‘missing women’, which resulted in skewed sex ratios, and argued that this has been one of history’s crucial development hurdles. Stephan Klasen, with various co-authors, used macroeconomic regressions to show that gender inequality has usually been associated with lower GDP growth in developing countries during the last few decades (Klasen and Lamanna 2009, Gruen and Klasen 2008). This resulted in development policies targeted specifically at women. In 2005, for example, UN Secretary General Kofi Annan stated that gender equality is a prerequisite for eliminating poverty, reducing infant mortality, and reaching universal education (UN 2005). In recent periods, however, a number of doubts have been made public by development economists. Esther Duflo (2012) suggested that there is no automatic effect of gender equality on poverty reduction, citing a number of studies. The causal direction from poverty to gender inequality might be at least as strong as in the opposite direction, according to this view.

…In a new study, we directly assess the growth effects of female autonomy in a dynamic historical context (Baten and de Pleijt 2018). Given the obviously crucial role of endogeneity issues in this debate, we carefully consider the causal nature of the relationship. More specifically, we exploit relatively exogenous variation of (migration-adjusted) lactose tolerance and pasture suitability as instrumental variables for female autonomy. The idea is that high lactose tolerance increased the demand for dairy farming, whereas similarly, a high share of land suitable for pasture farming allowed more supply. In dairy farming, women traditionally had a strong role, which allowed them to participate substantially in income generation during the late medieval and early modern period (Voigtländer and Voth 2013). In contrast, female participation was limited in grain farming, as it requires substantial upper-body strength (Alesina et al. 2013). Hence, the genetic factor of lactose tolerance and pasture suitability influences long-term differences in gender-specific agricultural specialisation. In instrumental variable regressions, we show that the relationship between female autonomy and human capital is likely to be causal (and also address additional econometric issues, such as the exclusion restriction, using Oster ratios, etc.). 

Age-heaping-based numeracy estimates reflect a crucial component of human capital formation. Recent evidence documents that numerical skills are the ones that matter most for economic growth. Hanushek and Woessmann (2012) argued that maths and science skills were crucial for economic success in the 20th century. They observed that these kinds of skills outperform simple measures of school enrolment in explaining economic development. Hence, in the new study we focus on math-related indicators of basic numeracy. We use two different datasets: first, we use a panel dataset of European countries from 1500 to 1850, which covers a long time horizon; second we study 268 regions in Europe, stretching from the Ural mountains in the east to Spain in the southwest and the UK in the northwest. 

Average age at marriage is used as a proxy for female autonomy. Low age at marriage is usually associated with low female autonomy. Age at marriage is highly correlated with other indicators of female autonomy, such as the share of female household heads or the share of couples in which the wife was older than the husband. Age at marriage is particularly interesting because of the microeconomic channel that runs from labour experience to an increase in women’s human capital. After marriage, women typically dropped out of the labour market, and switched to work in the household economy (Diebolt and Perrin 2013). Consequently, after early marriage women provided less teaching and self-learning encouragement to their children, including numeracy and other skills. Early-married women sometimes also valued these skills less because they did not ‘belong to their sphere’, i.e. these skills did not allow identification (Baten et al. 2017).

What do they find?

Figure 3 depicts a strong and positive relationship between average age at marriage and numeracy for the two half centuries following 1700 and 1800. Most countries are close to the regression line. Denmark, the Netherlands, Germany, Sweden, and other countries had high values of female autonomy and numeracy – interestingly, many of the countries of the Second Industrial Revolution of the late 19th century, rather than the UK, the first industrial nation. In contrast, Russia, Poland, Slovakia, Italy, Spain, and Ireland had low values in both periods.

In our regression analyses, we include a large number of control variables, such as religion, serfdom, international trade, and political institutions. We find that the relationship between female autonomy and numeracy is very robust.

We also study the relationship between female autonomy and human capital formation at the regional level in the 19th century. Numeracy and age at marriage (after controlling for country-fixed effects and other control variables) yield an upward sloping regression line (Figure 4). 

…In sum, the empirical results suggest that economies with more female autonomy became (or remained) superstars in economic development. The female part of the population needed to contribute to overall human capital formation and prosperity, otherwise the competition with other economies was lost. Institutions that excluded women from developing human capital – such as being married early, and hence, often dropping out of independent, skill-demanding economic activities – prevented many economies from being successful in human history.