Does Globalization Increase Economic Mobility?

A new working paper draws on Swedish manufacturing data between 1997 and 2013 to determine the effects of globalization on economic mobility. Defining globalization as “a reduction in trade costs” (pg. 22), the authors note,

Most workers land their first full-time job in their 20s and then spend 40 to 50 years in the labor market trying to earn a living. Over their careers, workers acquire new skills, which enables them to change jobs and (sometimes) occupations in order to increase job satisfaction and career earnings. It follows that a complete picture of the impact of globalization on a typical worker should take into account its impact on skill acquisition and the rate at which workers are able to secure better jobs (that is, economic mobility) (pg. 38).

The authors develop “a model of a jobs ladder in which workers gain skills on the job that qualify them for higher-paying jobs at more productive firms” (pg. 38). They explain,

Our main finding is that when trade costs are initially high, globalization increases economic mobility through two channels. First, the reduction in trade costs leads to more international engagement by firms. As the number of exporting firms grows, the ability of workers to gain skills that reduce trade costs is enhanced. This makes it easier for workers to qualify for jobs at the top of the jobs ladder. Second, since high-productivity firms gain disproportionally from falling trade costs, globalization increases wage inequality. And, as the gaps between the wages paid by different groups of firms increase, workers become more willing to (a) incur the moving costs associated with changing jobs and (b) expend effort to keep their skills from deteriorating. As a result, upward economic mobility rises and downward economic mobility (due to demotions or terminations) falls. These changes in economic mobility reduce the differences in expected lifetime incomes forecast by workers in high-wage and low-wage jobs, resulting in the possibility that inequality in lifetime incomes might fall with globalization (even though wage inequality is rising). Even the case in which globalization increases inequality in terms of lifetime incomes, the impact is smaller than its impact on wage inequality (pg. 39).

What’s more,

Employment is reallocated from firms that pay medium wage towards the extremes, with high-wage and low-wage employment both increasing. While it is tempting to interpret this reallocation of employment as an explanation of “job polarization” as described in recent empirical work (see Goos and Manning 2007; Goos, Manning, and Salomons 2009; Autor, Katz and Kearney 2006, 2008 and Autor and Dorn 2013), we believe that would be a mistake…Our results indicate that globalization can result in a shrinking middle-class within a given occupation, with increased export opportunities resulting in more firms willing to recruit the most experienced workers by paying the highest wage; while others react to increased competition from imports by re-orienting their hiring toward inexperienced low-wage workers. These results are not driven by outsourcing. Instead, they are completely driven by the manner in which globalization alters the networks that firms use to fill their vacancies (pg. 39-40).

This image has an empty alt attribute; its file name is employment-shares.jpg

When trade costs are high, “globalization allows [low-wage workers] to move up the jobs ladder more quickly and, as they reach higher and higher rungs, they enjoy the enhanced benefits of the higher real wages generated by freer trade. In this case, a focus on wage inequality can be misleading in that low-wage workers do not lose as much relative to others in the labor market as would be indicated by standard analysis” (pg. 28-29). However, when trade costs are already low, “[w]age inequality rises and the rate at which workers move
out of their entry level jobs slows.” However,

the proper way to measure the effect of globalization on a worker is to examine its impact on that worker’s expected lifetime real income. That measure considers both the change in real wages and the degree of economic mobility faced by that worker. Thus, we can get a better view of how globalization affects inequality by examining the changes in expected lifetime real incomes for workers in different labor market states…Inexperienced workers only hold low-wage jobs for a portion of their lifetime, moving on to much better jobs as they gain skills. As they mature, they benefit from the higher real wages paid to medium-wage and high-wage production workers if they can gain the proper skills and land better jobs. The fact that using current wages as a proxy for lifetime earnings can lead to misleading conclusions is not a new insight. This issue is well understood and heavily researched in many sub-fields of economics; but, as far as we know, it has not received much attention from those investigating the link between globalization and inequality (pg. 29-30).

This image has an empty alt attribute; its file name is inequality.jpg

The devil is in the details.

The Effects of Legalizing Immigrants

Image result for spain immigration

Over at SMU’s Texas-Mexico Center blog, I wrote,

Despite the recent political rhetoric and anti-immigrant sentiments, the economic benefits of immigration are well-established in the empirical literature. A 2011 meta-analysis by economist Michael Clemens found that dropping all current immigration restrictions would result in a doubling of world GDP.

A more recent analysis corroborated these findings, concluding that lifting all migration restrictions would increase world output by 126%. In 2015, migrants made up 3.4% of the world population yet contributed about $6.7 trillion to global output—9.4% of world GDP. The McKinsey Global Institute estimates that this is $3 trillion more than these migrants would have produced had they stayed in their origin countries. Even undocumented workers in the United States contribute about 3.6% of private-sector GDP annually—around $6 trillion dollars over a 10-year period. Granting these migrants legal status would increase their contribution to 4.8%. 

On this last point, a recent study explores the effects of immigrant legalization in Spain. The authors explain the background for the natural experiment:

In the early 2000s, Spain experienced an incredible boom in immigration. The share of immigrants in the working-age population increased from less than 2% in 1995 to around 10% in 2004. Many of these newly arrived immigrants lacked work permits. By 2004, there were close to 1 million undocumented immigrants in a country of around 43 million inhabitants.

Despite this large number of undocumented immigrants, the government at the time, led by Jose Maria Aznar (Popular Party) and with Mariano Rajoy in its cabinet, was unlikely to legalise the work status of immigrants. Traditionally, the Popular Party had been proposing tougher policies against immigration. Its main stance was to avoid implementing policies that could attract new waves of immigrants. In this context, in the early 2000s, immigrants were granted work permits mostly on the basis of family reunification.

On 14 March 2004, voters in the Spanish general election had to determine whether the Popular Party would continue in power or be replaced by the Socialist Party. In the week before the election, the outcome seemed clear: the polls were forecasting that Zapatero of the Socialist Party was trailing Rajoy by seven percentage points. 

Yet, something completely unexpected happened just three days before the election which, as shown by Garcia-Montalvo (2011), changed the final outcome: Madrid suffered the largest terrorist attack in Spanish history, a tragedy that was poorly managed by the Popular Party. As a result, the Socialist Party came to power, and one of the first policies it implemented was the legalisation of nearly 600,000 immigrants already living (and working illegally) in Spain.

Using administrative payroll tax revenues, the authors find

that the legalisation of immigrants’ work status increased revenues locally — i.e. at the province level — by around €4,189 per newly legalised immigrant. This amount is only 55% of what we would have expected if newly legalised immigrants had shared the same characteristics as previous contributors to the social security system and had enjoyed similar labour market experiences. Two factors may explain this. First, newly legalised immigrants were perhaps disproportionately low-skilled and had worse labour market experiences than natives. Second, the legalisation may also have affected previous workers.

…Using very detailed administrative and survey data on wages and employment, we show that the policy change disproportionately affected the labour market outcomes of workers in high-immigrant locations relative to low-immigrant locations. In particular, it worsened employment opportunities for both low-skilled natives and immigrants, while it improved them for high-skilled workers. Among low-skilled natives, those who lost their jobs were negatively selected — the policy change negatively affected employment prospects of native low-skilled workers at the bottom of the wage distribution. Putting together all the labour market changes and comparing them to payroll tax revenue changes, we show that this negative selection is crucial to fully understand the effects of the reform.

We also show that, following the reform, many immigrants moved from high- to low-immigrant locations. This is important since these immigrants also contributed to payroll tax revenues, but in traditionally low-immigrant locations. This, in turn, means that comparing local payroll tax revenues in high- relative to low-immigrant locations to evaluate the effect of the policy may underestimate the true impact of immigrant legalisation on payroll tax revenues. Once we take into account internal migration and selection, we argue that the true contribution was almost €5,000 per newly legalised immigrant, i.e. substantially higher than what we would have been able to estimate on the basis of local tax revenue data alone.

Incoherent Know-Nothings

Cards Against Humanity’s Pulse of the Nation poll from 2017 to 2018 has some pretty interesting, disturbing, and rather unsurprising findings about the American public:

Conflicting Views

39.1% of Democrats think that it’s wrong to negatively stereotype people based on their place of birth, but also think Southerners are more racist.

65.2% of Republicans think that people are too easily offended, yet find Black Lives Matter offensive.

64.6% of Democrats think that a woman has the right to do what she wants with her body, except when it comes to selling her kidney. Nearly half also believe a woman has the right to do whatever she wants with her body, except sell it for sex.

57.9% of Republicans think that people should be free to express their political opinions in the workplace, but athletes shouldn’t be allowed to make political protests at games.

Over half of Democrats think that men and women “are equal in their talents and abilities,” except when it comes to multitasking and empathy.

About 1/3 of Republicans think we should be more suspicious of foreigners, yet believe Putin when he says he didn’t interfere in the 2016 election. (You’re twice as likely to do this if you support Trump.)

Over half of Republicans believe nobody deserves a handout and that the government should do more to help small, working-class towns in America’s heartland.

About 1/3 of Democrats say that they trust the scientific consensus, just not when it comes to GMOs.

Political Ignorance

39% of Americans either think low GDP is better than high GDP or have no clue altogether.

The majority of Americans can’t name the three branches of government.

Only 12.7% of Americans can name a living, breathing economist. 55.9% can’t name a living economist, but think their opinions about economic policy are well-informed.

The richest 1% of Americans own 39% of the country’s wealth. Everyone overestimates the amount. If you’re a Democrat, you think it is 75 percent. If you’re a Republican, you think it’s 50 percent. Perhaps surprisingly, the more educated you are, the more likely you are to overestimate the amount.

Nearly half of Americans do not believe the U.S. has interfered with foreign democratic elections. You’re less likely to believe it if you’re Republican.

Other Stuff

Those who think “sex without love” is okay are far more likely to be pro-choice.

If you rely on “common sense” instead of empirical evidence, you’re likely older, less educated, and lack a Twitter account. You also are more likely support military action against Russia for their 2016 election interference. 

29% of Trump supporters would still stick with him in 2020 even if he murdered journalist for spreading lies.

Do Bumps in the Minimum Wage Increase the Number of Job Seekers?

Image result for minimum wage gif

Some argue that increasing the minimum wage will increase the number of job seekers and, consequently, employment. From a new NBER paper:

Do minimum wage increases affect search effort by job seekers?

…We investigate the effect of minimum wage increases on job search effort utilizing data from the Current Population Survey (CPS) and the American Time Use Survey (ATUS). We exploit the staggered nature of CPS and ATUS interviews and use an event-study approach, leveraging within-state variation in the adoption of minimum wage changes. We account for shocks affecting a particular state in a given year as well as month effects to control for seasonality, and individual demographic characteristics. Intuitively, we compare the outcomes in each month near the treatment date to the outcomes for otherwise-identical individuals in the same state and year whose survey period was not near a treatment date.

We find no evidence that the minimum wage has persistent effects on search effort; the likelihood of searching does not increase in the aftermath of minimum wage increases. However, there is a large yet transitory increase in the intensive margin of search effort, concentrated
in the month of the minimum wage increase, that fades almost immediately. There is no short-run increase in the employment rate nor changes in observable characteristics of searchers, suggesting that our results are not driven by changes in the composition of job seekers. These findings are robust to the inclusion of demographic controls, the duration of unemployment benefits, and month-by-year fixed effects that account for any idiosyncratic national-level variation in a given month. We also conduct a permutation test for our search duration results in which we randomly assign minimum wage increases across time periods and show that these results do not appear to be due to chance.

Our results call into question the assumption underpinning search-and-matching models as applied to analysis of the minimum wage – namely, that more workers will enter the labor market and each worker will search harder, increasing the returns to firm vacancy postings. Importantly, we find minimum wage increases do not induce individuals to begin searching. While we find that minimum wage increases yield significant increases in worker search effort on the intensive margin, they are transitory (pg. 2-3).

 

What Kind of Immigration Fuels Nationalism?

From a recent NBER Digest:

In Skill of Immigrants and Vote of the Natives: Immigration and Nationalism in European Elections 2007-16 (NBER Working Paper No. 25077), Simone Moriconi, Giovanni Peri, and Riccardo Turati explore the relationship between immigration and European elections. They develop an index of “nationalistic” attitudes of political parties to measure the shift in preferences among voters when confronted with influxes of skilled and unskilled immigrants. They find that larger inflows of highly educated immigrants dampen nationalistic sentiments, while larger inflows of less-educated immigrants heighten them. Their results imply that a more balanced inflow of high-skilled and low-skilled immigrants could attenuate voters’ nationalistic attitudes.

...The new study tracks voter attitudes and behavior for all political parties and elections in 12 European countries for a decade. It relies on demographic and political data from the European Social Survey and a number of other sources. In addition, the researchers collected and classified the political manifestos of 126 parties for 28 elections, focusing in particular on how frequently these materials mentioned nationalistic subjects, the European Union, and other indicators of where parties stood on the political spectrum.

The researchers found “that highly educated native voters are less nationalistic in their attitudes towards immigrants than less-educated natives. The data also show strong nationalistic sentiments in regional pockets in the United Kingdom, Ireland, France, Germany, Demark, Sweden, Norway and, especially, Italy.”

The results suggest that a 1 percent increase in the share of a country’s population who are immigrants in highly educated, highly skilled groups was associated with a 0.1 standard deviation voting change away from nationalism. An increase of comparable size in the number of less-educated and lower-skilled immigrants led to a 0.12 standard deviation voting change towards nationalism. The same patterns emerged when the researchers analyzed voter sentiment expressed in surveys. In this case, a 1 percent increase in high-skilled immigrants led to a 0.07 standard deviation decrease away from nationalism, while a 1 percent increase in lower-skilled immigrants lead to a 0.07 standard deviation increase in nationalism. The results were broadly similar regardless of whether the analysis focused on all immigrants or only on immigrants from non-EU nations.

Immigration is not only about ethnicity, but class as well.

What’s Behind Cuba’s Health Outcomes?

The above comes from Michael Moore’s Sicko. Cuba’s healthcare system is a common talking point among those of Moore’s persuasion. However, a recent study should give us pause regarding some of the overly positive claims about Cuba’s system. First, what people like Moore get right:

How is Cuba healthy while poor? Most attribute the fact to Cuba’s zero monetary cost health care system. There is some truth to that attribution. With 11.1% of GDP dedicated to health care and 0.8% of the population working as physicians, a substantial amount of resources is directed towards reducing infant mortality and increasing longevity. An economy with centralized economic planning by government like that of Cuba can force more resources into an industry than its population might desire in order to achieve improved outcomes in that industry at the expense of other goods and services the population might more highly desire (pg. 755).

However,

Centralized planning has disadvantages. Physicians are given health outcome targets to meet or face penalties. This provides incentives to manipulate data. Take Cuba’s much praised infant mortality rate for example. In most countries, the ratio of the numbers of neonatal deaths and late fetal deaths stay within a certain range of each other as they have many common causes and determinants. One study found that that while the ratio of late fetal deaths to early neonatal deaths in countries with available data stood between 1.04 and 3.03 (Gonzalez, 2015)—a ratio which is representative of Latin American countries as well (Gonzalez and Gilleskie, 2017). Cuba, with a ratio of 6, was a clear outlier. This skewed ratio is evidence that physicians likely reclassified early neonatal deaths as late fetal deaths, thus deflating the infant mortality statistics and propping up life expectancy. Cuban doctors were re-categorizing neonatal deaths as late fetal deaths in order for doctors to meet government targets for infant mortality.

Using the ratios found for other countries, corrections were proposed to the statistics published by the Cuban government: instead of 5.79 per 1000 births, the rate stands between 7.45 and 11.16 per 1000 births. Recalculating life expectancy at birth to account for these corrections (using WHO life tables and assuming that they are accurate depictions of reality), the life expectancy at birth of men by between 0.22 and 0.55 years (Gonzalez, 2015) (pg. 755).

But that’s not the only thing driving low infant mortality rates:

Coercing or pressuring patients into having abortions artificially improve infant mortality by preventing marginally riskier births from occurring help doctors meet their centrally fixed targets. At 72.8 abortions per 100 births, Cuba has one of the highest abortion rates in the world. If only 5% of the abortions are actually pressured abortions meant to keep health statistics up, life expectancy at birth must be lowered by a sizeable amount. If we combine the misreporting of late fetal deaths and pressured abortions, life expectancy would drop by between 1.46 and 1.79 years for men. In Figure 1 below, we show that that with this adjustment alone, instead of being first in the ranking of life expectancy at birth for men in Latin America and the Caribbean, Cuba falls either to the third or fourth place depending on the range (pg. 755-756).

The researchers explain, “Other repressive policies, unrelated to health care, contribute to Cuba’s health outcomes” (pg. 756) These include:

  • Restrictions in car ownership leading to low automobile fatalities.
  • Rationing combined with physically demanding transportation (e.g., cycling) contributing to reductions in obesity and deaths caused by diabetes, coronary heart diseases and strokes.

The researchers conclude,

Cuban mortality and longevity statistics appear impressive. They are a result of some combination of the government’s choice to allocate more resources into the health care industry (at the expense of other industries that could produce needed goods) and from coercive measures through both health delivery and economic planning that improve health statistics at the expense of other spheres of life.

Although the USA and other countries re-examine how to design health care delivery they should not uncritically accept the myth that the Cuban health care system has been the sole, or even the most important, cause of Cuba’s abnormally high longevity statistics. The role of Cuban economic and political oppression in coercing ‘good’ health outcomes merits further study (pg. 756).

The Benefits of Walmart

Image result for walmart

Walmart catches a lot of grief. For example, as reported by CNN, Bernie Sanders recently “introduced a bill, titled the Stop Walmart Act, that would prevent large companies from buying back stock unless they pay all employees at least $15 an hour, allow workers to earn up to seven days of paid sick leave and limit CEO compensation to no more than 150 times the median pay of all staffers.” Yet, many don’t consider the massive benefits produced by Walmart: 

A 2005 Global Insight study commissioned by Wal-Mart and overseen by an independent panel suggested that a new Wal-Mart would create, on net, 137 jobs in the short term and 97 jobs in the long term (Global Insight 2005: 2). Studying Pennsylvania counties, Hicks (2005, discussed by Vedder and Cox 2006: 110) found that the company led to a net increase of fifty new jobs with a 40% reduction in job turnover. Hicks (2007: 93-94) uses data from Indiana to estimate that Wal-Mart increases rural retail employment from 3.4% to 4.8% after correcting for endogeneity. After correcting for endogeneity of urban Wal-Mart entry, Hicks argues that Wal-Mart leads to a 1.2% increase in employment but points out that this estimate is statistically insignificant.

…Wal-Mart’s most obvious effect on the retail sector comes through its policy of Every Day Low Prices. Basker (2005b) and Basker and Noel (2009) estimate that WalMart has a substantial price advantage over competitors with the effect being that prices among incumbent competitors fall after Wal-Mart entry. Hausman and Leibtag (2007: 1147) argue that the compensating variation from Big Box retailers’ effect on prices leads to welfare increases of some 25% of total food expenditure for people who enjoy the direct and indirect effects of Big Box stores. Further, they argue (Hausman and Leibtag 2009) that the Consumer Price Index is over-estimated because it fails to account properly for price effects of supercenters, mass merchandisers, and club stores. Evaluating estimates of the price effects of Big Box retailers and adjusting for foreign sales, Vedder and Cox (2006: 18-19) argue that “the annual American-derived welfare gains are probably still in excess of $65 billion, or about $225 for every American, or $900 for a typical family of four.”

…Jason Furman (2005) called Wal-Mart a “progressive success story” because of its impact on prices. He notes that if the 2005 Global Insight estimate of annual average household savings of $2,329 is accurate, the annual Wal-Mart related consumer savings of $263 billion dwarfs Wal-Mart-generated reductions in retail wages of $4.7 billion estimated by Dube et al. (2005). Hicks (2007: 82) notes that reductions in nominal retail wages are likely offset by larger price reductions, which translates into higher real wages. Courtemanche and Carden’s (2011a) estimate of $177 per household in savings attributable to the effects of Wal-Mart Supercenters in 2002 multiplied by the 105,401,101 households in the 2000 census yields household savings of $18.7 billion, which is still substantially higher than Dube et al.’s estimate of lost wages. 

Hausman and Leibtag (2007: 25) argue that the compensating variation—i.e., welfare increase—attributable to supercenters, mass merchandisers, and club stores is some 25% of food expenditures. Since poorer households spend more of their income on food, the effect (as a percentage of income) is higher toward the bottom of the income distribution (Furman 2005: 2-3). Hausman and Leibtag (2007: 1172, 1174) further argue that compensating variation from access to non-traditional retailers is higher at lower income levels, which would make the effect even more progressive (pgs. 8-9).

A brand new study demonstrates even more benefits provided by Walmart:

We estimate the effects that Walmart Supercenters have on food security using data from the 2001–2012 waves of the December Current Population Study Food Security Supplement (CPS-FSS). Narrow geographic identifiers available in the restricted version of these data enable us to compute the distance from each household’s census tract to the nearest Walmart Supercenter. Our outcomes are counts of the number of affirmative responses on the household and child-specific portions of the food insecurity questionnaire, along with binary variables for household food insecurity, household very low food security, child food insecurity, and child very low food security. We estimate instrumental variables (IV) models that leverage the predictable geographic expansion patterns of Walmart Supercenters outward from corporate headquarters. Specifically, we instrument for Walmart Supercenters with the interaction of distance from Bentonville, Arkansas (Walmart’s headquarters), with time. For both households in general and children specifically, the results show that a closer proximity to the nearest Walmart Supercenter leads to sizeable and statistically significant improvements in all food security measures except the indicator for very low food security. Subsample analyses reveal that the effects are especially large for low-income households and children, though they are also sizeable for middle-income children.

As journalist John Tierney asked, “How could any progressive with a conscience oppose an organization that confers such benefits?”

Economic Growth and Corruption

In my latest paper in Economic Affairs, I wrote,

Drawing on the EFW Index, Brennan (2016a)…points to a strong positive correlation between a country’s degree of economic freedom and its lack of public sector corruption. Granted, a lack of corruption could very well give rise to market reforms and increased economic freedom instead of the other way around. However, recent research on China’s anti-corruption reforms (Ding et al. 2017; Li et al. 2017) suggests that markets may actually pave the way for anti-corruption reforms (pg. 425).

Furthermore,

Market liberalisation can also have indirect effects on war and violence. For example, Neudorfer and Theuerkauf (2014) explore the effects of public sector corruption on ethnic violence by analysing 81 to 121 countries between 1984 and 2007. They find that corruption has a robust positive effect…on the risk of ethnic civil war. When the evidence provided in the previous sections by Brennan (2016a) and Lin et al. (2017) is considered, we find that market liberalisation deters corruption and, consequently, ethnic violence (pg. 429).

Research suggests that economic growth may reduce corruption:

The traditional explanation for this relationship has been the theory articulated by Wolfenson above – corruption increases the cost and risk of business activity, thereby deterring investment and depressing growth that could have lifted citizens out of poverty (Mauro 1995, Wei 1999).

However, there is an alternative possibility that has received less attention among development practitioners and academics. The strong relationship between income and growth may result from exactly the opposite causal relationship – countries may be growing out of corruption (Tresiman 2002). Over time, economic growth reduces both the incentives for government officials to extract bribes and firms’ willingness to pay them. Some scholars of developed countries have discussed this possibility in terms of a ‘life cycle’ theory with corruption peaking at early stages of development and declining as countries industrialise (Huntington 1968, Theobald 1990, Ramirez 2013). However, there has been little work either testing for this empirical link from growth to corruption, or laying out the specific mechanisms that could generate the link.

The authors continue:

The key theoretical insight of our argument is that the share of bribes that officials will choose to extract as rents depends on a firm’s ability to move and set up business in a different location. Ask for too much, and firms that have the ability to do so, will simply pull up anchor and head to safer harbours. Because officials know this, they are likely to set a bribe amount that is just below the cost of moving.

Building on that insight, we show that as firms grow the cost of moving should decline relative to firm size. The fixed cost of moving becomes less expensive relative to revenue, and more and more firms have the opportunity to escape the bribe requests of officials in their locality. Corrupt officials faced with a sudden growth surge must lower their bribe rates, or face losing their key providers of employment and tax payers to competitors.

…The theory we propose has important policy implications. To the extent this theoretical mechanism is important, rather than focusing on politically difficult institutional changes to combat corruption, resources might be better spent on policies that facilitate capital mobility across subnational jurisdictions. Providing clear titles to business premises, for instance, enables entrepreneurs to sell and recoup the full market value of land. Such businesses are more mobile than renters or owners with insecure titles, who risk significant losses if they try to escape corruption by fleeing across the border.

Drawing on “an annual survey funded by USAID and administered by the Vietnamese Chamber of Commerce and Industry,” the researchers find “that the average bribe rate decreases as GDP per capita increases” and “that large firms actually pay lower bribe rates, which is what our theory predicts. Firms with higher revenues are more put out by a high bribe rate, since it increases the amount of bribes they must pay dramatically. To retain them then, officials must push their bribe rate lower.” 

Then, using “a census of firms conducted by Vietnam’s General Statistical Office (GSO) [to] calculate aggregate employment at the province-industry-year level,” the authors

show that exogenous industry-wide performance is indeed a strong predictor of a firm’s performance. A doubling of total employment in the industry is associated with a 1.6 percentage point reduction in the bribe rate, or about 42% of the mean level. Moreover, the effect is more pronounced for highly mobile firms. The magnitude of the effect of growth on bribe reductions is 17% larger for firms in possession of a Land Use Rights Certificate, which facilitates the sale of their business premises. Similarly, the effect is 20% greater for firms that already have branch operations in other provinces, and therefore possess knowledge and experience that could facilitate movement.

These effects survive a battery of robustness tests and alternative specifications, providing compelling evidence that growth can directly reduce corruption.

In short, economic growth can decrease corruption by undermining the power of officials to extract bribes. But this is likely part of a virtuous feedback loop. For example, a 2017 paper 

exploit[s] spatial variation in randomized anti-corruption audits related to government procurement contracts in Brazil to assess how corruption affects resource allocation, firm performance, and the local economy. After an anti-corruption crackdown, regions experience more entrepreneurship, improved access to finance, and higher levels of economic activity. Using firms involved in corrupt business with the municipality, we find that two channels explain these facts: allocation of resources to less efficient firms, and distortions in government dependent firms. The second channel dominates, as after the audits government dependent firms grow and reallocate resources within the organization (pg. 31). 

As I state in the beginning of my paper,

Of course, it is far easier to demonstrate correlation than causation, and while some studies do find markets playing a causal role in moral development, most simply establish a positive relationship. However, findings that ‘merely’ demonstrate positive correlations should be interpreted in light of the feedback loops: even if moral behaviours are foundational and give rise to market systems (instead of vice versa), market systems in turn reinforce these virtues by imbuing them with value. As Paul Zak (2011, p. 230) explains, ‘Markets are moral in two senses. Moral behavior is necessary for exchange in moderately regulated markets, for example, to reduce cheating without exorbitant transaction costs. In addition, market exchange itself can lead to greater expression of morals in nonmarket settings’ (pg. 423).

The Gender Wage Gap: Union Edition

Image result for union woman

Similar to previous research, a new working paper shows that the gender wage gap is driven largely by the amount of hours men and women choose to work. The authors draw on data from the Massachusetts Bay Transportation Authority (MBTA), explaining that the “bus and train operators are all represented by the same union, Carmen’s Local 589, and are all covered by the same bargaining agreement. The agreement specifies that seniority in one’s garage is the sole determinant of one’s work opportunities. Conditional on seniority, men and women face the same choice sets of schedules, routes, vacation days, and overtime hours, among other amenities” (pg. 2).

What do they find?

We show that a gender earnings gap can exist even in a controlled environment where work tasks are similar, wages are identical, and tenure dictates promotions. The gap of $0.89 in our setting, which is 60% of the earnings gap across the United States, can be explained entirely by the fact that, while having the same choice sets in the workplace, women and men make different choices. Women use the Family Medical Leave Act (FMLA) to take more unpaid time off than men and they work fewer overtime hours at 1.5 times the wage rate. At the root of these different choices is the fact that women value time and flexibility more than men. Men and women choose to work similar hours of overtime when it is scheduled a quarter in advance, but men work nearly twice as many overtime hours than women when they are scheduled the day before. Using W-4 filings to ascertain marital status and the presence of dependents, we show that women with dependents – especially single women – value time away from work more than men with dependents.

When selecting their work schedule for the following quarter, women try to avoid inconvenient days, like weekends, and shifts, like split-shifts, more than men. Prioritizing schedule related amenities over route quality-related amenities, women select routes with higher probabilities of assaults and collisions in order to avoid unfavorable schedules. When faced with having to work an unfavorable schedule, like a weekend, holiday, or split shift, women take more unpaid time off. Men also take more unpaid time off in those circumstances, but they more than make up for lost earnings with overtime. While constrained schedules lead to lower earnings for women, they result in higher earnings for men. In an effort to reduce absenteeism and overtime expenditures, the MBTA oversaw two policy changes: one that made it harder to take unpaid time off with FMLA and another that made it harder to be paid at the overtime rate. While the policy changes reduced the gender earnings gap from $0.89 to $0.94 and made it harder for operators to trade off regular hours for overtime, they also decreased women’swell-being by further constraining the work environment (pg. 34-35).

Texico: The Texas-Mexico Economy

Michael Cox and Richard Alm of SMU’s O’Neil Center have an essay in D Magazine based on the latest report from the Center. The two

imagine Texas and Mexico as one economy, connected by exports, imports, migration, cross-border business investments, transport infrastructure, tourism, and knowledge transfers. As a combined economy, Texas and Mexico churn out an annual GDP of more than $4 trillion, enough to rank as the world’s sixth-largest economy, just behind Germany and ahead of Russia.

We denote this sprawling and diverse economy by the portmanteau word: Texico. The name captures the reality that over the past quarter-century the Texas and Mexico economies have emerged as highly integrated, making an often-unsung contribution to Texas’ reign as America’s top-performing state economy.

The explain how trade has deeply integrated the two countries: 

An often-cited gauge of integration is trade—exports moving south, imports moving north. They totaled $188 billion last year, or more than 11 percent of gross state product, separating Texas from all other states in doing business with Mexico.

Texas companies are finding business opportunities in Mexico—among them, cosmetics-maker Mary Kay Inc. and telecommunications giant AT&T Inc., both based in Dallas-Fort Worth. At the same time, Mexican companies are heading northward and expanding their businesses, including Mission Foods in Irving and the movie theater chain Cinépolis in Addison.

The Texas and Mexico economies are more formidable combined rather than separate. Binational supply chains, for example, take advantage of low production costs in Mexico and highly skilled professional labor in Texas. The companies emerge more competitive in the global marketplace, able to sell their wares at a better price.

Automobile production comes to mind—for good reason. Plants in the Dallas-Fort Worth area are on the northern edge of the Texas-Mexico Automotive SuperCluster region, which includes close to 30 assembly plants and more than 230 parts suppliers in Texas and Mexico’s northern states.

Texas and Mexico have already profited a great deal from their binational economy, even though work began in earnest only recently. Mexico didn’t open its energy and telecom markets until just a few years ago. During negotiations that led to the North American Free Trade Agreement, Mexico clung to its monopolies in these industries. With its oil output falling, Mexico finally lifted its ban on foreign oil and gas companies three years ago. If all goes well, this should be a bonanza for Texas, with its deep roster of oilfield services and exploration companies. The telecom monopoly expired about the same time—and AT&T rushed in with its wireless service.

The annual report provides a few more interesting insights:

The report explains the slow convergence above: “Other factors like low levels of education shouldn’t be ignored, but the ongoing plague of corruption, cronyism and rising violence go a long way toward explaining why Mexican growth and income haven’t converged with the United States or kept pace with the likes of Chile, South Korea and China” (pg. 13). 

Cox and Alm conclude,

Texans are well aware of Mexico’s shortcomings, including corruption and drug-cartel violence. None of these problems will get any better by enacting policies that build barriers against Mexico and harm the Texas and Mexico economies. Perhaps Trump and Obrador will decide that the best course lies in expedient practicality—recognizing the fact that Texico has been working and building a large constituency. If these two leaders don’t make a mess of things, the businesses of Texas and Mexico can take it from there.