Category: Data Source

Data on private security forces

by  Tyler Cowen in

From a recent paper:

We bring novel data to bear on these questions, presenting the largest empirical study of private security to date. We introduce an administrative dataset covering nearly 300,000 licensed private security officers in the State of Florida. By linking this dataset to similarly comprehensive information about public law enforcement, we have, for the first time, a nearly complete picture of the entire security labor market in one state. We report two principal findings. First, the public and private security markets are predominantly characterized by occupational segregation, not integration. The individuals who compose the private security sector differ markedly from the public police; they are, for example, significantly less likely to be white men. We also find that few private officers, roughly 2%, have previously worked in public policing, and even fewer will go on to policing in the future. Second, while former police make up a small share of all private security, roughly a quarter of cops who do cross over have been fired from a policing job. In fact, fired police officers are nearly as likely to land in private security as to find another policing job, and a full quarter end up in one or the other. We explore the implications of these findings, including intersections with police abolition and the future of policing, at the paper’s close.

That is by Ben Grunwald, John Rappaport, and Michael Berg.  Via the excellent Kevin Lewis.

LDS and indeed USA fact of the day

by  Tyler Cowen in

Thus, approximately 1 out of 7 people who I classify as a Latter-day Saint attends [religious service] weekly.

And:

However, only 5% of Americans attend services “weekly”, far fewer than the ~22% who report to do so in surveys.

Oh, and this:

The religions with the longest average visit duration are Orthodox Christians (116 minutes), Latter-day Saints (115 minutes), and Jehovah Witnesses (115 minutes). The religions with the shortest average visit duration include Muslims (51 minutes), Catholics (66 minutes), and Buddhists (71 minutes). Jews (92 minutes) and Protestants (102 minutes) have average durations in the middle of the distribution.

That is from a new Devin G. Pope paper on religious attendance, measured by using cellphone data rather than self-reports.  Interesting throughout.

Fair coins aren’t fair

by  Tyler Cowen in

Many people have flipped coins but few have stopped to ponder the statistical and physical intricacies of the process. In a preregistered study we collected 350,757 coin flips to test the counterintuitive prediction from a physics model of human coin tossing developed by Diaconis, Holmes, and Montgomery (D-H-M; 2007). The model asserts that when people flip an ordinary coin, it tends to land on the same side it started — D-H-M estimated the probability of a same-side outcome to be about 51%. Our data lend strong support to this precise prediction: the coins landed on the same side more often than not, Pr(same side)=0.508, 95% credible interval (CI) [0.506, 0.509], BFsame-side bias=2364. Furthermore, the data revealed considerable between-people variation in the degree of this same-side bias. Our data also confirmed the generic prediction that when people flip an ordinary coin — with the initial side-up randomly determined — it is equally likely to land heads or tails: Pr(heads)=0.500, 95% CI [0.498, 0.502], BFheads-tails bias=0.183. Furthermore, this lack of heads-tails bias does not appear to vary across coins. Our data therefore provide strong evidence that when some (but not all) people flip a fair coin, it tends to land on the same side it started. Our data provide compelling statistical support for D-H-M physics model of coin tossing.

By František Bartoš, et.al., that is a paper from late last year.

350+ coauthors study reproducibility in economics

by  Tyler Cowen in

Jon Hartley is one I know, here is the abstract:

This study pushes our understanding of research reliability by reproducing and replicating claims from 110 papers in leading economic and political science journals. The analysis involves computational reproducibility checks and robustness assessments. It reveals several patterns. First, we uncover a high rate of fully computationally reproducible results (over 85%). Second, excluding minor issues like missing packages or broken pathways, we uncover coding errors for about 25% of studies, with some studies containing multiple errors. Third, we test the robustness of the results to 5,511 re-analyses. We find a robustness reproducibility of about 70%. Robustness reproducibility rates are relatively higher for re-analyses that introduce new data and lower for re-analyses that change the sample or the definition of the dependent variable. Fourth, 52% of re-analysis effect size estimates are smaller than the original published estimates and the average statistical significance of a re-analysis is 77% of the original. Lastly, we rely on six teams of researchers working independently to answer eight additional research questions on the determinants of robustness reproducibility. Most teams find a negative relationship between replicators’ experience and reproducibility, while finding no relationship between reproducibility and the provision of intermediate or even raw data combined with the necessary cleaning codes.

Here is the full paper, here are some Twitter images.  I have added the emphasis on the last sentence.

Disparities in psychological traits and incomes

by  Tyler Cowen in

There are pronounced racial, ethnic, and gender gaps in income in the U.S. We investigate whether these correspond with differences in competitiveness, risk tolerance, and confidence relative to performance in a large, stratified sample of the U.S. prime-age population. We find substantial differences in all three traits across Black, Hispanic, and White males and females. These traits predict individual income. Competitiveness and risk tolerance help explain the White gender income gap. Competitiveness also affects the Black-White income gap between men. Confidence about one’s performance helps explain a substantial and significant portion of all five race-gender income gaps with White men.

That is from a new paper by Aurélie Dariel, John Ham, Nikos Nikiforakis, and Jan Stoop.  The number of data points is 2,463.  Here is one sentence from the paper:

The sizes of the effects are substantial: individuals above the median in terms of competitiveness and risk tolerance,
for instance, have incomes that are 21.2% and 15.7% higher than those below the median, respectively, when jointly estimated. Confidence in relative performance is also associated with income: individuals in the upper and lower third of the distribution (the upper third being overconfident and the lower third being underconfident) have incomes that are 23.5% and 16.7% lower than the middle third, who are better at evaluating their relative performance.

And this:

We find that controlling for confidence substantially and significantly reduces the unexplained income gaps between White men and all of our other five REG groups; the effects range from 7.2% of the differential (White women versus White men) to 18.7% (Hispanic men versus White men). Only controlling for competitiveness significantly reduces the unexplained income gap between White women and White men by 5.9%, but increases the unexplained income gap between Black men and White men by 5.1%. Only controlling for risk tolerance, on the other hand, does not significantly affect any of the income gaps, with the exception of a (marginally) significant reduction of 4.1% in the gap between White women and White men. Jointly controlling for the three traits significantly reduces the unexplained income gap between Black women and White men (by 15.2%), Hispanic women and White men (by 11.5%), and White women and White men (by 15.0%). However, these traits do not explain the gap between Black men and White men, as the overconfidence and competitiveness effects go in opposite directions.

All worthy of a ponder.  I did find this result of particular interest:

On average, Blacks and Hispanics are 9.7% more competitive than Whites.

You will note this is based on self-reports.  While self-reports often are more reliable than outsiders might think, are they so reliable for making comparisons across different groups in this manner?  And the variable “confidence in relative performance” — might that be a proxy for other, unobserved but also quite real factors?

Via a loyal MR reader, and I commend the researchers for their courage, even if I am not convinced by everything they have done.

An RCT for income-sharing agreements

by  Tyler Cowen in

Is this the first one?

We conduct a survey-based experiment with 2,776 students at a non-profit university to analyze income insurance demand in education financing. We offered students a hypothetical choice: either a federal loan with income-driven repayment or an income-share agreement (ISA), with randomized framing of downside protections. Emphasizing income insurance increased ISA uptake by 43%. We observe that students are responsive to changes in contract terms and possible student loan cancellation, which is evidence of preference adjustment or adverse selection. Our results indicate that framing specific terms can increase demand for higher education insurance to potentially address risk for students with varying outcomes.

That is from a new NBER working paper by Sidhya Balakrishnan, Eric Bettinger, Michael S. Kofoed, Dubravka Ritter, Douglas A. Webber, Ege Aksu & Jonathan S. Hartley.

Netherlands fact of the day

by  Tyler Cowen in

The country, which is a bit bigger than Maryland, not only accomplished this feat but also has become the world’s second largest exporter of agricultural products by value behind the United States. Perhaps even more significant in the face of a warming planet: It is among the largest exporters of agricultural and food technology. The Dutch have pioneered cell-cultured meat, vertical farming, seed technology and robotics in milking and harvesting — spearheading innovations that focus on decreased water usage as well as reduced carbon and methane emissions…

The country has nearly 24,000 acres — almost twice the size of Manhattan — of crops growing in greenhouses. These greenhouses, with less fertilizer and water, can grow in a single acre what would take 10 acres of traditional dirt farming to achieve. Dutch farms use only a half-gallon of water to grow about a pound of tomatoes, while the global average is more than 28 gallons.

Here is the full article, via S.  The article is interesting throughout.  However here is a more recent piece on the Dutch nitrogen revolt.

New results on intergenerational progress

by  Tyler Cowen in

The full paper title is “Has Intergenerational Progress Stalled? Income Growth Over Five Generations of Americans,” by Kevin Corrinth and Jeff Larrimore.  Here is the abstract:

We find that each of the past four generations of Americans was better off than the previous one, using a post-tax, post-transfer income measure constructed annually from 1963-2022 based on the Current Population Survey Annual Social and Economic Supplement. At age 36–40, Millennials had a real median household income that was 18 percent higher than that of the previous generation at the same age. This rate of intergenerational progress was slower than that experienced by the Silent Generation (34 percent) and Baby Boomers (27 percent), but similar to that experienced by Generation X (16 percent). Slower progress for Generation X and Millennials is due to their stalled growth in work hours—holding work hours constant, they experienced a greater intergenerational increase in real market income than Baby Boomers. Intergenerational progress for Millennials under age 30 has remained robust as well, although their income growth largely results from higher reliance on their parents. We also find that the higher educational costs incurred by younger generations is far outweighed by their lifetime income gains.

The emphasis added is from me, not from the authors.  Via the excellent Kevin Lewis.

Non-binary gender economics

by  Tyler Cowen in

Economics research has largely overlooked non-binary individuals. We aim to jump-start the literature by providing data on several economically-important beliefs and preferences. Among many results, non-binary individuals report more gender-based discrimination and express different career and life aspirations, including less desire for children. Anti-non-binary sentiment is stronger than anti-LGBT sentiment, and strongest among men. Non-binary respondents report lower assertiveness than men and women, and their social preferences are similar to men’s and less prosocial than women’s, with age an important moderator. Elicited beliefs reveal inaccurate stereotypes as people often mistake the direction of group differences or exaggerate their size.

Here is the new NBER working paper by Katherine B. Coffman, Lucas C. Coffman, and Keith Marzilli Ericson.  P.s. comments are closed.

Which U.S. firms have grown in profitability?

by  Tyler Cowen in

China’s admission into the WTO in 2001 heralded a new era of globalization, increasing both import competition in domestic markets and foreign opportunities for US firms. In the aggregate, the average annual profitability of US public firms during the post globalization period (2003-2019) increased by 11.5% of the corresponding pre-globalization period (1984-2002) profitability. This increase in overall aggregate profitability was primarily driven by foreign profitability increasing by 47.4% for firms in the S&P 500 index, which are larger and have more intangible assets created by R&D and SG&A expenditures. In contrast, following globalization, the average aggregate domestic profitability of US firms remained flat, and firms employed more capital to generate sales. Firms with higher intangible assets benefited more from globalization.

That is from a new NBER working paper by Bullipe R. Chintha, Ravi Jagannathan, and Sri S. Sridhard.  When Average is Over was published about eleven years ago, in talks and media appearances I used to commonly draw a distinction between people/firms who are exporting their products — yes economists too — and those who are not.  Which category do you belong to?

U.S.A. yikes fact of the day

by  Tyler Cowen in

Between January 2016 and December 2022, the monthly antidepressant dispensing rate increased 66.3%, from 2575.9 to 4284.8. Before March 2020, this rate increased by 17.0 per month (95% confidence interval: 15.2 to 18.8). The COVID-19 outbreak was not associated with a level change but was associated with a slope increase of 10.8 per month (95% confidence interval: 4.9 to 16.7). The monthly antidepressant dispensing rate increased 63.5% faster from March 2020 onwards compared with beforehand. In subgroup analyses, this rate increased 129.6% and 56.5% faster from March 2020 onwards compared with beforehand among females aged 12 to 17 years and 18 to 25 years, respectively. In contrast, the outbreak was associated with a level decrease among males aged 12 to 17 years and was not associated with a level or slope change among males aged 18 to 25 years.

That is by Kao-Ping Chua, et.al., from the high-quality journal Pediatrics.  So that is how we respond to crises?  By doping up the young women?  Yikes!

Via the excellent Kevin Lewis.

Further data on alcohol use amongst American youth

by  Tyler Cowen in

This paper provides the first long-run assessment of adolescent alcohol control policies on later-life health and labor market outcomes. Our analysis exploits cross-state variation in the rollout of “Zero Tolerance” (ZT) Laws, which set strict alcohol limits for drivers under age 21 and led to sharp reductions in youth binge drinking. We adopt a difference-in-differences approach that combines information on state and year of birth to identify individuals exposed to the laws during adolescence and tracks the evolving impacts into middle age. We find that ZT Laws led to significant improvements in later-life health. Individuals exposed to the laws during adolescence were substantially less likely to suffer from cognitive and physical limitations in their 40s. The health effects are mirrored by improved labor market outcomes. These patterns cannot be attributed to changes in educational attainment or marriage. Instead, we find that affected cohorts were significantly less likely to drink heavily by middle age, suggesting an important role for adolescent initiation and habit-formation in affecting long-term substance use.

Here is the article by Tatiana Abboud, Andriana Bellou, and Joshua Lewis, via tekl once again.  People, you can make things easier for the political philosophers — why should they have to weigh liberty against utility?  Just give up drinking voluntarily.