Category: Law

Facial recognition isn’t just about China and airports

The child labor activist, who works for Indian NGO Bachpan Bachao Andolan, had launched a pilot program 15 months prior to match a police database containing photos of all of India’s missing children with another one comprising shots of all the minors living in the country’s child care institutions.

He had just found out the results. “We were able to match 10,561 missing children with those living in institutions,” he told CNN. “They are currently in the process of being reunited with their families.” Most of them were victims of trafficking, forced to work in the fields, in garment factories or in brothels, according to Ribhu.

This momentous undertaking was made possible by facial recognition technology provided by New Delhi’s police. “There are over 300,000 missing children in India and over 100,000 living in institutions,” he explained. “We couldn’t possibly have matched them all manually.”

Locating thousands of missing children is just one of the challenges faced by India’s overstretched police force in a nation of 1.37 billion people.

In spite of these practical benefits, I still do not favor facial recognition systems at the macro level.  India seems to be planning a big one:

…India’s government now has a much more ambitious plan. It wants to construct one of the world’s largest facial recognition systems. The project envisions a future in which police from across the country’s 29 states and seven union territories would have access to a single, centralized database.

Here is the full article with much more detail about the plans.

Harvard sentences to ponder

We show that Harvard encourages applications from many students who effectively have no chance of being admitted, and that this is particularly true for African Americans.

Here is the whole abstract, by Peter Arcidiacono, Josh Kinsler, and Tyler Ransom:

Over the past 20 years, elite colleges in the US have seen dramatic increases in applications. We provide context for part of this trend using detailed data on Harvard University that was unsealed as part of the SFFA v. Harvard lawsuit. We show that Harvard encourages applications from many students who effectively have no chance of being admitted, and that this is particularly true for African Americans. African American applications soared beginning with the Class of 2009, with the increase driven by those with lower SAT scores. Yet there was little change in the share of admits who were African American. We show that this change in applicant behavior resulted in substantial convergence in the overall admissions rates across races yet no change in the large cross-race differences in admissions rates for high-SAT applicants.

And from the paper’s conclusion:

If the goal of recruiting African Americans is not simply to increase the diversity of matriculants, but also to achieve racial balance in the admit pool and/or racial balance in admit rates, then the policy could be deemed a success. As an example, admit rates for African American applicants were twice as large as admit rates for Asian American applicants in 2000, but by 2017 were approximately the same. Why Harvard might careabout the racial distribution of admit rates and applicants is not obvious. What is clear is that each year there are a significant number of African American high school students who have a potentially false impression about their chances of being admitted to Harvard.

Here is the full paper.  And here is a recent paper by Howell, Hurwitz, and Smith, with related results.

The effect of district attourneys on criminal justice outcomes

In the United States, elected district attorneys’ offices prosecute over 85% of all felony cases, but we know little about their effect on local criminal justice outcomes. Using a newly-collected dataset of district attorney elections, I show that Republican district attorneys lead to a 18-21% increase in new prison admissions in the two years following their election, while nonwhite district attorneys lead to a 10% decline. In both cases, there are no significant effects on local crime or arrest rates. These results show that the identity of the local district attorney is an important determinant of incarceration rates.

Here is the paper, by Sam Krumholz, on the job market this year from UCSD, that is not his job market paper, here is his full portfolio, public economics and law and economics, to me one of the more interesting candidates this year.

Opioids and labor market participation

The onset of the opioid crisis coincided with the beginning of nearly 15 years of declining labor force participation in the US. Furthermore, the areas most affected by the crisis have generally experienced the worst deteriorations in labor market conditions. Despite these time series and cross-sectional correlations, there is little agreement on the causal effect of opioids on labor market outcomes. I provide new evidence on this question by leveraging a natural experiment which sharply decreased the supply of hydrocodone, one of the most commonly prescribed opioids in the US. I identify the causal impact of this decrease by exploiting pre-existing variation in the extent to which different types of opioids were prescribed across geographies to compare areas more and less exposed to the treatment over time. I find that areas with larger reductions in opioid prescribing experienced relative improvements in employment-to-population ratios, driven primarily by an increase in labor force participation. The regression estimates indicate that a 10 percent decrease in hydrocodone prescriptions increased the employment-to-population ratio by about 0.7 percent. These findings suggest that policies which reduce opioid misuse may also have positive spillovers on the labor market.

That is from a job market paper by David Beheshti at the University of Texas at Austin.

Economists and non-economists on elasticity

From a recent paper by Joanna Venator and Jason Fletcher:

In this paper, we estimate the impacts of abortion clinic closures on access to clinics in terms of distance and congestion, abortion rates, and birth rates. Legislation regulating abortion providers enacted in Wisconsin in 2011-2013 ultimately led to the closure of two of five abortion clinics in Wisconsin, increasing the average distance to the nearest clinic to 55 miles and distance to some counties to over 100 miles. We use a difference-in-differences design to estimate the effect of change in distance to the nearest clinic on birth and abortion rates, using within-county variation across time in distance to identify the effect. We find that a hundred-mile increase in distance to the nearest clinic is associated with 25 percent fewer abortions and 4 percent more births. We see no significant effect of increased congestion at remaining clinics on abortion rates. We find significant racial disparities in who is most affected by abortion clinic closures, with increases in distance increasing birth rates significantly more for Black, Asian, and Hispanic women. Our results suggest that even small numbers of clinic closures can result in significant restrictions to abortion access of similar magnitude to those seen in Texas when a greater number of clinics closed their doors.

There are (at least) two possible responses to such results, and that is without even getting into one’s underlying view of the ethics of abortion.  One is to say that a great deprivation has occurred because many fewer women end up having abortions.  Another response is to infer that the marginal value of the abortions could not have been so high to begin with, if the number drops off so readily.

The same issue comes up with Obamacare.  If the price of health insurance goes up, quite a large number of people decide to go without coverage.  Is the size of that number a measure of the human tragedy resulting from the price increase?  Or is it a measure of how little those people actually value health insurance?  Or somehow both?

I have yet to meet a person who can think through these issues rationally and absorb what is interesting and valid in each of those two perspectives.

Release Bad News on a Friday

Politicians have long known to release bad news on a Friday and it appears that pharmaceutical firms may do likewise.

Safety alerts are announcements made by health regulators warning patients and doctors about new drug-related side effects. However, not all safety alerts are equally effective. We provide evidence that the day of the week on which the safety alerts are announced explains differences in safety alert impact. Specifically, we show that safety alerts announced on Fridays are less broadly diffused: they are shared 34% less on social media, mentioned in 23% to 66% fewer news articles, and are 12% to 51% less likely to receive any news coverage at all. As a consequence of this, we propose Friday alerts are less effective in reducing drug-related side effects. We find that moving a Friday alert to any other weekday would reduce all drug-related side effects by 9% to 12%, serious drug-related complications by 6% to 15%, and drug-related deaths by 22% to 36%. This problem is particularly important because Friday was the most frequent weekday for safety alert announcements from 1999 to 2016. We show that this greater prevalence of Friday alerts might not be random: firms that lobbied the U.S. Food and Drug Administration in the past are 49% to 56% more likely to have safety alerts announced on Fridays.

From The Friday Effect in Management Science by Diestre, Barber and Santalo.

Hat tip: Kevin Lewis.

The impact of the Affordable Care Act

The Impact of the Affordable Care Act: Evidence from California’s Hospital Sector
(with Mark Duggan and Atul Gupta) R&R, AEJ: Economic Policy

The Affordable Care Act (ACA) authorized the largest expansion of public health insurance in the U.S. since the mid-1960s. We exploit ACA-induced changes in the discontinuity in coverage at age 65 using a regression discontinuity based design to examine effects of the expansion on health insurance coverage, hospital use, and patient health. We then link these changes to effects on hospital finances. We show that a substantial share of the federally-funded Medicaid expansion substituted for existing locally-funded safety net programs. Despite this offset, the expansion produced a substantial increase in hospital revenue and profitability, with larger gains for government hospitals. On the benefits side, we do not detect significant improvements in patient health, although the expansion led to substantially greater hospital and emergency room use, and a reallocation of care from public to private and better-quality hospitals.

That is the job market paper by Emilie Jackson of Stanford.

The Long-Term Effects of California’s 2004 Paid Family Leave Act on Women’s Careers

It is not obviously a winner policy, at least not from the point of view of boosting women’s labor market opportunities.  In fact it seems to harm them:

This paper uses IRS tax data to evaluate the short- and long-term effects of California’s 2004 Paid Family Leave Act (PFLA) on women’s careers. Our research design exploits the increased availability of paid leave for women giving birth in the third quarter of 2004 (just after PFLA was implemented). These mothers were 18 percentage points more likely to use paid leave but otherwise identical to multiple comparison groups in pre-birth demographic, marital, and work characteristics. We find little evidence that PFLA increased women’s employment, wage earnings, or attachment to employers. For new mothers, taking up PFLA reduced employment by 7 percent and lowered annual wages by 8 percent six to ten years after giving birth. Overall, PFLA tended to reduce the number of children born and, by decreasing mothers’ time at work, increase time spent with children.

That is from a new NBER working paper by Martha J. Bailey, Tanya S. Byker, Elena Patel, and Shanthi Ramnath.  And are you wondering why the number of children falls as a response?  Because the mother ends up staying home with them?  Or do mothers invest more in child quality, thereby lower quantity, as in a Becker model?  In any case a good question.

The Mexican drug war and its intensification

Event study estimates suggest that cartel presence increases substantially after 2010 in municipalities well suited to grow opium poppy. Homicide rates increase along with the number of active cartels per municipality, with higher increases when a second, third, fourth and fifth cartel become active in the territory. These results suggest that some of the increase in violence that Mexico experienced in the last fifteen years could be attribute to criminal groups fighting for market shares of heroin and not only to changes in government enforcement.

That is from a recent job market paper by Fernanda Sobrino of Princeton University.

The economics of California power blackouts

Once again there is a risk of fire so they are turning off the power in many parts of dry and windy northern California, for 2.7 million people.  From The New York Times:

“When you turn the lights out on 3 million people because you have to keep the power lines safe then there’s no reason you should be allowed to continue,” Mr. Court said.

Michael Lewis, PG&E’s senior vice president of electric operations, said the issue was safety.

“We would only take this decision for one reason — to help reduce catastrophic wildfire risk to our customers and communities,” Mr. Lewis said in a statement.

PG&E filed for bankruptcy in January after amassing tens of billions of dollars in liability related to two dozen wildfires in recent years. As speculation grew that its equipment might be the cause of the Kincade Fire, its stock price plummeted about 30 percent on Friday to $5.08, a small fraction of its 52-week high of $49.42.

I would think the market expectation is that if PG&E is allowed to continue, as is likely to be the case, that it slowly will claw its way back to profitability, given that this is a highly regulated sector with barriers to entry.  So the company is afraid of losing its expected remaining profit from further liability, and thus it plays it safe with power, too safe because they don’t suffer so much from the power blackouts.  Sadly, the retail customers do not have many other options.

One solution would be to remove the liability the company faces from the fires, or alternatively you could add a liability option of set of fines from power cuts (call them “breach of contract”).  Both changes would introduce greater symmetry into the liability equation, but of course the former would eliminate the incentives for fire safety and fire reduction and the latter might bankrupt the company or create unenforceable or undefinable legal obligations.  Still, it hardly seems the current arrangement can be first best.

How about raising rates?  And then spending more on capital improvements?  (do read the tweets behind that link):

And in those proceedings, there is an independent division of the CPUC (the ‘Office of Ratepayer Advocates’) that has typically argued against maintenance and safety expenditures, so that rates can be kept low

How about raising rates a lot?  But maybe it is too late for that.

Another option, which I do not feel I have enough information to assess, is to have the state government buy out the power company.  That is not usually a good idea but in this case there is at least a chance it could lead to superior incentives.  The resulting company would then be geared toward pleasing voters, hardly an ideal arrangement but possibly better than the current incentives toward excess safety and massive power cuts with no real chance of consumer backlash.  With government ownership, how would the state internalize the liability risk?  How much would state borrowing rates rise?

Have you seen good proposals for improving the incentives in this rather disastrous matter?

Middlemen, China story of the day

Five hitmen have been jailed for attempted murder, after each one avoided carrying out the contract themselves so they could make a profit.

Chinese businessman Tan Youhui was looking for a hitman to take out a competitor, Wei Mou, and was willing to pay 2 million yuan (£218,000) to get the job done.

The hitman that Mr Youhui hired, decided to offer the job to another hitman, for half the original price.

The second hitman then subcontracted to another hitman, who then subcontracted to a fourth, who gave the job to a fifth.

However, hitman number five was so incensed at how much the value of the contract had fallen, that he told the target to fake his own death, which eventually led to the police finding out about the plot, Beijing News reported.

Here is the full story, via Yana.

My Conversation with Henry Farrell

An excellent episode, here is the audio and transcript.  We ranged far and wide, starting with Huawei and weaponized interdependence, moving later to the Facebook supreme court, Karl Polanyi, Ireland, and Gene Wolfe and Philip K. Dick.  Here is one excerpt:

COWEN: Arguably, dominant firms are easier to regulate. And since you seem to favor some kinds of additional regulation on the major tech companies, does this mean we’re too worried about monopoly, that we actually want to keep around a few dominant firms, and that if we split them up into many small parts, there would be more chaos or more fake news or more privacy violations?

If some parts of what they do are bad, and you get more competition in the bad, don’t we just want to put in GDPR barriers to entry, not quite public utilities, but keep them big and fat and happy and somewhat not so dynamic, yes or no?

FARRELL: It depends on what you value.

COWEN: But what you value.

FARRELL: Yeah. Let me put the tradeoff to you this way. If you value security, if the highlight is on security, then the answer is, you probably want to keep big companies around because you’re going to want to impose broad standards. You’re going to want to create collective security goods, and the only actors that can really do that in a substantial way are big businesses of one sort or another.

If, alternatively, you value things like privacy and other kinds of rights, then you probably want to move towards an equilibrium in which there are far, far fewer big firms. So that’s where I see the fight being played out. I see the fight being played out between people who value security and people who value privacy. I think they point in somewhat different directions.

COWEN: And where are you on that spectrum?

FARRELL: Well, it depends on the time of the day, and I find myself —

COWEN: It is 2:22 p.m.

And:

FARRELL: Well, I guess the question for me is — and again, this is a wide open question because we simply don’t have enough good empirical research — but what is the relationship and the broader ecology between companies like 8chan and companies like Facebook? I suspect that companies like 8chan will be far, far less successful if there weren’t much bigger platforms like Facebook that they could effectively grow upon.

So here are the arguments, something as follows. If you think about 8chan, and if you think about 4chan before it, they were basically meme factories. They were basically these places where these bored individuals hung out. You also created these memes in a kind of process of frenzied Darwinian evolution, where you desperately want to make sure that whatever you had said was on the front page because otherwise it would disappear forever. So you’ve got this survival-of-the-fittest thing, where incredibly valuable or incredibly effective memes go out and begin to populate the entire space.

But you need two things for that to work. First of all, you need a process of generation, and secondly, you need some kind of process of dissemination. You need other platforms which have far greater reach, which can then allow for these memes to propagate through the atmosphere.

I suspect that if we were in a world in which everything was at the scale of 8chan, rather than having a mixture of companies at the scale of 8chan and companies at the scale of Facebook, that the likelihood of this stuff spreading and becoming epidemic across the entire community of internet users would be far, far less. Obviously, we would have other problems then. But I think that the problems that we would face would be a very, very different set of problems from the problems that we face in the current environment.

Finally:

FARRELL: Yes. [Gene] Wolfe misleads us systematically, and clearly Severian is not a reliable narrator, but then neither is Proust’s narrator either. I think that if you really want to understand where Wolfe comes from, it really is Proust. His writing style is Proustian. His concern with time, with how it is that time works, is quintessentially Proustian.

And you don’t look to Wolfe any more than you look to other science fiction for characterization. I don’t think that’s the particular strength. What you do look for is a kind of a sense of the world. And in Wolfe, in particular, he provides this real understanding of how it is that the workings of society, and interestingly, conservative understanding of the workings of society.

I think of him almost as being Proust in reverse. Proust is describing a world in which the modern world is overtaking aristocracy. And that clearly is one of the great problems of Proust, what is happening on the social level. You have all of these aristocratic understandings: the Merovingian, all of these histories, all of these castles, all of this wonderful art, and they are being replaced by the modern world with its telephones, with its electric lighting, and so on.

And how do you think about this? How would you try to preserve what was happening in the past? What Wolfe does, which I think is an extraordinarily interesting thing, which would be impossible for anybody who is not a science fiction writer, is to take that and to reverse this and to imagine a world in which modernity has disappeared.

Recommended!

What should I ask Shaka Senghor?

I will be having a Conversation with Shaka, no associated public event.  So what should I ask him?  Here is the main part of his Wikipedia page:

Shaka Senghor is director’s fellow of the MIT Media Lab, college lecturer, author, and was convicted of murder in American courts. As of October 2015, he also teaches a class as part of the Atonement Project, a partnership between him, the University of Michigan, and the MIT Media Lab. His memoir, Writing my Wrongs, was published in March 2016. Senghor was named to Oprah’s SuperSoul 100 list of visionaries and influential leaders in 2016.

And here is Shaka’s home page.  I thank you all in advance for your suggestions.

Ranking states by their degree of regulation

Now Mercatus has completed an analysis of state-level regulation. State RegData 1.0 analyzes the administrative codes of 46 states plus the District of Columbia, and the results are informative. The average state has 131,000 restrictive terms and about 9 million words in its code, which would require roughly twelve work weeks to read at a normal reading pace.

But there is huge variation. The least regulated state is South Dakota, with about 44,000 regulatory restrictions, while the most regulated state is California, with 395,000. All told, the least regulated states are South Dakota, Alaska, Montana, Idaho, and North Dakota, while the most regulated states are California, New York, Illinois, Ohio, and Texas.

Unfortunately, we weren’t able to include Vermont, New Jersey, Arkansas, and Hawaii. Arkansas is the easiest to explain: It has no administrative code, at least not yet. Its state agencies produce regulations, but until this year no one had ever bothered to compile them in one place. Fortunately — perhaps partly in response to RegData — legislation was passed this year to create an official Code of Arkansas Rules by January 2023, so Arkansas’s regulatory landscape will eventually come to light.

That is from James Broughel and Patrick A. McLaughlin, there is more at the link.