Update on Rapid Antigen Tests

In July of 2020 in Frequent, Fast and Cheap is Better than Sensitive, I wrote:

A number of firms have developed cheap, paper-strip tests for coronavirus that report results at-home in about 15 minutes but they have yet to be approved for use by the FDA because the FDA appears to be demanding that all tests reach accuracy levels similar to the PCR test. This is another deadly FDA mistake.

It’s depressing that we are still moving so slowly on these issues but the media has finally gotten on board. Earlier I mentioned David Leonhardt’s article. Here is Margaret Hartmann in the New York Magazine.

In many Asian and European countries, at-home COVID-19 tests are cheap and easy to find in stores. CBS News reported this month that home antigen tests are now used routinely in the U.K., where they are free and “readily available at pretty much every pharmacy in the country.”

The situation is drastically different here because U.S. health officials focused on getting people vaccinated against COVID-19 and never leaned into asymptomatic testing as a strategy to fight the pandemic. While some foreign governments moved quickly to encourage screening and subsidize the cost of at-home tests, the Food and Drug Administration’s approval process moved much more slowly.

….The FDA said it needed to ensure that the tests were accurate, but many scientists countered that the agency was letting the perfect be the enemy of the good.

Note also that this is a way of saying that the politicians have now also had it with the FDA:

In addition to ramping up production of tests already on the market, the government is also working to speed up the approval process. On October 4, the FDA authorized Flowflex, an at-home antigen test produced by ACON Laboratories that is expected to retail for around $10 per test. And on October 25, the Department of Health and Human Services announced that the FDA will streamline its authorization process, and the National Institutes of Health will spend $70 million on a new program to “establish an accelerated pathway” to aid test makers seeking approval for their products.

Labor Market Participation Rates and Male Incarceration Rates

In our textbook, Modern Principles, Tyler and I write

Another factor that may be important in explaining the decline in the labor force participation rate of less-skilled men is the rise in mass incarceration. The male incarceration rate in the United States increased from 200 per 100,000 in 1970 to nearly 1,000 per 100,000 at is peak in 2007, as shown in Figure 30.18. Incarceration doesn’t reduce the labor force participation rate directly because the rate is measured as the ratio of the labor force to the adult non-institutionalized population. But what happens to prisoners when they are released? It’s difficult to get a job with an arrest record let alone a prison record. In fact, due to occupational licensing, it’s illegal for ex-felons to work in many industries. Approximately 7% of prime-aged men have been incarcerated. Thus, the rising incarceration rates of the past could be causing some of today’s low labor force participation rates.

A recent paper provides some evidence: Felony history and change in U.S. Employment rates, estimates that “a 1 percentage point increase in the share of a state’s adult population with a felony history is associated with 0.3 percentage point increase in non-employment (being unemployed or not in the labor force) among those aged 18 to 54.”

Sex Differences in Work Aspirations

Another important paper from Stoet and Geary

We investigated sex differences in 473,260 adolescents’ aspirations to work in things-oriented (e.g., mechanic), people-oriented (e.g., nurse), and STEM (e.g., mathematician) careers across 80 countries and economic regions using the 2018 Programme for International Student Assessment (PISA). We analyzed student career aspirations in combination with student achievement in mathematics, reading, and science, as well as parental occupations and family wealth. In each country and region, more boys than girls aspired to a things-oriented or STEM occupation and more girls than boys to a people-oriented occupation. These sex differences were larger in countries with a higher level of women’s empowerment. We explain this counter-intuitive finding through the indirect effect of wealth. Women’s empowerment is associated with relatively high levels of national wealth and this wealth allows more students to aspire to occupations they are intrinsically interested in. Implications for better understanding the sources of sex differences in career aspirations and associated policy are discussed.

…it has been four generations since Miner’s [10] assessment of adolescents’ occupational interests and core sex differences have not changed much, despite dramatic social and economic changes since that time. Boys continue to express a greater interest in blue-collar and white-collar things-oriented occupations than do girls, and girls continue to show a greater interest in people-oriented
occupations.

…Policy makers have regularly expressed a desire to reduce the number of students choosing stereotypical careers (e.g., [49]) or to increase the number of girls aspiring to and women entering technical occupations, especially STEM occupations [22]. The results of this study and related ones reveal a policy-relevant conundrum [3,4,6,50]. Generally speaking, more developed and gender equal nations are better than less developed nations in attracting boys to more established things-oriented (often blue-collar) occupations, but they fail to attract girls to these areas. This problem is also occurring for the subset of things-oriented STEM occupations. In fact, the problem for STEM is even more profound, given that interest in STEM declines for both boys and
girls in more developed, innovative, and gender equal nations.

See also my previous post Do Boys Have a Comparative Advantage in Math and Science?

Hat tip: Steve Stewart-Williams.

Clement and Tribe Predicted the FDA Catastrophe

Paul Clement and Laurence Tribe

Laboratory developed tests are not FDA regulated–never have been–instead the labs are regulated under the Clinical Laboratory Improvement Amendments (CLIA) as overseen by the CMS. Laboratory developed tests are the kind your doctor orders, they are a service not a product and are not sold directly to patients. Labs develop new tests routinely and they do not apply to the FDA for approval. Despite this long history, the FDA has claimed that it has the right to regulate lab tests and they have merely chosen not to exercise this right for forty years. In 2015, Paul Clement the former US Solicitor General under George W. Bush and Laurence Tribe, considered by many to be the leading constitutional lawyer in the United States, wrote an article that rejected the FDA’s claims writing that the “FDA’s assertion of authority over laboratory-developed testing services is clearly foreclosed by the FDA’s own authorizing statute” and “by the broader statutory context.”

Despite lacking statutory authority, the FDA has continued to claim it is authorized to regulate laboratory tests. Indeed, a key failure in the pandemic happened when the FDA issued so-called “guidance documents” saying that any SARS-CoV-II test had to be pre-approved by the FDA. Thus, the FDA reversed the logic of emergency. In ordinary times, pre-approval was not necessary but when speed was of the essence it became necessary to get FDA pre-approval. The FDA’s pre-approval process slowed down testing in the United States and it wasn’t until after the FDA lifted its restrictions in March that tests from the big labs became available.

Clement and Tribe rejected the FDA claims of regulatory authority over laboratory developed tests on historical, statutory, and legal grounds but they also argued that letting the FDA regulate laboratory tests was a dangerous idea. In a remarkably prescient passage, Clement and Tribe (2015, p. 18) warned:

The FDA approval process is protracted and not designed for the rapid clearance of tests. Many clinical laboratories track world trends regarding infectious diseases ranging from SARS to H1N1 and Avian Influenza. In these fast-moving, life-or-death situations, awaiting the development of manufactured test kits and the completion of FDA’s clearance procedures could entail potentially catastrophic delays, with disastrous consequences for patient care.

Clement and Tribe nailed it. Catastrophic delays, with disastrous consequences for patient care is exactly what happened.

Addendum: See also my pre-pandemic piece on this issue, Our DNA, Our Selves.

Ports and Tolerance

An elegant essay by Saumitra Jha on why tolerance between Hindus and Muslims evolved in India’s port cities.

Port Sea Monument Gateway Of India Mumbai Historic

[W]here do institutions of tolerance emerge? Combining the historical accounts, the fieldwork, and the data, it became clear that such institutions develop in very specific places, where two conditions were satisfied. First, Hindus and Muslims needed to have incentives to work together: for example, engaging in business relationships that complemented each other, rather than competed against one another. Second, this complementarity had to be robust: it had to be difficult for one group to replicate or simply steal the source of the others’ complementarity.

One important set of examples of these were ports—like Mahatma Gandhi’s own hometown, Porbandar—that had traded to the distant Middle East during the medieval period. For one month a year, for close to a thousand years, Mecca had been one of the largest markets in the world during the Hajj—and one had to be Muslim to go to Mecca. This gave Muslims in ports—in India, but also on the African coasts, the Malay peninsula, and beyond—a strong advantage in overseas trade and shipping. And, yet, this advantage nonetheless benefited the communities they connected by sail.

Further this complementarity in overseas trade came from a trading network that was intangible, and so impossible to seize, and the scale of the Hajj was so large it was impossible for a Hindu to replicate. Not surprisingly, then—before being disrupted by European colonial interventions beginning in the 16th century—Muslims had dominated overseas trade across the Indian Ocean, from the coasts of Zanzibar to India, Malaysia and beyond, as far as China.

Ports emerged at natural harbors along India’s medieval coasts to accommodate these trading relationships. These ports also witnessed not just the emergence of rules but also beliefs and organizations that supported trade, inter-group trust, and religious tolerance. So much so, that even three centuries later—after Muslim trade advantages had ended due to European colonial interventions, and many of the ports themselves had silted up and become inaccessible to trade—this legacy of beliefs, norms, and organizations continued to shape the way people interacted with one another. The institutions of peace and tolerance outlived the economic incentives that had once sustained them.

Photo Credit: MaxPixel.

The Price of COVID-19 Risk In A Public University

Wow. Duha Altindag, Samuel Cole and R. Alan Seals Jr, three professors in the economics department at Auburn University, study their own university’s COVID policies. The administration defied the Alabama Governor’s public health order on social distancing and created their own policy which caused enrollment in about half of the face-to-face classes to exceed legal limits. Professors assigned to teach these riskier classes were less powerful, albeit they were paid more to take on the risk. I am told that the administration is not happy. I hope the authors have tenure.

We study a “market” for occupational COVID-19 risk at Auburn University, a large public school in the US. The university’s practices in Spring 2021 caused approximately half of the face-to-face classes to have enrollments above the legal capacity allowed by state law, which followed CDC’s social distancing guidelines. Our results suggest that the politically less powerful instructors, such as graduate student teaching assistants and adjunct instructors, as well as women, were systematically recruited to deliver their courses in riskier classrooms. Using the dispersibility of each class as an instrument for classroom risk, our IV estimates obtained from hedonic wage regressions show that instructors who taught at least one risky class were paid more than those who exclusively taught safe courses. We estimate a COVID-19 risk premium of $8,400 per class.

Causal Inference at Twitter

Twitter engineering had a nice tweet thread on how they use econometrics and causal inference:

 You may have heard about this year’s Economics Nobel Prize winners – David Card, Josh Angrist (@metrics52) & Guido Imbens.

Their publicly available work has helped us solve tough problems @Twitter, and we’re excited to celebrate by sharing how their findings have inspired us. Understanding causal relationships is core to our work on identifying growth opportunities and measuring impact.

This year’s winners laid the foundation for cutting-edge techniques we use to understand where Twitter can improve and how changes affect our platform experience.
To share a few exciting causal inference applications at Twitter:

While online experimentation is helpful to understand the impact of a product change, it may not be the most efficient way to measure long-term impact. We built a causal estimation framework on the idea of statistical ‘surrogacy’ (Athey et al 2016) – when we can’t wait to observe long-run outcomes, we create a model based on intermediate data.

Estimating Treatment Effects using Multiple Surrogates: The Role of the Surrogate Score and the Surrogate Index

Estimating the long-term effects of treatments is of interest in many fields. A common challenge in estimating such treatment effects is that long-term outcomes are unobserved in the time frame needed. We combine this framework with our online experimentation platform to form a feedback/validation loop and to help accurately infer product success. One of the challenges we face is understanding the impact of different actions at Twitter (likes, Retweets etc.) Engagement actions often occur sequentially and at different surface areas. How to disentangle the effect of multiple actions presents many challenges.
We use Double Machine Learning to understand the causal impact of engagement actions.

Our work leverages research by Chernozhukov et al. (2018), and is influenced by Imbens & Rubin (2015).

Causal Inference for Statistics, Social, and Biomedical Sciences
This framework helps the team to interpret search experiments and make Twitter a better place to serve the public conversation. These applications promote a better understanding of tradeoffs among competing signals, helping our engineering team to iterate fast under more principled measurement and decision frameworks, making Twitter a better platform to create and share ideas and information.

We’re grateful for the role that academic research plays in driving innovation across society. We couldn’t do this work without the methodological foundation of the winners’ work and contributions across academia. Work like this inspires product innovation and engineering ideas alike, and we look forward to all that is yet to come.

More details on Twitter Data Science work will be introduced in our upcoming Engineering Blog posts.

Make TeleMedicine Permanent

One of the silver linings of the pandemic was the ability to see a doctor and be prescribed medicine online. I used telemedicine multiple times during the pandemic and it was great–telemedicine saved me at least an hour each visit and I think my medical care was as good as if I had been in person. I already knew I had poison ivy! No need for the doctor to get it also.

Telemedicine has been possible for a long time. What allowed it to take off during the pandemic wasn’t new technology but deregulation. HIPAA rules, for example, were waived for good faith use of standard communication technologies such as Zoom and Facetime even though these would ordinarily have been prohibited.

The Federal Ryan Haight Act was lifted which let physicians prescribe controlled substances (narcotics, depressants, stimulants, hallucinogens, and anabolic steroids) in a telemedicine appointment–prior to COVID an in-person appointment was required.

Prior to COVID Medicaid and Medicare wouldn’t pay for many services delivered over the internet. But during the pandemic the list of telemedicine approved services was expanded. Tennessee, for example, allowed speech therapists to bill for an online session. Alaska allowed mental health and counseling services and West Virginia allowed psychological testing to be delivered via telemedicine. Wisconsin allowed durable medical equipment such as prosthetics and orthotics to be prescribed without a face-to-face meeting.

Another very important lifting of regulation was allowing cross-state licensing which let out-of-state physicians have appointments with in-state patients (so long, of course, as the physicians were licensed in their state of residence.)

The kicker is that almost all of these changes are temporary. Regulatory burdens that were lifted for COVID will all be reinstated once the Public Health Emergency (PHE) expires. The PHE has been repeatedly extended but that will only push off the crux of the issue which is whether many of the innovations that we were forced to adopt during the pandemic shouldn’t be made permanent.

Working from home has worked better and been much more popular than anyone anticipated. Not everyone who was forced to work at home because of COVID wants to continue to work at home but many businesses are finding that allowing some work from home as an option is a valuable benefit they can offer their workers without a loss in productivity.

In the same way, many telemedicine innovations pioneered during the pandemic should remain as options. No one doubts that some medical services are better performed in-person nor that requiring in-person visits limits some types of fraud and abuse. Nevertheless, the goal should be to ensure quality by regulating the provider of medical services not regulating how they perform their services. Communications technology is improving at a record pace. We have moved from telephones to Facetime and soon will have even more sophisticated virtual presence technology that can be integrated with next generation Apple watches and Fitbits that gather medical information. We want medical care to build on the progress in other industries and not be bound to 19th and 20th century technology.

The growth of telemedicine is one of the few benefits of the pandemic. As the pandemic ends, let’s make this silver lining permanent.

Worrying Sentence(s) of the Day

NYTimes: An examination of hundreds of health departments around the country shows that the nation may be less prepared for the next pandemic than it was for the current one.

…State and local public health departments across the country have endured not only the public’s fury, but widespread staff defections, burnout, firings, unpredictable funding and a significant erosion in their authority to impose the health orders that were critical to America’s early response to the pandemic.

People have had it. Let’s hope we aren’t tested again soon.

Be Green: Buy a Coal Mine!

It’s time to reup the idea of buying coal mines and shuttering them. I wrote about this a few years ago based on Bard Harstad’s piece in the JPE and it came up again on twitter so I went looking for a coal mine to buy. Here’s a coal mine for sale in West Virginia for only $7.8 million! According to the ad, the mine produces 10,000 tons of coal monthly and has reserves of 8 million tons. Now here are some back of the envelope calculations.

(Warning: There may be errors since there are a lot of unit conversions. I invite someone with expertise in the industry to do a more serious analysis.)

Each ton of coal burned produces about 2.5 tons of carbon dioxide (you get more carbon dioxide since the carbon combines with oxygen). Sources: 2.86 short tons. 2.086 short tons.

It costs about $100 to sequester a ton of carbon dioxide for a long time.

Thus, 10,000 tons of coal burnt monthly produce 25,000 tons of carbon dioxide that it would cost $2.5 million a month to sequester. Or buying the mine pays for itself in reduced C02 emissions in about 3 months.

Ordinarily buying up the supply would increase supply elsewhere but coal mines are going out of business–thus no one is investing much in building new coal mines. The supply curve, therefore, is inelastic. In addition, you could buy up the right to mine in precisely those countries that are not committed to reducing coal mining. Indeed, you could buy the right to mine costly-to-exploit coal deposits–those deposits are cheap (since they are costly to exploit) and by taking them off the market you are making the supply curve even more inelastic so you aren’t encouraging much additional supply. Imagine, for example, that coal mining will be banned tomorrow. Thus, companies will be producing all-out today but that means you could reduce a lot of carbon emissions by buying the right to mine from the most expensive producers (who will sell cheap) and you won’t appreciably increase the incentive to mine. Indeed, on the margin, a higher price of energy might even do more to increase alternative sources of power like solar, especially if you buy thermal coal where there are lots of substitutes (there are fewer substitutes for coke coal.) See the Harstad paper and references in my earlier post.

Thus, buying a coal mine and leaving the coal in the ground looks like a cost-effective way of sequestering carbon dioxide.

Addendum: There are also some crazy “use it or lose it” laws that say that you can’t buy the right to extract a natural resource and not use it. When the high-bidder for an oil and gas lease near Arches National Park turned out to be an environmentalist the BLM cancelled the contract! That’s absurd. The high-bidder is the high-bidder and there should be no discrimination based on the reasons for the bid. See this Science piece.

H/t: Austin Vernon.

Hunting the Satanists

Michael Flynn, the former Trump National Security Advisor and QAnon promoter, is now being accused by QAnon of being a Satanist.

…Flynn’s trouble started on Sept. 17, when he led a congregation at Nebraska pastor Hank Kunneman’s Lord of Hosts Church in prayer. Flynn’s prayer included invocations to “sevenfold rays” and “legions,” two phrases that struck some of Flynn’s followers as strange.

…As video of the prayer circulated in online conspiracy theorist groups, the references to “legions” and “rays” soon sparked speculation among Flynn’s right-wing supporters that their hero had been lured to the dark side. Always on the lookout for the Satanic influence they imagine lurks at the heart of the world, they claimed that Flynn had secretly been worshiping the devil. Worse, since the congregation was repeating the prayer after Flynn, the rumor went, he had duped hundreds of Christians into joining the ritual.

…Flynn isn’t the first right-wing figure tied to QAnon to see its acolytes turn on him. Oklahoma Senate candidate Jackson Lahmeyer, whose challenge to Sen. James Lankford (R-OK) has been endorsed by Flynn, appeared at an April pro-QAnon conference with Flynn in Tulsa.

A few months later, however, Lahmeyer posted a seemingly innocent picture of his daughter wearing red shoes—apparently unaware that QAnon followers consider red shoes to be yet another sign of their imagined Satanic sex-trafficking cabal. Lahmeyer was soon caught up in a QAnon controversy of his own.

“Unfortunately, I have to say it because people are asking me,” Lahmeyer wrote in a Facebook post. “I’m in no way involved in Child Sex Trafficking, pedophilia or devil worship.”

Now, here’s another story–this one about an email sent by a Yale law student from the Native American Law Students Association (NALSA) to fellow classmates. The email in question reads:

SUP NALSA,

Hope you’re all still feeling social! This Friday at 7:30 we will be christening our very own (soon to be) world=renowned NALSA Trap House….by throwing a Constitution Day Bash in collaboration with FedSoc. Planned abstractions include Popeye’s chicken, basic-bitch-American-themed snacks (like apple pie, etc.), a cocktail station, assorted hard and soft beverages, and (most importantly) the opportunity to attend the NALSA Trap House’s inaugural mixer!

Hope to see you all there!

The email seems to me like a light-hearted invitation to a party but, of course, not being one-of-the-elect I can’t read the secret, esoteric meaning. According to Yale’s Diversity office the email was actually a coded message to celebrate white supremacy with a blackface party.

Just 12 hours after the email went out, the student was summoned to the law school’s Office of Student Affairs, which administrators said had received nine discrimination and harassment complaints about his message.

At a Sept. 16 meeting, which the student recorded and shared with the Washington Free Beacon, associate dean Ellen Cosgrove and diversity director Yaseen Eldik told the student that the word “trap” connotes crack use, hip hop, and blackface. Those “triggering associations,” Eldik said, were “compounded by the fried chicken reference,” which “is often used to undermine arguments that structural and systemic racism has contributed to racial health disparities in the U.S.”

Eldik, a former Obama White House official, went on to say that the student’s membership in the Federalist Society had “triggered” his peers.

…Throughout the Sept. 16 meeting and a subsequent conversation the next day, Eldik and Cosgrove hinted repeatedly that the student might face consequences if he didn’t apologize—including trouble with the bar exam’s “character and fitness” investigations, which Cosgrove could weigh in on as associate dean.

…When the student hadn’t apologized by the evening of Sept. 16, Eldik and Cosgrove emailed the entire second-year class about the incident. “[A]n invitation was recently circulated containing pejorative and racist language,” the email read. “We condemn this in the strongest possible terms” and “are working on addressing this.”

The two cases illustrate that the worldview of QAnon and Yale’s diversity office are surprisingly similar. Both see a world in which Satan, literal or metaphorical, is an active force in the world corrupting individuals and institutions. Satan is powerful but hidden. He only reveals his influence when the corrupted slip-up and by the incorrect use of a word, phrase, or gesture reveal their true natures.

Since Satan is powerful and hidden the good people must constantly monitor everyone. The moment a slip-up is spotted, no matter how small, the corrupted must be denounced because anyone who even unwittingly associates with the corrupted will themselves become corrupted. “Legions”and “rays”? Satanist! “Trap House.” Satanist! “Red shoes.” Sex-trafficker! “Federalist Society.” Satanist society! Repeating the prayer? Duping hundreds of Christians into joining the ritual! Attending a party? We condemn this in the strongest possible terms! Condemn the non-believers to HELL! It’s all the same.

The other similarity, of course, is that both views are disturbingly common and completely bonkers.

Photo Credit: Wikipedia.

Stripe v. Elrond! Crypto and the Payments System

Recently Elrond, the blockchain startup for which I am an advisor, bought a payments processor (conditional on approval from the Romanian government). On the same day, Stripe, the payments processor, announced that they are moving into crypto. None of this is coincidental. Elrond understand that the payments market is a multi-trillion dollar opportunity. Stripe knows that crypto innovation could undercut them very quickly if they aren’t prepared.

How did Stripe turn into a multi-billion dollar firm almost overnight? Obviously, Stripe is a great firm, led by the brilliant Collison brothers, CEO Patrick Collison and President John Collison. But it’s also important to understand that the payments market in the United States is a $100 trillion dollar market. Yes, $100 trillion. Any firm that captures even a small share of this market is going to be big. Credit cards are actually a small part of payments, about $7 trillion with roughly a 2% transaction fee or a $140 billion market. (Quick check. Credit card companies had 2020 revenues of $176 billion).  ACH debit and credit transfers are the big market, $65 trillion, which at a .5% transaction fee amounts to a $325 billion market (this is retail price, wholesale is lower). Thus, payments revenue is on the order of $465 billion. A small share of $465 billion is a very big market (and that is just the US market).

Now consider the following. Crypto payments are in principle at least an order of magnitude cheaper than ACH payments. On Elrond, for example, a very fast and low cost blockchain compared to Ethereum or Bitcoin, someone recently transferred $17.5 million for less than a penny. Moreover, crypto payments are global while every other payments system gets much more expensive as you cross borders. I recently sent $1500 to India and it cost me $100 in transaction fees! To be sure, payments made through the banking system have to obey “Know Your Customer” regulations and also include invoicing and billing services which adds both to value and cost. The main reason, however, that payments through the banking system are expensive is because the banking system rails are taped together with two hundred years of spit and duct tape.

Crypto payments are the future. Stripe knows it. Elrond knows it. The race is on.

The First Nobel Prize for Marginal Revolution University!

The Nobel Prize in economics this year goes to David Card, Joshua Angrist and Guido Imbens. I describe their contributions in greater detail in A Nobel Prize for the Credibility Revolution.

It’s also fun to note that Joshua Angrist mostly teaches at MIT but he also teaches a course on Mastering Econometrics at Marginal Revolution University so this is our first Nobel Prize! Here is Master Joshua on instrumental variables.

A Nobel Prize for the Credibility Revolution

The Nobel Prize goes to David Card, Joshua Angrist and Guido Imbens. If you seek their monuments look around you. Almost all of the empirical work in economics that you read in the popular press (and plenty that doesn’t make the popular press) is due to analyzing natural experiments using techniques such as difference in differences, instrumental variables and regression discontinuity. The techniques are powerful but the ideas behind them are also understandable by the person in the street which has given economists a tremendous advantage when talking with the public. Take, for example, the famous minimum wage study of Card and Krueger (1994) (and here). The study is well known because of its paradoxical finding that New Jersey’s increase in the minimum wage in 1992 didn’t reduce employment at fast food restaurants and may even have increased employment. But what really made the paper great was the clarity of the methods that Card and Krueger used to study the problem.

The obvious way to estimate the effect of the minimum wage is to look at the difference in employment in fast food restaurants before and after the law went into effect. But other things are changing through time so circa 1992 the standard approach was to “control for” other variables by also including in the statistical analysis factors such as the state of the economy. Include enough control variables, so the reasoning went, and you would uncover the true effect of the minimum wage. Card and Krueger did something different, they turned to a control group.

Pennsylvania didn’t pass a minimum wage law in 1992 but it’s close to New Jersey so Card and Kruger reasoned that whatever other factors were affecting New Jersey fast food restaurants would very likely also influence Pennsylvania fast food restaurants. The state of the economy, for example, would likely have a similar effect on demand for fast food in NJ as in PA as would say the weather. In fact, the argument extends to just about any other factor that one might imagine including demographics, changes in tastes and changes in supply costs. The standard approach circa 1992 of “controlling for” other variables requires, at the very least, that we know what other variables are important. But by using a control group, we don’t need to know what the other variables are only that whatever they are they are likely to influence NJ and PA fast food restaurants similarly. Put differently NJ and PA are similar so what happened in PA is a good estimate of what would have happened in NJ had NJ not passed the minimum wage.

Thus Card and Kruger estimated the effect of the minimum wage in New Jersey by calculating the difference in employment in NJ before and after the law and then subtracting the difference in employment in PA before and after the law. Hence the term difference in differences. By subtracting the PA difference (i.e. what would have happened in NJ if the law had not been passed) from the NJ difference (what actually happened) we are left with the effect of the minimum wage. Brilliant!

Yet by today’s standards, obvious! Indeed, it’s hard to understand that circa 1992 the idea of differences in differences was not common. Despite the fact that differences in differences was actually pioneered by the physician John Snow in his identification of the causes of cholera in the 1840 and 1850s! What seems obvious today was not so obvious to generations of economists who used other, less credible, techniques even when there was no technical barrier to using better methods.

Furthermore, it’s less appreciated but not less important that Card and Krueger went beyond the NJ-PA comparison. Maybe PA isn’t a good control for NJ. Ok, let’s try another control. Some fast food restaurants in NJ were paying more than the minimum wage even before the minimum wage went into effect. Since these restaurants were always paying more than the minimum wage the minimum wage law shouldn’t influence employment at these restaurants. But these high-wage fast-food restaurants should be influenced by other factors influencing the demand for and cost of fast food such as the state of the economy, input prices, demographics and so forth. Thus, Card and Krueger also calculated the effect of the minimum wage by subtracting the difference in employment in high wage restaurants (uninfluenced by the law) from the difference in employment in low-wage restaurants. Their results were similar to the NJ-PA comparison.

The importance of Card and Krueger (1994) was not the result (which continue to be debated) but that Card and Krueger revealed to economists that there were natural experiments with plausible treatment and control groups all around us, if only we had the creativity to see them. The last thirty years of empirical economics has been the result of economists opening their eyes to the natural experiments all around them.

Angrist and Krueger’s (1991) paper Does Compulsory School Attendance Affect Schooling and Earnings? Is one of the most beautiful in all of economics. It begins with a seemingly absurd strategy and yet in the light of a few pictures it convinces the reader that the strategy isn’t absurd but brilliant.

The problem is a classic one, how to estimate the effect of schooling on earnings? People with more schooling earn more but is this because of the schooling or is it because people who get more schooling have more ability? Angrist and Krueger’s strategy is to use the correlation between a student’s quarter of birth and their years of education to estimate the effect of schooling on earnings. What?! What could a student’s quarter of birth possibly have to do with how much education a student receives? Is this some weird kind of economic astrology?

Angrist and Krueger exploit two quirks of US education. The first quirk is that a child born in late December can start first grade earlier than a child, nearly the same age, who is born in early January. The second quirk is that for many decades an individual could quit school at age 16. Put these two quirks together and what you get is that people born in the fourth quarter are a little bit more likely to have a little bit more education than similar students born in the first quarter. Scott Cunningham’s excellent textbook on causal inference, The Mixtape, has a nice diagram:

Putting it all together what this means is that the random factor of quarter of birth is correlated with (months) of education. Who would think of such a thing? Not me. I’d scoff that you could pick up such a small effect in the data. But here come the pictures! Picture One (from a review paper, Angrist and Krueger 2001) shows quarter of birth and total education. What you see is that years of education are going up over time as it becomes more common for everyone to stay in school beyond age 16. But notice the saw tooth pattern. People who were born in the first quarter of the year get a little bit less education than people born in the fourth quarter! The difference is small, .1 or so of a year but it’s clear the difference is there.

Ok, now for the payoff.  Since quarter of birth is random it’s as if someone randomly assigned some students to get more education than other students—thus Angrist and Krueger are uncovering a random experiment in natural data. The next step then is to look and see how earnings vary with quarter of birth. Here’s the picture.

Crazy! But there it is plain as day. People who were born in the first quarter have slightly less education than people born in the fourth quarter (figure one) and people born in the first quarter have slightly lower earnings than people born in the fourth quarter (figure two). The effect on earnings is small, about 1%, but recall that quarter of birth only changes education by about .1 of a year so dividing the former by the latter gives an estimate that implies an extra year of education increases earnings by a healthy 10%.

Lots more could be said here. Can we be sure that quarter of birth is random? It seems random but other researchers have found correlations between quarter of birth and schizophrenia, autism and IQ perhaps due to sunlight or food-availability effects. These effects are very small but remember so is the influence of quarter of birth on earnings so a small effect can still bias the results. Is quarter of birth as random as a random number generator? Maybe not! Such is the progress of science.

As with Card and Kruger the innovation in this paper was not the result but the method. Open your eyes, be creative, uncover the natural experiments that abound–this was the lesson of the credibility revolution.

Guido Imbens of Stanford (grew up in the Netherlands) has been less involved in clever studies of empirical phenomena but rather in developing the theoretical framework. The key papers are Angrist and Imbens (1994), Identification and Estimation of Local Treatment Effects and Angrist, Imbens and Rubin, Identification of Causal Effects Using Instrumental Variables which answers the question: When we use an instrumental variable what exactly is it that we are measuring? In a study of the flu, for example, some doctors were randomly reminded/encouraged to offer their patients the flu shot.  We can use the randomization as an instrumental variable to measure the effect of the flu shot. But note, some patients will always get a flu shot (say the elderly). Some patients will never get a flu shot (say the young). So what we are really measuring is not the effect of the flu shot on everyone (the average treatment effect) but rather on the subset of patients who got the flu shot because their doctor was encouraged–that latter effect is known as the local average treatment effect. It’s the treatment effect for those who are influenced by the instrument (the random encouragement) which is not necessarily the same as the effect of the flu shot on groups of people who were not influenced by the instrument.

By the way, Imbens is married to Susan Athey, herself a potential Nobel Prize winner. Imbens-Athey have many joint papers bringing causal inference and machine learning together. The Akerlof-Yellen of the new generation. Talk about assortative matching. Angrist, by the way, was the best man at the wedding!

A very worthy trio.

The Future is Getting Farther Away

In Launching the Innovation Renaissance I said that “If total factor productivity had continued to grow at its 1957 to 1973 rate then we today would be living in the world of 2076 rather than in the world of 2014.” Sadly, the future is continuing to recede. Consider the graph below. If growth had continued at the rate expected by the CBO in 2005 then we today would be living in the world of 2037 rather than in the world of 2021. (n.b. I am eyeballing.)

By the way, don’t blame the forecasters. The forecast was reasonable, the reality is below expectation.

Hat tip: Matt Yglesias reupping a graph originally produced by Claudia Sahm who I thought had a different interpretation but maybe not!.