Monday, 28 January 2019

How re-usable shopping bags are like a regressive tax

A few weeks back, the New Zealand Herald reported:
As of July 1, retailers will no longer be able to sell or give away single-use plastic shopping bags.
Instead, people will have to purchase reusable bags that are more than 70 microns thick.
This follows Cabinet's decision last year to go ahead with a mandatory nationwide phase-out of single-use bags...
A regulatory impact assessment of the ban, written by the Ministry for the Environment, warned there was a risk it could impact on the poor the hardest.
"Requiring consumers to pay up-front for new multi-use shopping bags could disproportionately affect lower-income consumers," the assessment said.
A tax is described as regressive if lower income consumers (or households) pay a higher proportion of their income in the tax. In this case, assuming that higher-income and lower-income households use the same number of shopping bags (or at least, that shopping bag use increases with income, but at a diminishing rate, which seems reasonable [*]), then lower-income households will end up paying more of their income in shopping bag costs. Even though the bag cost is not strictly a tax, it is still a regressive cost imposed disproportionately (relative to income) on lower-income households by the government. The shopping bag cost is regressive.

MBIE proposed this solution, according to the article:
"This could be mitigated by retailers allowing consumers who have Gold Cards or Community Service Cards a discount or exceptions."
The ministry also proposed partnering with food banks and different donors to distribute free multi-use bags with food parcels.
That would ensure that most low-income households won't be paying a higher proportion of their income on shopping bags than high-income households. However, does that also open a possibility for black market shopping bag arbitrage?


[*] Higher-income households might use more shopping bags because they regularly buy more shopping items, or because they are less careful with the bags they do have (because the cost of replacing the bags is less consequential to them). However, it seems unlikely that households with double the income would use twice as many shopping bags (they would probably use less than twice as many), hence the 'increase at a diminishing rate).

Tuesday, 22 January 2019

The impact of an online minimum wage

Yesterday, I wrote about the latest research on the Seattle minimum wage. The post was based on this NBER working paper by Jardim et al. (also ungated here). One thing I forgot to mention about that paper by is that it has an excellent discussion of the problems for measuring impacts of minimum wage changes, including:
Starting in any baseline period, some workers may witness natural wage gains that render the increased minimum irrelevant upon phase-in. Others will disappear from the data entirely, and it isn’t clear whether these departures should be coded as transitions to non-employment or to employment not tracked by one state’s [unemployment insurance] system. And using an early baseline quarter causes the analysis to overlook a key segment of the low-wage labor market: those who have only recently entered it. For these reasons analysis of a cohort tracked from too early a point in time may reveal very small impacts on wages, employment, and earnings.
One of the key results from that paper was that the minimum wage benefited only the more experienced workers, while less experienced workers were the same or worse off. Presumably, the more experienced workers are also likely to be the most productive, so are worthwhile for employers to continue to hire even at the higher minimum wage. Interestingly, this result is corroborated by another recent paper in a completely different context, by John Horton (New York University). Horton looked at the impact of imposing a minimum wage experimentally in an online job market:
The experiment in this paper was conducted in a large online labor market. In this market, a would-be employer writes job descriptions, labels the job opening with a category (e.g., “Administrative Support”), lists required skills, and then posts the job opening to the platform website. Workers generally learn about job openings via electronic searches. Workers submit applications, which generally include a wage bid (for hourly jobs) or a total project bid (for fixed-price jobs) and a cover letter... After a worker submits an application, the employer screens his or her applicants and can decide to make an offer or offers.
Employers on the platform were randomly allocated to different minimum wages (or no minimum wage), and were not aware they were in an experiment. When workers submitted their bids, if their bid was below the minimum wage (for treated employers) they were instructed to increase their bid (the workers were also not aware there was an experiment going on). Horton found that:
Imposing a minimum wage raised the wages of hired workers, but this imposition also reduced hiring, albeit not by very much. In contrast, hours-worked fell sharply, with reductions as large as 30% in some sub-populations of job openings expected to pay low wages. Large reductions in hours-worked occurred even in sub-populations that saw no reduction in hiring. Presumably some of the reduction in hours-worked was caused by employers economizing on labor, and perhaps from improved worker morale. However, hours-worked also likely fell because treated employers hired substantially more productive workers, with productivity measured by pre-experiment worker attributes.
Productivity was measured by the average past wage rate of the worker on the platform, but was robust to some alternative measures. Overall, the results suggest that increased minimum wages increase the earnings of those workers who continue to be employed, and reduces hours worked. Those are both pretty standard explanations from a supply-and-demand model of the labour market. That employers concentrate hiring on higher-productivity workers is a little more surprising (why wouldn't the employers try to do that already?), but not if you consider that those high-productivity workers might already be working their preferred number of hours and are only be willing to work more if wages are higher.

In any case, this latest evidence gives more food for thought in terms of the trade-offs associated with having a higher minimum wage.

[HT: Marginal Revolution, back in July]

Read more:

  • Seattle's minimum wage, revisited
  • The latest evidence supports negative employment effects of the minimum wage
  • Monday, 21 January 2019

    Seattle's minimum wage, revisited

    As I noted in a post back in 2017, the initial research on Seattle's minimum wage showed that, although workers earned more per hour, they worked fewer hours and so the net impact on earnings was basically zero. That was based on this NBER Working Paper (ungated here) by Ekaterina Jardim (University of Washington) and others. Now, Jardim et al. have re-visited the effects of the minimum wage in a new NBER working paper (which appears to be ungated, but just in case it is also available here), as the New York Times reported back in October:
    Seattle increased its minimum wage for large employers to $11 an hour, from $9.47, in April 2015, then to $13 for many of those same employers in January 2016. (The minimum wage increased by less for small employers, and for large employers that contributed toward workers’ health coverage.)
    In their latest paper, which has not been formally peer reviewed, Mr. Vigdor and his colleagues considered how the minimum-wage increases affected three broad groups: People in low-wage jobs who worked the most during the nine months leading up to and including the quarter in which the increase took effect (more than about 600 or 700 hours, depending on the year); people who worked less during that nine-month period (fewer than 600 or 700 hours); and people who didn’t work at all and hadn’t during several previous years, but might later work. The latter were potential “new entrants” to the ranks of the employed, in the authors’ words.
    Usually, I'm not in favour of sub-group analysis, because too many researchers use it as a fishing expedition to try and find something to say, when their initial analyses show up nothing (or when the results contradict the researchers' priors). In this case, the sub-groups make some sense. Experienced workers (the first group, who worked a lot) are likely to be more productive, and so after the minimum wage increase, you'd expect employers to be more willing to keep them on. The second group is inexperienced (or at least less experienced), so you would expect them to be less affected than the first group. And the third group you would expect to be affected most of all. And indeed, that is what Jardim et al. found, as the New York Times reported:
    The workers who worked the most ahead of the minimum-wage increase appeared to do the best. They saw a significant increase in their wages and only a small percentage decrease in their hours, leading to a healthy bump in overall pay — an average of $84 a month for the nine months that followed the 2016 minimum-wage increase.
    The workers who worked less in the months before the minimum-wage increase saw almost no improvement in overall pay — $4 a month on average over the same period, although the result was not statistically significant. While their hourly wage increased, their hours fell substantially. (That doesn’t mean they were no better off, however. Earning roughly the same wage while working fewer hours is a trade most workers would accept.)
    It’s the final group of workers — the potential new entrants who were not employed at the time of the first minimum-wage increase — that Mr. Vigdor and his colleagues believe fared the worst. They note that, at the time of the first increase, the growth rate in new workers in Seattle making less than $15 an hour flattened out and was lagging behind the growth rate in new workers making less than $15 outside Seattle’s county. This suggests that the minimum wage had priced some workers out of the labor market, according to the authors.
    “For folks trying to get a job with no prior experience, it might have been worth hiring and training them when the going rate for them was $10 an hour,” Mr. Vigdor speculated, but perhaps not at $13 an hour.
    However, that isn't the end of the story. The paper of course has more detail:
    ...workers initially employed at low wages in Seattle enjoyed significantly more rapid hourly wage growth over the quarters following both the first and second minimum wage increases in April 2015 and January 2016. While these workers experienced a modest reduction in their hours worked, on net their pretax earnings increased an average of around $10 per week....
    Essentially all of the earnings increases accrue to the more experienced half of the low-wage workforce. The less experienced half saw larger proportionate decreases in hours worked, which we estimate to have fully offset their gain in wages, leaving no significant change in earnings. More experienced workers were also more likely to supplement their Seattle income by adding hours outside the city. 
    So the average gain in earnings was just $10 per week, and the gains were concentrated among the more experienced workers, while less experienced workers saw no change in income. And, to top it off, the more experienced workers' gain in earnings partly arose because they were more likely to work some hours outside the area where the minimum wage applied. So the minimum wage reduced their working hours but forced many of them to work a second job. That doesn't sound like a resounding endorsement of the increased minimum wage, unless the aim is to incentivise employers to shift their workforce to more experienced (and probably more productive) workers.

    [HT: Marginal Revolution, back in October]

    Read more:

    Sunday, 20 January 2019

    It's January, so that means rising house rents are in the news

    There are few certainties in life: death, taxes, and January news items about rent increases (as I've highlighted last year, and in 2017, 2016, and 2015. I'm not sure I have anything new to say on the topic, but then again, neither do the media. So, here's this year's article from the New Zealand Herald, about Auckland and Wellington (see also this article as well):
    An Auckland agent says more prospective tenants are flocking to house viewings than ever before as a photo posted to social media showed up to 50 people snaking their way through a Remuera rental this week.
    One city tenant, aged in his 60s, said he and his wife were shocked when the first home they viewed - a three-bedroom Greenlane house with a rent of $830 per week - also had 50 people present.
    The man - who only wanted to be known as Chris - had been keen to move out of his current rental because the landlord tried to bump the rent up by $20 per week.
    But after seeing so many people in Greenlane, Chris and his wife - who own a property in Australia - decided to stay in their current rental and negotiate a lesser rent increase of $10 per week.
    I have just two words to say: excess demand. Ok, I lied. I have more than two words. Excess demand happens when the market price is below the market-clearing equilibrium price. So, if you have lots of people looking for houses to rent, and not enough houses to go around, it's because the rents are too low. Normally, we would expect markets to adjust, but for some reason the rental market seems to be perpetually in a state of excess demand. In a 2016 post, I pondered whether efficiency rents might be an explanation for the persistently low rents:
    If the landlord instead offered an efficiency rent (a rent below the equilibrium market rent), then they would have many potential tenants applying for the property, allowing the landlord to pick the best (the least likely to damage the property). It also gives the tenants an incentive to look after the property after signing the tenancy agreement, because if they don't they get evicted and have to find another place to live at a much higher cost.
    Maybe landlords offer efficiency rents already and we just don't realise it?
    This year, I had a student working on this topic for their ECON499 research project, based on a survey of landlords. I'll post in more detail on the results later, since I am re-analysing the data, but the student's simple analysis showed that landlords are more likely to offer below market rents to existing tenants, but offer market rents to new tenants. At least, that's what the landlords say they do. Which suggests that efficiency rents are not a strong explanation for the excess demand in the rental market. Maybe landlords are not good at setting rents? Seems unlikely - most landlords have a lot of experience and many properties.

    Maybe what we're observing isn't really excess demand at all? If we think about the traditional excess demand in the goods market, the number of goods available is not enough to satisfy all buyers. So, some buyers go away from the market empty-handed, even though they were willing to pay the market price. However, in the rental market, most 'buyers' are already tenants somewhere else, so they aren't going away empty-handed - they are just going back to their previous property (as the example of Chris from the quote above shows).

    A better way to think about this might be a search model of the rental market, in the same way that search models have been applied to understanding labour markets (where workers might be searching for a new job even though they already have one). Something to think about in future. In the meantime though, expect a flood of media coverage about rents every January.

    Read more:

    Saturday, 19 January 2019

    $500,000 endorsement for the new BMS(Hons)

    One of my administrative roles for 2018 was being the Qualification Convenor for the University of Waikato's Bachelor of Management Studies with Honours (BMS(Hons)). I took the role on somewhat reluctantly at the start of the year, because I was already the Qualification Convenor for our other business degree, the Bachelor of Business. However, the Dean shoulder-tapped me for the role and it turned out to be quite rewarding.

    Most of the time, these admin roles are wearisome (or worrisome), but the reason this one was particularly rewarding was because I was able to lead a significant re-development of the BMS(Hons) degree (more on that later in the post). The re-developed degree was approved in November and the first students are enrolling in it right now.

    The new BMS(Hons) received a strong endorsement from the EQUIS peer review panel (one of our accrediting agencies) last year. But it has also just received another strong endorsement, in the form of a $500,000 donation, as the New Zealand Herald reports (see also the University press release, NZ Management, and SunLive):
    Business entrepreneur-philanthropist and Just Water founder Tony Falkenstein has pledged $500,000 to Waikato University's school of management to enable students to travel overseas and experience global business innovation.
    Falkenstein, who founded the business school at Onehunga College and is a champion of innovation and entrepreneurship, said he was "putting my money where my mouth is".
    The donation, which will be $100,000 a year for five years with the option of continuing another five years, is the largest donation of its kind to directly benefit students since Waikato University's management school was founded in 1972, said dean professor Tim Coltman.
    The overseas study tour funds will support the university's four-year Bachelor of Management Studies with Honours degree.
    Students will visit one or more countries with the opportunity to visit successful businesses.
    The international study tour makes the BMS(Hons) unique among New Zealand undergraduate business or commerce degrees, making it more like an MBA (Waikato MBA students also do an international study tour as part of their degree). However, it's not the only thing that makes the new BMS(Hons) stand out. The BBus degree is nested within the BMS(Hons), so students benefit from a common core and the same exposure to discipline-specific skills development as other business students.

    However, while the BBus is three years, the BMS(Hons) is four years. That extra year (which is actually spread across the final two years of the four-year degree) includes additional work-integrated learning (such as internships), the international study tour, and several papers designed to build students' transferable skills (sometimes also referred to as soft skills). These are the skills that are most in demand from employers, and the papers devoted to those skills include Managing People and Projects; Negotiating and Communicating Effectively; Data-Informed Decision-Making; Business Ethics and Managing Risk; and Leading for an Uncertain Future. Students in the BMS(Hons) also do the long-running Strategic Management capstone paper, which includes the case competition (which has now been running for some 20 years).

    It isn't clear just yet how enrolments are split between the BBus and BMS(Hons), but I'm confident that, once the message gets out students will recognise the value that extra year provides. And $500,000 tells me that at least one employer already recognises that value.

    Friday, 18 January 2019

    Gaming the Lambda School system and avoiding the tax

    The New York Times reported last week:
    Now, Silicon Valley is backing a novel idea that proposes to rewrite the economics of getting an education.
    The concept is deceptively simple: Instead of charging students tuition — which often requires them to take out thousands of dollars in loans — students go to school for free and are required to pay back a percentage of their income after graduation, but only if they get a job with a good salary.
    The idea, known as an Income Share Agreement, or I.S.A., has been experimented with and talked about for years. But what’s happening at Lambda School, an online learning start-up founded in 2017 with the backing of Y Combinator, has captivated venture capitalists...
    At Lambda, students pay nothing upfront. But they are required to pay 17 percent of their salary to Lambda for two years if they get a job that pays more than $50,000. (Lambda says 83 percent of its students get a job with a median salary of $70,000 within six months of graduating.) If they don’t get a job, or their salary is lower, they pay nothing. Payments are capped at $30,000, so a highly paid student isn’t penalized for success, and if a student loses a job, the payments pause.
    It is a model that so far has been aimed at vocational education but has the potential to end the crushing cycle of student debt and change the way schools think about students.
    “It aligns the incentives fully,” said Mr. Lewis, the venture capitalist.
    I'm not sure whose incentives are aligned here, but I doubt they are aligned fully in the way Mr Lewis thinks they are. Sure, the Lambda School has a strong incentive to make their graduates employable, and employable at a high salary, in order to maximise their return in the two years that the graduate is paying the 17 percent Lambda tax.

    However, the student's incentives are not necessarily aligned with those of Lambda. They'd rather avoid paying that Lambda tax, and the more income they earn, they more they have to pay to Lambda. On the surface, the student still maximises their income by finding the highest paying job after graduation.

    However, I'd say this system also creates some strong incentives for students in the opposite direction to what Lambda wants. There is an incentive for the student to engage in moral hazard. Moral hazard the tendency for someone who is imperfectly monitored to take advantage of the terms of a contract (a problem of post-contractual opportunism). In this case, the student takes advantage of their free tuition, but then tries to avoid paying the 17 percent Lambda tax.

    There are several ways students could game the Lambda system, by their actions after completing their education. For instance, the student could choose to take the first two years off after graduating, thereby paying Lambda back nothing. Of course, the student also gives up two years of earnings, so that isn't very attractive. And in any case, surely Lambda has thought of that already and there is a contract clause preventing it (say, the students have to pay 17 percent of their first two years' working after graduation).

    So, maybe the student could work as a volunteer, or in the Peace Corps? Or go on a Mormon mission? The student is still giving up their salary for two years, but they are adding to their CV and might earn more once they start working, and they don't have to pay anything to Lambda. Of course, Lambda could no doubt contract around those situations as well (e.g. 17 percent of the student's first two years of paid work after graduation).

    So, what if the student joins a start-up and is paid a nominal salary in cash, which is less than the $50,000 threshold, and receives the rest of their salary as stock options? Again, Lambda could contract for that eventuality. So, the smart student play here might be to find an employer who would agree to some form of deferred salary payment, payable after more than two years and thereby not subject to the 17 percent Lambda tax. Or they go into self-employment and pay themselves the $50,000 salary, but the remainder in dividends. Or... you can probably see by now there are lots of alternatives to get around the Lambda tax.

    Of course, not all students would engage in this opportunistic behaviour. But it's clear to me that Lamba's system is too easily exploited. One solution for Lambda (if they don't want to move to a formal student loan system) would be to have some minimum required payment, which would be paid back regardless of how many years it takes, and 17 percent for two years otherwise. That makes it less attractive for students to engage in moral hazard to avoid the Lambda tax.

    [HT: Marginal Revolution]

    Wednesday, 16 January 2019

    This couldn't backfire, could it?... Cane toad edition

    One of my favourite topics to write about (and to talk about in my ECONS102 class) is unintended consequences. There is a famous story about cobras that exemplifies the topic, as I wrote about in 2015:
    The government was concerned about the number of snakes running wild (er... slithering wild) in the streets of Delhi. So, they struck on a plan to rid the city of snakes. By paying a bounty for every cobra killed, the ordinary people would kill the cobras and the rampant snakes would be less of a problem. And so it proved. Except, some enterprising locals realised that it was pretty dangerous to catch and kill wild cobras, and a lot safer and more profitable to simply breed their own cobras and kill their more docile ones to claim the bounty. Naturally, the government eventually became aware of this practice, and stopped paying the bounty. The local cobra breeders, now without a reason to keep their cobras, released them. Which made the problem of wild cobras even worse.
    So, I was interested in this article in The Conversation last week by David Smerdon (University of Queensland), which references to cobra story, as well as this one about rats in Vietnam:
    When the colonial government built a sewerage system under Hanoi early in the 20th century, it inadvertently helped create a rat plague. Its solution was a cash-for-rats scheme - though to save the government having to dispose of hundreds of thousands of rat carcasses, it only required collectors turning in a rat’s tail to claim their bounty...
    The consequences this time were not only the creation of pop-up rat-breeding farms, but also hordes of tail-less rats roaming the city streets.
    Seen from 2019 New Zealand, this story is hilarious. But, the point of Smerdon's article is this:
    [Australian Senator Pauline] Hanson’s proposal involves paying welfare recipients 10 cents for each toad they collect (alive) and hand over to their local council. The council would then kill the toads humanely in large freezers.
    The senator is right to be concerned about the cane toad problem. Introduced in the 1930s as a biological fix to control native beetles eating sugar cane crops, the animals have prospered with devastating impact on native flora and fauna. It’s estimated there are now more than 200 million across Queensland and northern New South Wales.
    So, cane toads are almost a double example of unintended consequences, given the reason they were introduced into Australia in the first place! In any case, even though paying a bounty for cobras, rats, ragweed, possums, or cane toads sounds good in theory, surely we could learn the folly of this approach from historical examples?

    Read more:

    Wednesday, 9 January 2019

    Does studying economics make you sexist?

    Economics has a gender problem. If you don't believe it, just watch this webcast of the session "How can economics solve its gender problem?", from last week's AEA meeting in Atlanta. Or the several posts I have written on the gender gap (see the list at the end of this post).

    The question of whether economics is sexist is pretty much settled (the answer is, yes it is). The relevant question now, is why? Is it because studying economics somehow makes students more gender-biased (maybe because they encounter fewer female peers, or fewer female faculty members)? Or is it because students who are already more gender-biased (perhaps they have more socially conservative views) are more likely to study economics?

    Another paper presented at the AEA meetings, by Valentina Paredes (Universidad de Chile), Daniele Paserman (Boston University), and Francisco Pino (Universidad de Chile) provides an initial answer to the question of whether studying economics makes students more sexist. They used survey data from 3228 Chilean students, of which about 26% were enrolled in 'Business and Economics'. Their sample includes both first-year students, and senior undergraduate students (Years 2 to 6), so they are able to see how gender attitudes differ between economics and other students, and how this difference changes between those just starting out as students, and those who are late in the university education (this is what economists refer to as a 'difference-in-differences' research design). They use several measures of gender-biasedness, including a composite score of all of their various measures. They report that:
    For eight of the nine individual measures, we find that B&E students are more gender biased than non-B&E students, and five of the differences are statistically significant at the 5% level (in the specification with controls). The bias of B&E students appears to be particularly pronounced in the IAT-career score, hostile and benevolent sexism, and the gender roles-normative measure.16 The raw gap in the aggregate gender-bias score 0.19 standard deviations, and it falls to 0.14 standard deviations after inclusion of controls. In other words, the gap between B&E and non-B&E students is about one quarter as large as the gap between men and women. This is a sizeable difference.
    So, there is more gender bias among Business and Economics students than among other students. But, is this a result of studying economics, or were they more biased to begin with? Paredes et al. report that:
    The difference-in-differences estimator is positive for all measures (in the version with controls), but never statistically significant. Therefore the evidence in support of the training hypothesis is rather mixed.
    I wouldn't call that evidence mixed. I'd call it unconvincing. It appears that students choosing to study Business and Economics are already more biased before they start their studies. However, there are some interesting gender differences:
    For women, there is some evidence of selection: economics students are more gender-biased upon entry, and the gap increases at most moderately in subsequent years. For male students, on the other hand, we observe a moderate-sized but insignificant gap at entry, a large and statistically significant gap among upperclassmen, and a large and statistically significant difference-in-differences estimate.
    So, it appears that studying economics might increase gender bias, but only for male students. Paredes et al. go on to look at some possible mechanisms that might explain the difference. The proportion of female peers doesn't explain it, and neither does differences in political ideology. However:
    The share of female faculty has a strong and statistically significant association with the gender bias score for both first year and upperclass male students, but is almost unrelated to female students’ score. The difference-in-differences estimate for male students drops from 0.175 to 0.105, implying that about 40% of the B&E-non-B&E gap for male students can be explained by differential exposure to female faculty.
    A substantial proportion of the effect of studying economics on gender-biased attitudes among students can be explained by differences in exposure to female faculty. In other words, this suggests that having more female faculty is a good thing for reducing gender bias in economics. However, as noted in the webcast I linked above, this leads to a chicken-and-egg problem - the solution to reducing gender bias is to have more female economists, and to have more female economists, we need less gender bias. And given that the key difference is between first-year students and older students, we need more female faculty teaching introductory economics (am I talking myself out of my current teaching allocation?).

    Of course, there are some caveats to the Paredes et al. paper. Chile has a huge gender gap problem, as noted in the paper:
    The raw gender gap in monthly wages is 31.7%. Even among full-time employees, the gender gap in hourly wages is 19.4%. The United Nations Gender Inequality Index ranks Chile 65th out of 188 countries (the U.S. is ranked 43), while the World Economics Forum Gender Gap Index ranks Chile 63th of 144, and only 117th in the Economic Participation and Opportunity subcategory.
    That might contribute to the results in a way that makes them less representative for other countries where gender gaps are not quite as pronounced. There is also the problem that there is a gender bias only for female first-year students (but not males), which leads to the statistically insignificant effect overall when both groups are combined. There doesn't seem to be a good explanation for why female students would exhibit gender bias on entry to university, but not male students. So, there is definitely scope for more of this type of research.

    [HT: Marginal Revolution]

    Read more:

    Sunday, 6 January 2019

    Parachutes may be ineffective as safety devices

    A lot of what you read about research in mainstream media is taken from press releases (by universities or research institutes), or at best, is taken from the abstracts of research papers. Randomised controlled trials are the gold standard in evaluation research (especially in health), so many readers might not take a critical eye when reading the studies. This would be a mistake.

    A good example of where a lack of critical reading would go horribly wrong is this article from the Christmas issue of the British Medical Journal, by Robert Yeh (Harvard Medical School) and co-authors, entitled "Parachute use to prevent death and major trauma when jumping from aircraft: randomized controlled trial". Here's the first sentence of the conclusions from the abstract:
    Parachute use did not reduce death or major traumatic injury when jumping from aircraft in the first randomized evaluation of this intervention.
    One could conclude from that sentence that parachutes are ineffective as safety devices. However:
    Compared with individuals screened but not enrolled, participants included in the study were on aircraft at significantly lower altitude (mean of 0.6 m for participants v mean of 9146 m for non-participants; P<0.001) and lower velocity (mean of 0 km/h v mean of 800 km/h; P<0.001).
    The authors tried to enrol people into their randomised controlled trial by asking people:
    ...whether they would be willing to be randomized to jump from the aircraft at its current altitude and velocity.
    They would be randomised into making the jump either with, or without, a parachute. The only people willing to be randomised in the study, unsurprisingly, were those on stationary aircraft on the ground. This limits the external validity of their sample a little, but does allow them to conclude that:
    ...although we can confidently recommend that individuals jumping from small stationary aircraft on the ground do not require parachutes, individual judgment should be exercised when applying these findings at higher altitudes.
    This study was, of course, a response to an early BMJ article on parachutes from 2003, which concluded that:
    ...the effectiveness of parachutes has not been subjected to rigorous evaluation by using randomised controlled trials.
    Well, now the effectiveness of parachutes has been evaluated, and found wanting. Read both papers (they're open access); they're hilarious (as has been the case with many past papers in the Christmas issue of the British Medical Journal, such as this one I blogged about in 2017).

    [HT: Thomas Lumley at StatsChat, whose pet hate is journalists quoting uncritically from press releases or abstracts of non-peer-reviewed papers]

    Friday, 4 January 2019

    Gender differences in multiple choice answering

    In a post in 2017, I discussed multiple choice and constructed response questions in exams. One of the key points of that post was to highlight the gender difference in performance in multiple choice questions (female students do worse, but they do better on constructed response questions). It turns out that result is fairly common in the academic literature, but the reason why female students don't perform as well as male students on multiple choice questions is less clear. Maybe it is that female students don't respond well to high pressure or competitive situations (and multiple choice questions are reasonably high pressure). Women are more risk averse than men, so maybe it is related to that? Female students might be more likely to skip questions in order to avoid the risk of losing marks [*]. Also, men are more overconfident than women, so maybe it is related to that? Again, male students might be less likely to skip questions, because they are less likely to be unsure they have the right answer.

    So, I was quite interested to read this 2017 article by Gerhard Riener and Valentin Wagner (both Düsseldorf Institute for Competition Economics), published in the journal Economics of Education Review (sorry, I don't see an ungated version, but it looks like it might be open access anyway). Riener and Wagner conduct an experiment with 2060 German secondary school students across 89 classes in 25 schools, who each sat a maths test (based on the Math Kangaroo questions).

    There were three levels of questions (easy, worth 3 marks; medium, worth 4 marks; and difficult, worth 5 marks). Students lost one mark for every question they got wrong. So, there is an incentive not to guess the easy questions if you don't know (because there is a 1/5 chance of getting three marks, and a 4/5 chance of losing a mark, so the expected value of guessing is -0.2 [0.2 * 3 + 0.8 * -1]). There is no incentive either way for the medium questions (the expected value is 0 [0.2 * 4 + 0.8 * -1]), and a positive incentive to guess on the difficult questions (the expected value is 0.2 [0.2 * 5 + 0.8 * -1]). So, Riener and Wagner expect students to skip more of the easy questions, and fewer of the difficult questions. It turns out that wasn't the case, since they:
    ...find that the number of skipped questions is increasing in difficulty.
    I'm not surprised by this at all. Students don't understand expected value, so it wouldn't surprise me that they didn't realise that there was a positive expected value for guessing, for the difficult questions. And the positive expected value is a relevant consideration for a risk neutral decision-maker. Since there were only 14 questions in the test (and only 4 difficult questions), then it is relatively risky to guess on one of them, in terms of the impact on overall score.

    That wasn't the main results from the paper though, which concerned the gender differences, and the experiment. The experiment was that students in some classes were rewarded, if their test score was better than their earlier mid-term result (they didn't know about the experiment until after the mid-term, so there is no risk of the students engaging in strategic behaviour). The reward (according to Riener and Wagner, a source of extrinsic incentive to do well in the test) was one of: (1) a medal, awarded in front of the rest of the class; (2) a letter to their parents, praising their good performance; (3) a "no homework" voucher, which entitled them to take a day off homework; or (4) a "surprise gift" (which was actually a combination of (1) and (2)).

    They find that:
    Females always tend to skip more questions than males regardless of whether they are incentivized or not. However, incentivized pupils tend to skip fewer questions than non-incentivized pupils...
    ...girls in our low stakes baseline treatment skip significantly more questions than boys... However, the gender gap depends on item difficulty. While girls skip as many questions as boys when items are easy... they skip significantly more questions for medium... and difficult questions... Interestingly, providing extrinsic incentives for performance, and hence increasing the stakes, closes the gender gap in skipping test items. 
    So, providing an incentive closed the gender gap. They also find that the gender gap is only present in academic high schools (Gymnasium) and not in vocational high schools (Gesamtschule, Realschule, and Hauptschule). However, I'm not convinced by those results, as the vocational students sat an easier test, with fewer questions used in the analysis, which renders the results not comparable with the academic high school students.

    Riener and Wagner argue that their results are:
    ...suggestive evidence that the gender gap could be explained by a stereotype threat. Girls in high school only skip significantly more questions if the questions are difficult, although the attractiveness of answering is higher for difficult questions than for easy questions. Further support for a stereotype threat explanation is the fact that the gender gap vanishes if the difficulty of the task is made less salient (shifting the focus of pupils to winning an extrinsic reward).
    It's possible that this is stereotype threat. However, it is also possible that the types of rewards that they were offering were the types of rewards that girls value more than boys (especially given that the students got to choose their preferred reward). Perhaps the reward increased the stakes of the test for girls, but not for boys? In any case, it's hard for me to see how the offering of the reward reduces stereotype threat. So, although they were able to eliminate the gender difference in performance, I don't think this paper really helps us get to the bottom of why female students usually perform worse than male students in multiple choice questions.


    [*] This applies if there are negative marks for a wrong answer. It is a bit harder to argue this when there is no penalty for a wrong answer (or at least, no difference in the penalty between skipping and getting the answer wrong).

    Read more:

    Wednesday, 2 January 2019

    Book review: Foolproof

    One good thing about the Christmas break is that I have a lot of time for catching up on reading. So, it only took a couple of days for me to read Greg Ip's Foolproof: Why Safety Can Be Dangerous and How Danger Makes Us Safe.

    I quite enjoyed the book, as at its heart it is about moral hazard and especially about unintended consequences, which regular readers of this blog will recognise is one of my favourite topics. However, Ip continually returns to financial crises, which nicely knits the book together into a coherent narrative. If you've already read books about the Global Financial Crisis, then I doubt you would gain much from reading this one, but Ip writes it in a very accessible way, and the links to moral hazard and unintended consequences mean that almost any student of economics will learn a lot, along with the general reader. The overall message is that because our economic institutions (like central banks) work hard to reduce risks to the economy, we end up taking more risks, so that when the system fails, it does so catastrophically.

    Although the book has financial crises as an underlying theme, I generally enjoyed some of the other examples more (although I must admit, a lot of the stuff about the Volcker years in the U.S. was new to me). For example, this bit on floods:
    Gilbert White, an obscure government geographer who had been pursuing graduate studies part-time at the University of Chicago, noticed that the frenzy of levee and dam building in the 1930s had not solved flooding; instead it had created a new problem: more homes, factories, and farms had sprung up on the floodplain, so more destruction ensued when floods overtopped the levees.
    And this bit on helmets in the National Hockey League (NHL):
    Helmets became mandatory for new National Hockey League players in 1979. Thereafter the number of head fractures went down, while the number of spinal injuries went up. The conclusion of several specialists was that a more aggressive style of play, perhaps encouraged by the wearing of helmets and full face masks, was causing players to hit one another harder in ways that made spinal injuries more likely.
    The lessons from these other examples though, are targeted towards finance, such as this on the 1987 stockmarket crash:
    Nonetheless, the crash taught an important lesson about insurance against financial catastrophes. It works when only a few people buy it; when everyone does, it not only makes the catastrophe more likely, it threatens the survival of the system...
    Just as flood and earthquake insurance enable more people to live in flood- or earthquake-prone regions, insurance against market disruptions enables more investors to pile into those markets and perversely make the event more likely and more severe. Portfolio insurance had enabled this with stocks in 1987, and now CDSs [Credit Default Swaps] did the same with mortgages.
    Ip doesn't get everything right though, at least from my perspective. In one section, he does a poor job of explaining Prospect Theory, and consequently the book stumbles over the distinction between loss aversion and risk aversion. Similarly, most economists would disagree with Ip that risk aversion is characteristic of behavioural economics, and not traditional economics. However, the general reader will not notice these errors.

    As you might expect of a book about financial crises, Ip does come up with solutions. He concludes that:
    Our goal should be to eliminate big disasters, not small ones, to accept a bit more risk and instability today in return for more reward and stability in the long run.
    An analogy here (surprisingly not used in the book) is allowing children to take risks and hurt themselves a little, so that they can learn about their own limits and avoid catastrophe in the future. A few small boo-boos in the financial system should help us to reduce the chance of serious life-threatening events in the future. Overall, this is a good book for those interested in financial crises, but who don't want to do a deep dive into a book heavy on theory.

    Tuesday, 1 January 2019

    Economics students are extraverted, disagreeable, but emotionally stable

    I've written a couple of times about personality differences of students by academic discipline (see here and here). In one of the studies I discussed, students high in extraversion were more likely to choose to study law, or business and economics, and less likely to choose science, technology, engineering and mathematics (STEM). However, both posts were based on a single study.

    A 2016 systematic review by Anna Vedel (Aarhus University), published in the journal Personality and Individual Differences (ungated version here), summarises the results of twelve studies, across seven countries (in Europe, Israel and the U.S.), and including 13,389 students. All studies compared students' Big Five personality traits by discipline, although only four studies separated out economics students (and in one of those, "economics" included marketing, management, accounting, and public administration). Vedel finds that:
    Economics and Business scored consistently lower than other groups [for neuroticism]...
    Economics, Law, Political Sc., and Medicine scored higher than Arts, Humanities, and Sciences [for extraversion], and the differences often represented medium effect sizes...
    Humanities, Arts, Psychology, and Political Sc. scored higher than other academic majors [for openness], and effect sizes were often moderate or even large in comparisons with Economics, Engineering, Law, and Sciences.
    Law, Business, and Economics scored consistently lower than other groups [for agreeableness], and a few medium effect sizes were found in comparisons with Medicine, Psychology, Sciences, Arts, and Humanities...
    Arts and Humanities scored consistently lower than other academic majors [for conscientiousness], and medium effect sizes were found in comparisons with Sciences, Law, Economics, Engineering, Medicine, and Psychology.
    If we take those results as representative (which might not be as big a stretch as it sounds, as the effect sizes were reasonably consistent across studies), then economics students are more extraverted, but less neurotic (alternatively, more emotionally stable) and less agreeable than other students. The big question now is, how do we leverage those traits to improve student outcomes, or to provide better advice to students?

    [HT: Marginal Revolution, back in August]

    Read more: