Selection into the sample on the variable of interest: A Rant

First, a disclaimer:  I dislike mike the mad biologist.  I blame men like him for Trump being president because of his relentless attacks on ‘this woman, I would support any woman for president, just not this one” and the fact that he’s one of those assholes who does absolutely zero work and again, attacks people who are actually doing something as always doing the wrong thing.  Those men make doing things that they claim to support more difficult and emotionally draining while not lifting a finger themselves.  The only reason I ever look at his blog is because he’s on Bardiac’s blogroll and occasionally he’ll have one of those post headlines that I have to click.  Previously I thought he was just an unselfaware misogynist blowhard.  Now, I’m realizing that he’s … also not got very good statistics training.  (Interestingly, I’ve seen a recent survey that finds that people who are explicitly racist also just tend to be wrong about other things unrelated to their racism.  Not saying that translates to implicit misogyny, but… )

Ok, so here’s the post in question.  In it, he claims that a survey of people who aren’t getting vaccinated proves that time pressure and inability to take a day off work are not reasons.  (Therefore any policies targeting making getting vaccinations easier or getting time off work are wasted effort.)

The problem?  If you click on the survey, it says it has a 6.1.% weighted rate for taking the survey.  (So only about 6 out of every 100 people they sent the survey out to actually responded.)  Also, it is an online survey.

If that 6.1% were a randomly selected sample of a general population, there wouldn’t be a problem.   The problem is when the selection into the sample is based on the outcome that you’re measuring.  In this case, if you’re measuring people who don’t have time or ability to get vaccinated, well, likely they don’t have time or ability to take a survey either.

The Census Bureau isn’t stupid– they know this is a problem and they have lengthy documentation about the non-response bias in the sample generally.  They make it clear what you can trust the results for and what you can’t, as well as the limits of their weighting schemes.  The survey isn’t completely useless, but it is only externally valid for the groups that were surveyed!

I had been planning to use this little example of “sample selection on the Y variable” in my stats class this fall, but now I can’t because his response was so ironically ignorant that I have to blog about it instead.

Here’s his response:

The low income people who are supposed to be burdened by the time constraints also don’t report access as an issue compared to other factors. Are there any data that could convince you, or will the answer always be the same?

So– I guess I was right about his complete lack of self-awareness.  Can you imagine being convinced to make a huge policy change based on one extremely selected survey?  The people who ran the survey would never ever want you to make a policy change based on this result!

The answer to the question of what would be convincing (taking it seriously rather than just an accusation of me being set in my ways):

1.  A nationally representative sample that found the same result.  The US government has some of these, where you are required to take the survey and they have much better response rates (not perfect, but much better).  This survey is not one of them.
2.  A sample representative of the population who we are trying to target with our policies (ex. a state going into a few factories with onsite vaccination clinics before expanding the program).
3.  Multiple biased surveys that are biased in different ways (that is, are not biased on the outcome variable of not having enough time).

No good policy maker would make policy based on such flimsy evidence as the survey mike the mad biologist presents.  In fact, we rarely make big changes based on the result of any one study, unless it is the only study available or is the only well-designed large experiment available.  And even then, good policy makers keep their eyes out for new evidence and try not to do huge national things when the evidence is scant.  Ideally we’ll have a largescale randomized controlled trial, but failing that we’ll take a series of mixed methods– qualitative information, event studies (these two are the easiest and cheapest to do but can be biased depending on how they’re done), natural experiments, and so on.  Ideally we’ll have information about heterogeneity– we think, for example, that the effects of the Affordable Care Act and the effects of universal health insurance were different for Oregon compared to Massachusetts compared to Wisconsin or Tennessee.  And that could be because they have different populations and different starting environments, or it could be that each of these states had a different methodology used to study it with different biases.

Unlike Mike the Mad Biologist, every single thing I do (in research and in teaching!) has the potential of helping or harming someone’s life.  I have to be extremely careful.  I don’t make policy recommendations until the bulk of evidence supports those recommendations.  Because, getting back to that first disclaimer– I’m actually out there doing stuff, not just complaining about the people who do things.

So yeah, I teach my students about how not to use samples that are selected on your variable of interest.  It’s a more challenging concept than people say, lying about their weight or height, but it is an extremely important one.  I have a lot of students who go out and design/make/evaluate policy when they graduate.  Hopefully the lessons I give them remain with them.

Yes, Asian Americans also face discrimination in the US

Hate crimes against Asians and Asian Americans have been getting some news lately.  Not as much news as they ought to be, especially if you don’t follow any Asian American people on social media.

Our university, like many in the country right now, is only hiring for positions that increase the diversity of the faculty.  Unlike many universities that are only responding superficially to the Black Lives Matter movement, ours is taking a broader view and not just hiring people who do various versions of “African American Studies” or whatever else I’ve been seeing on job ads.  (Research suggests that doing a big cluster hire AND taking into account diversity as a structural whole will likely be more successful in retention and improved climate than the places that hire one black person to teach all diversity studies classes and then forget about diversity as soon as BLM is out of the news cycle.)

Our percent of Asian American graduate students and of Asian American faculty are much smaller than the percent of the population (at the undergraduate level we’re closer to matching).  As a large R1 public state school, we are supposed to look like our constituents and we don’t.

When we get “diversity points”* for students and faculty, in general, international students count as “international” regardless of their race or ethnicity and only domestic students (citizens, permanent residents, and those who could be included in DACA) count as diverse, meaning they qualify for university-level scholarships only available to increase diversity in our graduate student body.  Because our numbers are low for Asian Americans, our university gives us “diversity points” for recruiting and enrolling Asian Americans, but we are supposed to try to keep international within a certain percent so that we have international exposure but aren’t entirely international.  For this hire focused on increasing diversity, those guidelines have been relaxed (I assume because immigrant faculty are assumed to become US Citizens) and Black and Hispanic international faculty are considered to be increasing diversity for this targeted hire.

So we had two international candidates and the third candidate was Asian American.

The worst thing about discrimination against Asian Americans is that so many people believe that it doesn’t exist.  They look at numbers of Asian Americans at Berkeley or Harvard and argue that quotas should go in the opposite direction– that Asian Americans should be limited rather than encouraged.  That we should discriminate against them rather have affirmative actions for them.  They are expected to be “higher quality” if a place hires them over a White applicant– if equal, people believe the White person should be hired.  This is a kind of discrimination that plenty of people who believe they are Woke have– they deny that Asian Americans face discrimination and they put them in the “model minority” box.  One of my grad school professors told me that the hardest candidate for him to place (because of discrimination) was any Asian American male.  And indeed, it’s a bit jarring realizing how few Asian Americans we have in graduate school in my field outside of California.

Worse, not only do they treat all Asian Americans as the same (ignoring that Asia is an entire continent and that Asian Americans have widely varied histories depending on when, where, and why their families settled here and what their families experienced once here).  But they treat all Asians IN America as the same.

So when I noted that we only get maybe one Asian American graduate student every couple of years and that’s a known problem and we’re actually doing better with Hispanic and Black students since our last major intervention… several of my colleagues shook their heads.  We get TONS of Asian students from South Korea and China(!)  (!!!!!) they argued.  And I couldn’t even.  But instead I gently argued that no, Asians from Asia and Asian Americans have not had the same life experiences and are treated differently by admissions.  I did not mention that before that last admissions intervention, our graduate admissions officer referred to Asian Americans as “Oriental” (which is not a slur when referring to carpets but most definitely is when referring to PEOPLE) and worse, color coded them yellow (zie also coded Native Americans as red– I am not making this up).  I mean, is there any question about why we have so few Asian Americans in our graduate programs?

This is not to say that international faculty aren’t discriminated against or that Black and Hispanic faculty aren’t discriminated against.  They obviously are.  But the most insidious thing about discrimination against Asian Americans is the general belief that it doesn’t exist or that if it does exist it’s ok because: “model minority.”

In any case, we decided all three candidates were acceptable.  We’re making offers in order of the quality of their work and how well they fit gaps in our curriculum.  But that was after a lot of discussion about the intention of the cluster hire (the university had provided guidance that they found all three candidates equally acceptable in terms of increasing diversity, but some of my colleagues wanted to argue about that even though we currently have no Asian Americans in our department**).

*diversity points is an amorphous concept– basically we get audited every few years and if we are too white/non-Hispanic, we get dinged, and we get dinged the more off from the state averages we are for each major group.  Being dinged generally means having to write up a plan to fix the problem; I was in charge of one of the committees when we were doing really badly almost a decade ago as a newly minted associate professor.  Also there are scholarships from the university that only underrepresented minorities are eligible for.  At our uni, Asian-Americans are underrepresented at the graduate level.

**the department that shares a building with us has an Asian American professor of Japanese descent and an international professor from China.  A couple of my colleagues regularly mix them up with each other even though they are *nothing* alike in any way other than both having black hair.  They don’t even dress similarly (one is more business and the other business casual).

There are better options than increasing the minimum wage

Recently Jen Doleac (a leading economist in the study of crime and a wonderful person*) came under fire on twitter for saying that if the minimum wage were increased to $15/hr where she lives in Texas, she would hire fewer undergraduate RAs (currently $10/hr) and more pre-docs (people who have graduated college and plan to go to graduate school in the future) at $22/hr, resulting in fewer hires overall.  (Note to people who live in Blue Coastal cities– remember that cost of living in the South is a lot lower than where you live.  $15/hr seems pretty high to me given how low rents are where I live– $500/mo for a studio, much less if you’re willing to share a multiple bedroom place with others, and you still get your own room.)

What bothers me about this, other than Jen being a stand-up person who cares a LOT about people and genuinely wants to do things that help rather than hurt people**, is that the attacks on her are a knee-jerk reaction to a policy option that we still don’t know for sure will help or hurt people overall, and the mitigation of potential hurts is not being taken seriously because somehow raising the minimum wage is supposed to solve problems and not create new ones.  Ignoring potential problems not only doesn’t make them go away, but it prevents us from having comprehensive policy solutions.

The problem with most conservatives is that they want money and power and don’t care who gets hurt (also they have stronger levels of disgust and really hate moral hazard).  The problem with many knee-jerk ultra-liberals is that even though they care a lot, they often advocate for things that sound good but actually hurt the populations they’re trying to help.  (This is why Public Finance is such an important class for everyone!)

We don’t know if raising the minimum wage helps or hurts, because if minimum wage gets too high, employment will drop.  Yes, in our current environment it is probably too low now.  Yes, $100/hr is probably too much and will result in more robots and fewer jobs.  (Or in the undergraduate RA case– fewer paid RA positions and more unpaid class credit positions.)  Is $15/hr the right number?  $25?  $10?  It is really hard to say.  There are well over 500 papers examining the effects of the minimum wage.***  We do know that when firms have monopsony power****, employment goes up when the minimum wage is raised a little or in environments that can bear the increase.  It seems likely that a minimum wage increase may have short term benefits to employment outcomes, but then decreases employment over the long term (that is, firms don’t fire people, but they don’t hire as much when people leave when the wages are higher).  There’s a lot of newer evidence showing heterogeneous effects of minimum wage increases and it helps white dudes most and in some cases hurts black men and sometimes women.  So even if it helps on average, there’s a good possibility that large increases in the minimum wage will hurt our most vulnerable.

What does work?  Well, the evidence for the EITC is extremely strong.  Even researchers and policy makers who were EITC deniers back in the 1990s have come around because the evidence is so strong.  (What is EITC?  it’s basically a tax program that subsidizes wages for low income people.  Sometimes knee-jerk folks think, this is the government subsidizing big companies, and they are not wrong.  But are you trying to punish companies or are you trying to help people?  What is the end goal?  In theory, the EITC could be combined with a stronger corporate tax, but in reality…)  So Jen Doleac says, we need to expand the EITC.

But the problem with the EITC (as with increased minimum wages) is that you have to work to get it.  There are a lot of people who, for whatever reason, cannot produce enough to make even $7.50/hr.  You’ve probably worked with people who actively decrease other people’s productivity (possibly even had a boss like that).  Then there are people who are just not capable of regular work.  Now, we do have a disability program and TANF and WIC and etc. but this is not a comprehensive social safety net.  I fully believe that in a country as rich as the US even people who are not capable of being productive enough for a firm to hire them at minimum wage should be fed and sheltered and clothed and provided healthcare.

So what is better than increasing the minimum wage?

Basic Income.  Give everybody enough income that they can live a minimal subsistence lifestyle without working.  Yes, there will be moral hazard– lots of people will work less under this system.  Even potentially productive people.  I don’t care.  Some people shouldn’t be working (though because discrimination is real, under our current system that doesn’t necessarily map to the people who aren’t working) and I am ok with this (not ok with discrimination though).  As long as the basic income isn’t too high, most people will still want to work at least somewhat because most of us don’t want minimal subsistence lifestyles.  Then you can combine basic income with EITC, single payer health insurance, and stronger corporate taxes and there’s no longer even a need for minimum wage or disability or TANF or a whole lot of other expensive red-tape programs.   Of course it’s easier to change one program than to change a number at the same time.  And it’s likely corporate taxes would be dropped later without changes to basic income or EITC, which could be problematic.  We don’t know that much about the effects of basic income because it’s never really been tried on a large scale, but in theory the effects aren’t ambiguous like they are for minimum wage.  Productivity will drop but average well-being will improve and children will benefit from the stability.  (There’s details… how much should geography be taken into account etc. Subsistence income in Manhattan is very different than say, rural Montana.)

Now, from a political standpoint, you may say that increasing the minimum wage is an easier sell than completely restructuring our welfare system to basic income, and you would be right.  Basic income is probably less feasible.  Expanding EITC may be more feasible, and may be a better option, but, of course, it does have the problem that it’s the government subsidizing firms that provide low wages.

I hate the way that well-meaning people often label people as terrible because they don’t think about the potential negative consequences of the programs they’re advocating for.  When I was living in paradise, I was constantly bombarded with requests to sign petitions in favor of rent control.  To help the poor people, they would say.  But what happens with rent control?  You end up with more condos, fewer rental units, and what rental units there are the landlords become abusive and stop keeping up the residences.  (And you end up with current tenants making a mint subletting…)  In the end, some people benefit enormously, but generally not the poor people who are likely to need to move from time to time for work.  What is a better policy?  To remove the rules that buildings can’t be more than 2 or 3 stories and allow high rise apartments to be built.  But nobody wants that!  It’s really a policy to protect people with safe jobs already living there and to keep outsiders out.  There’s simple economic theory behind it, just as there is for minimum wages (same graph, actually, taught in the same section of Micro 101), but there’s also robust empirical evidence that rent control never helps the people it claims to help.  Rent control hurts poor people.  Allowing more units to be built helps poor people.  But “people who care” listen to the NIMBYs with their specious arguments and want rent control, not high-rises.

Some economists are bad people.  I cannot disagree that there are some people who choose this field because they really identified with the simplistic assumptions of Econ 101 and then they never took a public finance course because they don’t believe there’s a role for government or that that could be a useful course.

But there are legitimate reasons that economists who do care about things will disagree on the minimum wage and the cut point for the minimum wage.  (And a lot of this is going to depend on geography!  Housing costs alone vary enormously.)  Bringing these potential problems up does not make an economist a bad person.  Good economists who are good people want to do things that genuinely help and they do not want to accidentally hurt the people they are trying to help.  Good economists care about unintended consequences and they advocate for policies that help the most vulnerable.  (Bad economists just care about reducing moral hazard at the expense of total social welfare.)

Don’t just knee-jerk say someone is a bad person because a tweet on a complex issue doesn’t agree with Bernie Sanders.  One of my colleagues from Vermont says that nobody ever goes to Bernie for help because even though he talks a good game, he doesn’t actually do anything.  Complex issues are complex.  There are trade-offs.  And you can’t demonize someone for bringing up the trade-offs.  Because when you do that, you end up hurting the people you want to help.  Republicans have blinders on.  Democrats don’t need them too.  Evidence-based policy for all!

*I don’t know her very well, but I have seen her present several times.  She’s an extrovert and very good with faces and names, so she knows everyone and everyone knows her.  I am basing her wonderful personness on the body of her work and her interpretations of said work.  She asks important policy questions that show she cares about people and doesn’t start with bigoted assumptions like some economists do.  As a caveat, I didn’t much care for her naloxone paper when it came out but I can’t remember why not.  I think there were competing papers and I found the other one more convincing, but I’m not sure.

**She has a wonderful Ban the Box paper that shows that it decreases hiring among young black men but increases it among women and older black men… and of course white men.  IIRC, she has another paper on how algorithms and machine learning predicting recidivism will be racist if their inputs are racist.

***Many of the people who study minimum wage are jerks to each other.  There’s two big camps that have nasty arguments because results are extremely sensitive to specification choices and I recently told my managing editor I did not want any more minimum wage papers as co-editor.  People working on minimum wage are mostly white dudes because nobody else wants to get into the fray.  But there are a lot of them.  Note that because there are over 500 papers, you can find a paper that supports whatever your agenda is!  And the top economists disagree with which ones are better quality (and yes, there’s a correlation between their political ideology and whether or not they believe they should include state specific time trends, for example).

****Monopsony power is like monopoly power except the firm is the only purchaser of inputs, for example, it’s the only employer in town.

Disclaimer:  I am currently paying undergrads $14/hr which is higher than the university average, new grad students $15/hr (I have to get special permission for this!), and experienced grad students $18/hr (for some reason the top salary allowed for experienced student workers is higher than for starting student workers).  I don’t have pre-docs or post-docs.

I imagine this post might spark some discussion, but my Monday only has an hour free between in-person class and zoom meetings, so I might not be able to moderate/respond until late or even Tuesday depending on how exhausted I am and how much attention I have to pay at the meetings.

How to write a referee report

Seriously cribbed from A Guide for the Young Economist.  This is how the majority (though not all) of economists do it, and when I’ve reviewed for other fields I’ve been complimented on the organization, so I don’t think you can go wrong using this format even if you’re not an economist.

The letter to the authors

Start with a paragraph called “Summary”.  There’s some disagreement if these are still needed or just waste time, but I think if you do the summary paragraph right, it can be useful to both the editor and the authors.  The summary should NOT just be a restatement of the abstract.  It should be a summary of what the actual paper is about, and not what the authors think it’s about.  So, for example, if it’s an experiment, you would have a sentence saying what you think the authors are trying to do (Ex. The authors explore the effect of salt water vs. fresh water on underwater basket weaving.)  Then you say briefly what they actually did.  (The authors did a randomized controlled experiment using a student population in which…)  You might end with a statement about how they extrapolate their claims to a broader issue.  Often the abstract doesn’t actually fit what the paper is about– it makes much larger claims about what happened.  The summary should be neutral and describe what the authors actually did.  This is helpful to the editor to know what the paper is actually about before they read through it, and helpful to the authors in case what your understanding of the paper is about is different than what they intended.  They can fix their writing to make the paper more clear.  I find it helpful to focus on the method section and tables only for this part.

Then create three sections:

Major:

Minor:

Minutia:

Major should include things that you think must be fixed before the paper is published.  If the paper is a reject, then this is where reasons for rejection would go.  If the paper is an R&R, these are the items that must be addressed for sure.

Minor:  This is where smaller questions go.  You might have things that need clarification, things that are incorrect, additional robustness checks that are not make-or-break but should be addressed, and so on.  It is helpful to include page numbers with these.

Minutia:  This is where all the typos, page proofing, etc. stuff should go.  You’re recognizing that they’re small mistakes that the authors will want to fix, but that they’re not big deals.  These definitely need page numbers.  (If one of your recommendations is “spellcheck” because there are multiple spelling problems, I would put that under minor, as opposed to saying “should it be here instead of hear on the first sentence of page 28?” which would be a minutia, but YMMV.)

Always be polite in your referee report, even if the paper is ridiculous.  Do not make a reject/R&R recommendation within the paper.  (Also:  as an editor I can say for certain that positive letters don’t always lead to R&R recommendations and negative letters don’t always mean the person recommended reject.  It’s insane how some people can say different things depending on the audience.)

Advice is generally that you do not have to spend as much time on reject papers as you do on R&R– some people will say just stating the major points is enough if you plan to reject.  As a reviewer I generally try to give advice for making the paper better should it go to another journal or should the editor disagree with my assessment, but sometimes a paper is just not publishable so it doesn’t matter if they never fix the typo in footnote 17 even if I found it.  I’ve found editing at a lower tier journal that reviewers tend to over recommend revise and resubmit (they’ll be like, “the paper says that correlation is causation, but if they could only get at causation, this would be a great paper, R&R”), and the explanations people give me are much more important than their actual recommendations.  My colleague who edits at a top journal says reviewers over reject (“this is the best paper I’ve ever read, Reject”), so the explanations are important.  When I was editing a top field journal, reviewers tended to get it “right” on average.

The letter to the editor

You will also generally have a letter to the editor.  I find the best editors letters provide a concise summary of the letter to the authors and possibly elaborate on the context of your comments– Basically reiterate the major points that led to your decision of reject, or explain what must be fixed before publishing.  If you don’t have much to say because it’s an obvious accept, use this space to fight for the paper.  You don’t have to be anonymous in the letter to the editor, so you can say more things that put it into context or explain what you’re not sure about because it’s not your area of expertise or what you are sure about because you are an expert.  If you’re not sure if it should be R&R or Reject, here is a good place to say so and explain why– what are the pros and cons?  These pros and cons should also be in your letter to the authors, but you can provide more context in your letter to the editor.  You can also put disclaimers in the letter to the editor like, “I didn’t realize when I accepted this paper that it was written by a former coauthor” or “I reviewed this paper for a top journal earlier and recommended it be sent to this journal instead.” Some dudes who read this blog think that there should never be anything said to the editor that isn’t in the letter to the authors, but I strongly disagree.  I appreciate the reiteration of the major points of the review (especially since some people don’t use must be fixed as their delineation between major and minor sections, but instead use difficulty of fixing etc.) and any context that I should know about (and I really don’t need to know about that typo on footnote 17 unless the paper is a revise and resubmit, but not everybody keeps those things to the minutia section).

Special topic:  Top journals

For top journals (for which I have not yet been an editor but have done a lot of reviewing), you may want to keep in mind the following points:

1. Is it clean/well-done?  (This is the bare minimum)
2. Is it Novel? (Doesn’t always have to be, but it helps a lot… though you can’t be too novel or it gets rejected because it’s “not economics” even if it actually is.  grrrr.)
3. Does it make a major theoretical and/or empirical contribution to the field?  (Sometimes papers don’t need to add to empirics, but they do need to have a theory base even if not literally a theoretical model.)
4. Is it Important/ of general interest? (This is highly subjective and where many of my papers strike out because it turns out they’re ahead of their time.  grrr.)

Update:  Here’s xykademiqz on the same topic for her science field.

Do you do a lot of referee reports?  How does your field handle them?

Ask the grumpies: Should economists not teach anything about race?

SLAC prof asks:

In a tweet, Trevon Logan says

Message to 99.7% of Econ professors: This coming academic year, do NOT attempt to teach about race, racism, or racial inequality. Just don’t. Even if you’ve published papers on it. Respect this as a body of knowledge and refer students to appropriate academic units.

— Trevon D Logan (@TrevonDLogan) June 21, 2020

The whole thread has more information.  It makes me want to give up.  He says economists do race all wrong.  What do you think?  And what does one need to do/know to be qualified to teach about race?

Ok, so first off:  I am not black.  Also I know and hugely respect Trevon Logan and his work (and I’m fairly sure we referee each other’s papers and I’ve always been impressed with his!)

But I disagree with him.  I think this is ok for two main reasons:

First, I have had a relatively large number of black (mostly female) students, many of whom have taken some of these cross-campus classes he discusses, and they have always asked me for more on race, not less.  You just cannot teach health economics without discussing disparities (and many of the big papers in this area are from epidemiologists and demographers, not economists).  You cannot teach labor economics without having a huge section on discrimination, and while many of the white male economists working in this area have blinders on, it is fairly easy (if you have been listening to people, or if you’re female/minority) to point that out and modify their theories into something more realistic and less bigoted.  Like, of course taste-based discrimination exists, we don’t have competitive markets, duh.  (And current US events during my last semester’s class made it very clear that discrimination can lead to monopoly power, not just be a consequence of it.)  Theories of statistical discrimination should include incorrect stereotypes because we don’t have perfect information, honest to FSM.  Your (not privileged white male) students can generally point out these flaws themselves just using their own experiences and common sense.  You cannot teach public finance without talking about the political economy of race and how these programs affect different groups.  Heck, Political Economy is less than half a class without discussing race.  Similarly, Law and Economics (even if you’re planning on limiting to patents and contract law, race is still a factor!).  Sports economics!  You just cannot do justice to any subject that affects money or people without discussing how race impacts it.  So I include these topics and every year my students have more ideas for things to add.  (Like yes, in health economics we do need to talk about how white doctors have used black women’s bodies and DNA without their permission, you are absolutely right.  That would be a great addition to the Tuskeegee paper we already discuss.)

Second, I have listened to the troubles of our young black female faculty across campus (I was on a university-level thing to improve things, which we sort of did but also mostly didn’t … in any case, we did a lot of listening in addition to convincing the university to allow salary equity bumps and a few other things) who primarily teach these classes that Dr. Logan is suggesting we send our econ majors to.  It is really unfair to them to inundate them with mostly white male econ majors who have been taught that it’s just fine to play devil’s advocate and haven’t really examined their implicit biases at all.  I have enough trouble breaking them in in my intro stats classes.  Can you imagine how disruptive they would be in a discussion based class with women and minorities from what they consider to be lesser majors?  That is going to have huge negative spillovers.

I have other reasons to disagree which may be less ok, and I would modify his advice some.  (Note that since I wrote this post– several other people in the comments of the twitter thread have made these or similar suggestions.)

First off, I agree with him 100% that most of the white dudes in econ who gatekeep and work on racial discrimination start from racist assumptions and for many of them, their main goal is to show how it is Black people’s fault (or women’s fault etc.) for not being more like White men.  It’s only recently that economics has started thinking that no, maybe Black people and women are rational, they’re just playing a different game.  This problem can easily be solved by just saying, “Don’t teach any papers on race by white men (or by Roland Fryer who may be black but has serious issues).”  You can even modify this advice to “Teach only papers on race by black scholars (except not Roland Fryer).”  There’s plenty of great work by black scholars and some by other minorities and women that don’t start with racist assumptions or trying to bend evidence to “prove” racist ideas.  There are even textbooks and summary articles that would be great for lower-level undergraduate classes (William Darity Jr. is a good author/editor to start with).

And there are a LOT of white economists who could themselves benefit from reading this work.  Maybe they should start with So you want to talk about race and/or White Fragility and following Black scholars on twitter.  Then they can move on to articles in academic journals.

In terms of whether or not economists think about discrimination incorrectly… some of them do, but I think we benefit from looking at how different social sciences deal with race and discrimination.  NONE of them give a complete picture.  The assumptions and questions asked are different.  We gain tremendously from thinking about these different viewpoints and different ways of modeling.  (I took Race and the Economy from an amazing Black woman and she incorporated overviews from other fields in the class.  It can be done.)  I could go into huge detail about this, but that would get too long… suffice to say that these different viewpoints complement each other; they are not substitutes.  An economist can learn a lot from how anthropology, sociology, psychology and other fields conceptualize discrimination and other questions involving race.  (Insert rant about irritating white male gate-keepers in labor economics here who think innovation and interdisciplinarity is incorrect.)

Maybe the better advice would be for economist professors themselves to take a few classes across campus, or at the very least, read a textbook from another field, before adding race to their classes.  They should also read up on how to make their classroom more inclusive so that students don’t feel scared to speak up when the professor screws up.

As for me, I have been including race in my classes since I started and I cannot imagine stopping now.  The more I teach, the more I listen to my students, and the more I learn from them, which helps students the next time I teach.  It is a learning process for everybody.  Did I have some cringeworthy moments when I first started, probably, but minority students have been gentle with me and each year I’ve learned more and gotten better and future students benefit from that.

Update:  The more I talk with my colleagues interested in adding a race unit in their classes, the more I’m convinced that my suggestion about only using papers written by minorities is the correct one.  I had no idea that people didn’t know Becker was a huge racist misogynist jerk(!)  I mean, I thought everybody knew that.  People knew it back when he was still young, like decades ago.  So no, DO NOT read Becker in the raw original.  Many of his theory structures are lovely, but read them with the sexist and racist assumptions removed by someone else; there are great minority scholars who have explained the baseline theories and added to them, so go with them.  (William Spriggs talks about some of the problems still inherent today.)

I swear, my colleagues are all going to give up and just end up covering Bertrand and Mullainathan, though I did convince one to try Quillian et al. (in PNAS) instead.  Look, it’s not that B&M isn’t a great paper, it is, but the really horrible overlying thing is that it got into the AER because everybody, including labor economists who should have known better, thought this was the first time a correspondence audit had been done, completely ignoring ALL of the correspondence audits done by Black scholars or non-Americans– I learned about them in my undergrad economics class on Race in the economy.  What I mean is, I’m fairly sure that racism is the reason those earlier audits by black people aren’t known at all.  Quillian and coauthors do a good job of collecting them and plotting their results over time.  (It should have been published in Science, but the racist editor overruled like 7 referees who all said it was must publish.)  Quillian is also white, but he’s a sociologist, so maybe he gets a pass?  Plus he’s very nice.  I’m not sure if there are any minorities in the “et al” portion.  (Plus the econometrics textbook we use has B&M as one of their datasets and students replicate all the ttests and regressions, so it’s not adding that much for our majors.)  Any time I explain this to a White labor economist they get really mad at me because B&M is somehow the first hardcore proof they’ve ever seen about racism against black people other than those small scale in-person audits from like the 70s that somehow Jim Heckman “disproved”  in the 1990s (spoiler:  he didn’t really).

Update 2:  Last night we talked to a number of students and alumni (mostly underrepresented minorities) and they said to be careful to make sure that the lesson is integrated into the curriculum, and to not just have it as a separate unit unconnected with the rest of the class.

Ask the grumpies: How to subvert the Tragedy of the Commons

Leah asks:

Is there any way to subvert the tragedy of the commons, or are we doomed to that fate? I seem to remember learning some examples way back when I took environmental economics but they all escape me . . .

I just happen to teach a class on this!

The first way we learn about is with government setting property rights and facilitating costless Coasian bargaining.  In the canonical example, there’s a river and a factory and a fisherperson.  The factory pollutes the river which kills some fish.  If the factory owner owns the river, then the fisher can pay it to pollute less.  If the fisher owns the river, then the factory owner can pay them to allow some pollution.  There’s problems with this solution when there’s not costless bargaining, when there’s multiple fishers (that can cause a holdout problem) or multiple factory owners (and they don’t know which ones are causing the pollution), but that’s the “preferred” government intervention when it works because it leads to the least amount of deadweight loss.

Fancier versions of this solution include things like the government setting a specific number of pollution credits and allowing firms to bargain over them.  That’s the idea behind Cap and Trade.

When the Coasian solution is difficult to implement, generally because of bargaining problems or informational aysmmetry, the government can step in a bigger way.

First:  The government can mandate that firms not be allowed to pollute more than a certain amount or fish more than a certain amount or hunt more than a certain amount.  Associations can also take the role of government in order to say, prevent over-fishing, though it’s often harder for a non governmental association to enforce these kinds of mandates.  Mandates are most enforceable when there’s jailtime associated (not just a shell company going bankrupt), though that tends to be unpopular.

Second:  The government can tax things like pollution or things that cause pollution.  Think gasoline taxes or hunting fees.

Third:  The government can subsidize companies to not pollute or to not fish etc.  This option tends to be the most popular with industry.

All of these methods have situations in which they work better or worse than the other solutions.  With nuclear waste, you want a mandate because even a little bit of waste is bad.  With air pollution you might want a tax or subsidy or cap and trade system.  The government can make money with taxes or by selling property rights in a Coasian situation.  Companies tend to lobby for subsidies which makes them more politically feasible.

So the short answer is:  yes, government can subvert the tragedy of the commons.  Market failure is why there is an economic role for government and the tragedy of the commons is one of the causes of a main source of market failure (negative spillovers).  But we need political will for it to work.

Replacing a misogynistic mnemonic

It is easy to get backwards when doing t-tests, especially when you’ve first started.  You have to remember that big t and small p-values mean to reject.  Of course, if you’re a little dyslexic (undiagnosed) like me, when you get confused, you can go back and re-figure out that you want small amounts in the tails and the t-slice to be far away from the mean making it large etc., but that is really time consuming, especially if you’re in the middle of a lengthy problem set or a timed exam.

Many years ago, one of my students shared a dirty misogynistic mnemonic.  When p is low, she said, reject the ho.  This is clever and funny because p is both probability and slang for penis.  Ho is both how a null hypothesis is written and a derogatory term for female prostitute.  When one’s penis is at low mast, it makes sense that said penis-holder might not be purchasing the services of a prostitute.  And it rhymes.

But it’s both dirty (I got into trouble for saying “prick” meaning “jerk” early in my classroom career) and one should not be using derogatory terms for prostitutes or women anywhere, much less in an institution of higher learning.

More recently, I was explaining my conundrum in office hours and one of my sunshiny students came up with a much better mnemonic.  It isn’t quite as clever, but it’s just as memorable and it still rhymes.  “When p is lo, reject H-O!”  Like a cheerleading chant (aich – oh).

It makes me much happier.

What are some good non-racist, non-misogynist, non-ablist, non-patriarchy mnemonics that you know?

Ask the grumpies: whether or not to purchase insurance

H.I.P. Person (Home Insurance Purchasing Person) asks:

How do you decide if you need insurance for something? We are updating our home owners coverage and they are peddling the following:

1) service line coverage –  up to $10,000 per event
2) systems coverage (a/c, hvac, water heater, furnace, and the like) – up to $50,000 per problem
3) sewer back up (?? not sure of max coverage)

We live < 75 miles from the coast; <20 miles from two major water coastal inlets but  not in a flood zone. What hurricane related insurance should we own? The policy comes with wind/water but not flood and they don’t really want to sell us flood (they are offering #3 instead).

Are any of these worth it?  What price point would make them worth it or not?

DISCLAIMER:  We are not financial advisors.  Get advice from real professionals or do your own research before making important monetary decisions.

So in economics, you purchase insurance when the expected utility of the insurance is greater than the cost of the insurance.  So, assuming you knew how your utility function was shaped (the important part for this purpose is that you know your coefficient of risk aversion, in this case specifically how much you hate the possibility of loss), and you know the probability of a bad thing happening, you just multiply that probability by your expected utility plugging in the amount you’d be out if the bad thing happens and then add the probability that the bad thing doesn’t happen and multiply that by your expected utility plugging in the amount you’d have if the bad thing didn’t happen.  Problem is… in reality, we usually don’t know the probabilities of these negative events occurring (or even what those negative evens could be!) and we definitely have no clue about what our utility functions look like.

So how does one decide what to get insurance on in reality?  Well, if you have rough ideas of probabilities, you can look at the expected value of something happening.  Expected value is like expected utility, but it tells you what the break-even point is assuming you have no emotions.  You’re neither a gambler nor risk averse.  But risk averse you can look at those numbers and think to yourself, “How do I feel about this calculation?”  In general, the insurance company is out to make money, so they’ll be charging more than the expected value of the thing happening, but does the amount more that they’re charging seem reasonable to you?

If you don’t know the probabilities of something bad happening, you can still play with worst case scenarios:  If the bad thing happens, what would you have to pay out of pocket to fix it without insurance?  What would you have to pay out of pocket with their insurance?  Is the peace of mind for the difference between those two numbers worth what they’re charging?  (And again, if you have rough ideas of how probable these events are, you can factor that in as well).

Another thing to do is to google around, preferably with reputable sites, to see what kinds of insurance are usually a good idea and what kinds are generally scams.  You can ask people around you too, though people often do things that don’t make sense if there’s good marketing on the part of the insurance company or if they’re more credit constrained or risk seeking than you are.

An important thing to note is that you want insurance to insure against risk.  You don’t want it as a pre-payment for things you’re going to pay for eventually.  You don’t want a high monthly cost if you can avoid it by sharing some of the pain should disaster strike through a higher deductible.  The goal isn’t to save money, it’s to smooth your consumption over good and bad states of the world.  Don’t try to beat insurance– in the best states of the world you give them money and they never give you money.  But you want it there when disaster strikes if you can’t handle the disaster on your own.

If it were me, I’d definitely eschew the service line coverage unless it was really cheap– we can afford a 10K emergency.  Since the limit is capped, it is not actually useful insurance unless paying up to 10K would be devastating.  (Capped insurance is often a red flag– if they stop paying after a certain amount is spent you may be better off self-insuring because if something really terrible happens they’re not going to be much use, and self-insuring means you’re not paying for the additional administrative costs of going through them.)

Systems coverage again, we probably wouldn’t pay… replacing a/c, hvac, water heater, furnace etc. is all stuff that has to get done some time anyway (and none of these should cost 50K, unless that’s including the damage after a water heater explodes or something) so paying them is like pre-paying for bills you’re going to have, but it is likely you’re going to have less choice about how to make those replacements and you’ll be paying their administrative costs over what you’d be paying if you did these things yourself.  And, again, it’s capped.  This is unlikely to be good insurance.

Sewer back-up is the only one of these that doesn’t sound 100% scammy.  Check all your other insurances to see if this is covered under them.  Estimate how much a sewer backup could destroy.  Make sure that it’s unlimited covered and not capped and there are no other strings attached.  Think about the probability of this happening.  Look at how much they’re charging for it.  Then go with your gut.

We have no idea about hurricane insurance.  The internet has a bunch of pages about it, noting you should get wind and flood on top of home, but I’m not getting a good idea of what numbers you should be looking at or even how to make that calculation.  You may also want to look into the different deductibles they offer, because none of these sound particularly cheap, but it may be that if you have a high deductible the insurance cost will be more reasonable (assuming you can afford the deductibles).

Good luck!

Grumpy Nation, do you have any better advice for H.I.P.?

 

Ask the grumpies: Privatizing nation’s air controllers?

Crone asks:

opinion on privatizing our nation’s air controllers. I oppose but was told the whole system should be moved to computer based GPS system and then Highways in the Sky for planes could be free form making flights faster and private industry can do this more rapidly than government. (I was in social situation so could not say I have never known a single computer system that did not ‘go down’ or ‘have ‘undocumented features ‘ so how would that work…) The topic of pipelines that ‘will not fail but ALL LEAK at some time’ had already come up.~~ I had been assured I was wrong on that point and ignorantly female. SO, back to air controllers: If this would be profitable for private companies to do why isn’t it done profitably or better by public government?

OMG, the WORST IDEA. OMG OMG OMG.

Well, it’s only the WORST IDEA if you think that airplane passengers are more important than prisoners. If you think prisoners are people too and should have rights, then privatizing prisons is actually the worst idea and this is only second worst. I guess there’s also privatizing foster care systems… if you think all people are equal then that might be slightly above air traffic control but still below prisons in potential harm done by privatization. (Foster care systems empirically aren’t as bad as prison systems, even though the potential is there to be as bad. This has to do with better state oversight.)

I had a section on privatizing public systems in one of my classes last semester and students brought in stuff– if I’d known it would come up as an ask the grumpies I’d have taken a picture of the whiteboard commonalities of when it works and doesn’t that we came up with. It can be ok, but it depends on a lot of stuff and it really shouldn’t be something where you know, people could die.

Ugh, so no, not you being ignorantly female. There’s a reason there’s a role for government for various systems.

We generally think that there is a potential role for government intervention when there is market failure in the competitive markets.  One form of market failure comes from monopolies.  Something like air-traffic control is what we call a “natural monopoly”– natural monopolies occur when it just doesn’t make sense for more than one company to be in one market.  A lot of utilities are in this kind of situation– where it doesn’t make sense for two companies to lay down pipes or what-have-you.  (You can also have government-private partnerships, where, for example, the government owns the rail-lines but allows different companies to pay to use them.)  Air traffic control is an example of a natural monopoly.  At an airport, it makes sense for only one company to do the air-traffic control.  Any more could lead to planes, for example, hitting each other.

The government in this situation could still allow private contractors to bid on the ability to be that one company doing all the air traffic controlling.  Unfortunately, air traffic control benefits a lot from experience and there are switching costs when an old company leaves and a new one takes its place.  Those switching costs could lead to not just inefficiency but also death.   Finally, oversight is really important with privatization.  Unlike the government, companies can just go out of business when they cut costs so much that people die, so they don’t have as much of an incentive to stay safe when it means cutting into profits.  Government can combat that by making it costly for them to cut corners before someone ends up dead, but that oversight comes at a cost.  Those costs could be large enough (and the possibility of bribes could be high enough) that it makes sense for the government just to do it itself.

 

What does statistical significance mean?

One of my students sent me this article because we spend some time in class covering Type 1 and Type 2 errors.

All the .05 threshold means is that you have a false positive 1/20 times.  A .005 threshold would say you’re getting a false positive 1/200 times.  So by moving to a .005 threshold, you’re less likely to get a false positive.  That’s good, right?  In common parlance, we’d be less likely to send an innocent person to jail.

Well, that depends.  At the .005 threshold, we’re more likely to get a false negative than you would at the .05 level.  That means we’d be more likely to get a guilty person go free.  (Indeed, the only guaranteed way to send no innocent people to jail would be to send nobody to jail.  I, for one, am happy that folks like Charles Manson are behind bars.)

It isn’t as easy as saying, oh we should just switch to .005.  When you adjust the p-value you’re making trade-offs between type 1 and type 2 error.  With a lower p-value threshold you’re going to be getting a lot more false negatives even with fewer false positives.  What we always need to be cognizant of when we’re doing policy is that significance isn’t everything– we also have to think about what the damage is if this information turns out to be incorrect.  For example, doctors recommend that pregnant women should heat up cold cuts if they’re worried about listeria, which is a very low probability event but if it happens it’s horrible.  It’s pretty easy to avoid room temperature cold-cuts for 9 months, so unless there’s some other difficulties attached to diet, women will probably follow this recommendation.  (And if one accidentally eats room temperature coldcuts while pregnant, one shouldn’t freak out because the probability of getting listeria is very low!)  But if we’re talking about something like doing chemotherapy or surgery, that’s a much more onerous action and we might want to be more sure we need it before going ahead with it.

Another thing to note is that the article talks about how physics and genetics have already made this switch, while most social sciences haven’t.  One big difference between the fields that have made the switch and the fields that have not is how easy it is to get large samples.  A larger sample size will make it so your sample behaves more and more like the population that you’re trying to study.  We can reduce both Type 1 and Type 2 errors simply by increasing the sample size.  So why don’t we do that?  Well, it turns out that increasing the sample size can be very very expensive when you’re dealing with people and behavior.  Sometimes doing the study with a large enough sample to get 80% power and an alpha of .005 might be more expensive than just throwing that same money at the intervention you’re trying to decide about, whether or not it actually works.  There probably is some resistance because people in these fields want to be able to publish their 5% results, but that’s not the main or only reason we haven’t yet made the switch.  Research is complicated and expensive and we have to make trade-offs.

The context for these really does matter, and you shouldn’t necessarily put off making policy choices just because your sample size is too small to get significance (or to make policy changes just because you have significance).  You always have to be aware of the costs and the benefits.

 

(Incidentally, in case he comes across this, Hi Dan!  I’m assuming that the reporter greatly simplified your arguments here because I know you must know this stuff.)