Two recent Econlib posts of note

1. David Henderson discusses Scott Winship's skepticism about claims of the costs of inequality.

It's a good post to take a look at just to get the whole scope of the issue, but I think there are a lot of problems with Winship's claims and the bashing of several left-leaning economists is entirely inappropriate and unsupported by what he presents. He puts a lot of emphasis on work on inequality by the CBO and work by Burkhauser. These analyses are fine, but hardly the CBO/Burkhauser research speaks to fundamentally different questions than the stuff the list of economists he bashes at the beginning usually cite. It is household adjusted, which is good to look at but different, and it's post-transfer, which is good to look at but different.

If we are thinking about redistributive policy that latter point - that the CBO/Burkhauser numbers are post-transfer - is extremely relevant. One can't very well argue that redistributive efforts are unnecessary because inequality is illusory if your proof that inequality is illusory demonstrates that it's illusory because of redistributive efforts! Indeed, that may even show that the existing redistributive policies were necessary and did their job!

Inequality is a very complicated issue to discuss, and there are lots of different questions that people are legitimately interested in - is the wage distribution fair? do families have access to the resources they need? does redistribution work? is inequality a problem when you consider prospects over the life course?

All of these are important - but all call for looking at very different data. Too much of the accusations around inequality are derived from muddying the fact that people seem to be interested in fundamentally different questions.

*****

2. Bryan Caplan discusses job mismatch.

I think people need to approach this discussion very carefully too. This issue is actually exactly what the Sloan Foundation has funded me for three years to look into. The job relatedness questions he cites should not be interpreted as "mismatch", in my opinion. For one thing we should be skeptical about exactly what they're capturing. Mark Regets, an economist at NSF who works exclusively with the NSCG data that Bryan cites, always likes to refer to bizarre cases that come up in the data where surgeons report that their job is unrelated to their education (i.e. - medical school). Unfortunately, he refused to disclose the identity of these survey respondents so I could avoid these surgeons (something about privacy, etc.). More importantly for the purposes at hand, it raises questions about how people are interpreting the survey. Maybe most of what the surgeon does he feels he learned on the job or through inservices - that school only contributed a small component to his knowledge. Maybe these people are focusing on the detailed technical knowledge they gained outside of school and are ignoring the fact that school taught them how to "think like an economist" or "think like a lawyer" - a very important contribution even if people pick up the details of their work after school. It also has to do with conceptualization of what a certain major "does". An English major writing proposals for a big contractor may say that her job is unrelated to her major because she feels like she's gone off of the editor/journalist/writer track she thought she wanted. Of course that's nonsense. In all likelihood she does her job much better than someone that doesn't have the background in writing and editing that she does.

Another reason to not interpret these questions as evidence of mismatch is that mismatch has a specific meaning in economics related to a poor search-and-matching process in the job market. So do we think that everyone working outside their field is poorly matched to their job? Of course not. Lots of people choose to work outside their college major for lots of different reasons. This is especially important for interpreting the wage premium associated with working in a job related to your major. If you work outside of your major by choice (in other words - it's not a matching problem), it's likely because you are very passionate about this opportunity that has come up. Passion in the labor market leads to compensating differentials: they don't have to pay you as much because you are doing what you love.

There's also a question of the underlying skills distribution within a major as an explanation for the earning differential. If there are fewer engineering jobs than engineering majors, the best engineers may be more likely to get the engineering jobs (if they want them). These engineering majors are going to be more productive than the other engineering majors, so of course they are going to enjoy a premium. You would see this even if the cost of mismatch were zero. The NSCG data used in the paper that Bryan cites has no grades or ability information that I am aware of to account for this, and the notes on the wage regressions don't seem to account for it either.

So approach with caution.

Our Sloan work specifically concerns the "loose coupling" between major and occupation. Most people assume the science and engineering labor force is a pretty smooth pipeline, or if they know that people work outside of the field they assume it's always bad or a divergence. We're looking into the decision making behind working out of field. This should be the subject of one of my dissertation chapters (and then hopefully that will be an article), as well as a bunch of other papers we're putting together. We'll be using the NSCG, as well as the Baccalaureate and Beyond survey, which does have data on grades.

Best line yet in the microeconometrics seminar, plus another great guest lecturer

"As everyone knows, Russell Crowe invented game theory"

Ya, I should probably watch that again soon.

*****

In addition to the normal seminar discussion leadership, we had a lecture by Aman Ullah of UC Riverside on non-parametric econometrics. I thought his criticism of parametric models was a little too harsh, but asked a question that he thought was a good one that was framed a little nicer. One of the things he raised was the specification of the earnings equation by Heckman and his students as a quadratic function of experience. He talked about how he re-estimated it non-parametrically and found a quartic relation - and that this was later confirmed by a more exhaustive parametric analysis which tested a quartic relation against quadratic and other options.

That's all fine, and a nice exercise. But I noted that they didn't use a quadratic relation because they were too lazy to do non-parametrics. They used it for theoretical reasons based in Becker's human capital theory. Then I asked whether any theoretical framework for explaining the quartic relation (he had come up with this in the late 80s) had been developed, because if this is really robust it ought to tell us something about our labor theory. Not that he was aware of, he said.

This is an important point, though. The explanation offered at the time of his paper was the impact of World War II (I believe... maybe Vietnam): a cohort effect. That's fine, but that doesn't suggest there's anything wrong with the quadratic specification - that suggests that there's something wrong with an omitted variable that you ought to have conditioned on.

That's a huge difference! It's the difference between revising human capital theory and not revising human capital theory.

My concern with non-parametrics is that it gives the illusion of being agnostic when actually you are making some very consequential assumptions about whether its OK to be indifferent between a conditioned and unconditioned binary relationship. Now that's not necessarily true of newer advances in non-parametrics, which he talked about as well. Non-parametric methods can be used to estimate "functional coefficient" models that do condition on other variables and in that sense offer the opportunity to get more meaningful stories back into the analysis (albeit not by assuming a parametric specification). He's used these functional coefficient models to look at some of Card's work on the returns to schooling.

One question I didn't ask because time was running over was about the role of non-parametrics in non-experimental techniques that rely on a particular functional form. It's one thing to be agnostic about theory in an empirical question, but if you're using a particular non-experimental technique you cannot be agnostic about the assumptions required for that technique if you're implementing it. The specific example I have in mind is the popularity of using non-parametric methods like local linear regressions to estimate regression discontinuity models. It's always seemed to me that these pose serious risks around violating the local continuity assumptions of RDD, which is critical for identifying the model. I'm going to email this question to him today - because it has implications for one of my dissertation chapters.

I used to talk a lot on here about the problems posed by what I called temporal autarky...

...but sometimes unidirectional time is a good thing (also see Season 5 of Fringe):
 

For you guys that like betting on predictions...

... Kate and I have a bet. Based on what I've heard about how mothers carry the baby and a few other factors, I am guessing we're going to have a girl. She says a boy. The loser has to partake in an activity of the winner's choice.

On the ride in this morning I would be doing a yoga class and she would go to a historical site with me (I had mentioned Mount Vernon).

Now she tells me it will be pedicures...

...the stakes have been raised, so I think I'll have to figure out something a little less user-friendly than Mount Vernon (that's to their credit, of course - they are one of the best in that department... but it won't do in a bet where I'm facing the prospect of a pedicure).

We have awesome baby names for both, so either way it will be fine. And if it's a boy he'll come into this world knowing his father has beautifully maintained toes (pedicures are toes, right?).

Thoughts on the Chavez school econ project

I had a great time yesterday morning at the Cesar Chavez charter school. I was at the Parkside campus, in Anacostia. The set-up was a little different than I expected. They had us each sit with a group of four or five students. The students took turns presenting their reports/proposals, and then we asked them questions about it, lead a discussion, and evaluated them. I heard from nine students overall. The teacher assigned them which policy proposal to defend, which made for a nice diversity of arguments. Many of them disagreed with what they were assigned and did a good job explaining why in the discussion section.

One of the things that I found fascinating was the difference between my first session of four students and my second session of five students in how they thought about the social insurance purpose of Social Security. The first group was adamant about this, to the point of being very skeptical about differentially taxing or providing benefits to wealthier beneficiaries or even raising the payroll tax cap (it wasn't clear to me when they were presenting that they all understood how the cap worked, though, but we went through that in the discussion). This did not surprise me - this was part of their Great Depression studies in their history class, so they had been talking a lot about the purpose of the program, which was social insurance.

The second session was very different - very willing to consider the redistributive aspects of Social Security, and take them into account when thinking about policy. There was one student that seemed to be familiar from outside school with the disability side of the program, and that colored a lot of the discussion.

I obviously consider Social Security a social insurance program, but of course I didn't propound on that or anything. In the first group we talked how we might address poverty issues if we don't do it through Social Security and the in the second group we talked about the same thing (if you couldn't address poverty through Social Security how would you do it?).

I did raise privatization in both groups (these kids are so young! they were in kindergarten when that was the hot topic!). They all shot that down pretty quickly even though many of them didn't even know what it was at first (and no, I didn't push them). I think the recent crash (which they were well aware of) had a lot to do with that. They saw right through the risks involved with moving money out of the trust fund. I actually had to coax them into thinking about why anyone would propose it in the first place - the point that there are actually years when financial markets do well too and that in the long run it does well!

We also got into discussions about COLA (the second session actually had a proposal around COLA - the first didn't). We didn't get into chained CPI or any of that - the proposal was just framed as "slowing the cost of living adjustment", but we did talk about the differences in the prices of the things that seniors spend on and the things that non-seniors spend on, and the implications for COLA. I've had an ongoing facebook discussion with a fellow GW alum about this - my view is that chaining is a great idea - the problem really is in the basket used (regular CPI probably isn't the best). I get fussy when progressives dump on chaining when the real problem is the basket. Chaining is a good idea.

Ryan on Pete Boettke

Here.

Very good thoughts, although I'm not sure Boettke is going the full Sapir-Whorf. Certainly a dominant discourse shapes the way we think about things even if the fact that Eskimos have fifty ways to say "snow" doesn't mean as much as some people like to think it does. It's almost tautological - we think in sentences and the sorts of sentences we put together a lot are going to have an influence on the thoughts we think a lot.

What's wrong is the extreme nature and false dichotomy of the sentences Boettke attaches to economic scientists of the twentieth century.

Two links on children and the crisis

- Simon Johnson on the impact of austerity on children's programs. Anyone who thought that Thomas Sowell's article just nailed Obama because it pointed out that Obama ignored a "solution" to the problems posed by austerity by rejecting the GOP bill needs to read this article. The GOP bill was rejected because it made one part of sequester less biting and arbitrary and another part of sequester more biting and arbitrary. A solution to sequester would be not having sequester.

- Julia Isaacs, of the Urban Institute, has a report on the incidence of unemployment from the perspective of children.

A question that always pops into my head when reading stuff like this from Peter Boettke

(this)

Can he even conceive of an economist that deeply respects the thinking and legacy of Keynes, Hayek, Buchanan, and Friedman or does he just dismiss that as a nonsense position?

In the realm of the history of economic thought - which I've been moving out of lately for career-building reasons but would love to do more with at some point - the thing I probably hate most of all is turning the history of thought into a cartoonish battle of the titans. That's really not how intellectual history works, IMO.

Rosnick on Work Hours and Climate Change

David Rosnick (CEPR, fellow GW alum), a reader of the blog but who usually sticks to emailing me comments, has a new paper in the Real World Economics Review on work hours and climate change.

I imagine this is something Keynes would have really gotten on board with if he knew anything about climate change.

Even aside from working hours, this is a good argument for telecommuting too.

Lots of good working papers today!

These ones look particularly interesting to me:

- The Great Reversal in the Demand for Skill and Cognitive Tasks by Paul Beaudry, David A. Green, Benjamin M. Sand

- Bubbles, Crises, and Heterogeneous Beliefs, by Wei Xiong

- Rethinking Elderly Poverty: Time for a Health Inclusive Poverty Measure?, by Sanders Korenman, Dahlia Remler

- A Spatial Approach to Energy Economics, by Juan Moreno Cruz, M. Scott Taylor

- Subjective and Objective Indicators of Racial Progress, by Betsey Stevenson, Justin Wolfers [the subjective well being gap has closed a lot since the 1970s, and what remains of that gap is pretty much explained by objective differences - so closing the rest of the subjective well-being gap will probably have to rely on improving economic performance].

- Inefficient Hiring in Entry-Level Labor Markets, by Amanda Pallais [This is relevant to the minimum wage discussion with Gene Callahan the other day, and another explanation for why unemployment is higher for low skill or young workers who have trouble signaling their productivity]