Paying Fixed Costs Weblogging, Quant. Economics Using Python Edition

I successfully made my first annotated graph of something using pandas, datetime, matplotlib, and some other libraries and pulling data from FRED without ever looking at it. It’s pretty exciting. Here’s the graph (click to enlarge):


I’m not really contributing very much to this argument (Dean Baker, Matias Vernengo, and Media Matters are already all over this), but the last hour spent playing with matplotlib.pyplot.annotate had to produce something.

In other news, Quantitative Economics Using Python (Stachurski and Sargent) is a blast and the most fun I’ve had learning at least since May.

Terrified that we might be wrong about the shutdown webblogging, rd. 1

A lot of reporting on the shutdown has focused on the irrationality of the GOP position on several fronts: first, the irrationality of their demands; second, their shifting targets; and third, how the GOP has become a party of petty extortionists. I think all of these approaches miss the point when a much simpler explanation is available. It’s easy to rationalize Ted Cruz’s and the House GOP’s actions if you guess that they don’t actually want the government open.

Or, simplified, this is the situation everyone seems to be imagining:


Both sides want an open government, each side has some sort of binary issue it cares about that the other hates, there’s an issue of degree where there’s room for each side to give a little, and a deal looks not only possible but likely. If this were your starting position, you’d be very confused about why Republicans seem uninterested in making “reasonable” demands.

If, however, this is more appropriate:


…we have a much more challenging situation. There’s nothing both sides want, and since an open government is something only Democrats want (and it’s not unreasonable to think of an open government as the  biggest thing on their list of demands), it requires a roughly comparable concession in order for both sides to feel like they’ve gotten something. If Democrats don’t understand that they’re the only side that wants an open government, they will continue fundamentally to misunderstand what the GOP is up to.

But why the House GOP would want a closed government probably requires some explanation. I have two. First, the GOP still labors under the myth of ever-expanding government payrolls and expenditures, while neither of these appears to be accelerating:

It hasn’t mattered at any point during the recession whether this myth was true. If this is the world in which the GOP lives and both need to be cut, a government shutdown gives them larger reductions of federal employment and expenditure than any deal they could possibly extract.

Second, inasmuch as Ted Cruz is proud of pissing off everyone he didn’t “push into traffic and walk away” from, there’s little reason to believe that he has any plan related to actual policy. It’s inane, posture-heavy and substance-light, but if Cruz is contemplating his future as a non-Senate candidate, this is an episode he can point to in which he “broke with the Washington establishment” or something equally hollow. Further, election season is a long time and probably at least one more debt ceiling crisis away. If Anthony Weiner can make a serious push for NY Mayor, I don’t know why we’d expect serious long-term consequences for Cruz, while, with the benefit of some distance in time, he can probably spin his intransigence into “principle.”

There’s a possible third reason in which Rs just want to throw 800,000 employed people — or about 5 months of average job growth — out of the labor force for a lil bit in an effort to keep things awful for the 2014 midterms, but that sounds almost evil, so I’m going to leave that one.

None of this suggests irrational behavior. The Republicans responsible for the shutdown are less baffled by the way public opinion has turned against them than they are preoccupied with furious masturbation to their destructive fantasies. There’s no reason to believe that explaining the political realities of the shutdown to them will change anything, because they aren’t concerned about the political realities (and they already know). The assumption here that the obstructionists don’t know what they’re doing is lazy and destructive to good analysis, and deals predicated on the belief that Democrats and (radical) Republicans want one of the same things will invariably fail unless Boehner and a collection of Senate Republicans in an instant take back control of their party.

Failed State Index and Gas Prices

I was annoyed about a Washington Post article yesterday with the headline “Mounting Tensions with Syria Sink U.S. Stocks.” The reasoning was

“If Syria becomes drawn out and becomes a long-term issue, it’s going to show up in things like gas prices,” said Chris Costanzo, investment officer with Tanglewood Wealth Management.

What was confusing for me was why conflict in Syria hadn’t already shown up in gas prices, which I suppose are somehow deterministic on the value of U.S. stocks. Are investors unfathomably short-sighted? Do they not read news? Are they simply bad at forming expectations? Anyway, I wasn’t sure I bought the logic.

I started hunting around for data to use and found the Fund for Peace’s Failed States Index, which has to be one of the most depressing datasets to assemble. The data only go back to 2007 with any degree of completeness. I weighted values based on percentage of total oil export for each country obtained from this Wikipedia list and graphed average failed state index of oil exporting countries against U.S. gas prices (obtained from FRED). This is what I wound up with:



Admittedly, the FSI is only reported annually, is a pretty short dataset at this point, and may be poorly constructed (though I don’t know why we’d assume this), but this doesn’t look like much of a trend anywhere. If investors react to anything more stable than immediate stimuli, the logic of “Aaahhh! Tension!” –> gas prices sucks.

A better analysis would use information about from whom the U.S. imports oil, rather than simply take global information and hope for the best, but that’s what people do in real research projects.

Rick… Baumolling.

I don’t think that pun works. Anyway, per this:

Cost disease jokes are COMING. To save time later when arguing that I was into the Arcade Fire cost disease as an explanation of rising tuition costs outpacing broad inflation long before they won a Grammy Wonkblog started on it, here’s a very brief discussion of it from the Spring 2012 Undergraduate Economics Club Debate:

The theory of cost disease is that countries with high productivity in the manufacturing sector must have higher than average wages in the service sector. The reason behind this correlation is that high productivity in manufacturing allows the manufacturing sector to offer higher wages, which the service sector must match in order to attract talent. It isn’t chauvinism to argue that the “high productivity” description would apply to the United States. The theory is simple, so the question is whether the data bear the theory out.

The chart behind me1, from a paper from the College of William and Mary called “Explaining Increases in Higher Education Costs,” shows price levels in several sectors from 1949 to 1989. The four categories whose costs grew the most in the period are legal services, higher education costs, brokerage services, and physicians fees. Of those four, the only exception to the accelerated price schedule is brokerage services, which displayed the phenomenon for the first half of the period and then fell off. The reason its price fell is simple. With the increase in the number of over-the-phone and computer-mediated trades, brokers’ productivity in the period skyrocketed, bringing the field’s wages closer to high productivity industries. The other industries, each requiring highly-educated service professionals, depending on what you think of lawyers, doctors, and university personnel, have not had the same success in adapting new technologies into the delivery of their products.

The problem they face is the same as that faced by a quintet playing a half-hour piece, requiring 2.5 man hours of labour. They could increase their productivity by playing faster, but “any attempt to increase productivity here is likely to be viewed with concern by critics and the audience alike.”2 In the analogous case, professors could teach twice as many students at a time, but not without sacrificing the quality of the product. As a result, despite a relatively stable marginal product of labor for professors, the price of what they provide must rise in order to keep them from wandering off to higher-productivity and higher-paying fields.

1 cost disease chart

2 Baumol’s joke, quoted here

Anyway, Wonkblog makes much prettier charts than the one above and will probably have a whole ton more data (we were limited to 5 minutes per speech and had a lot of other things to get through). I think this is a grab at cred in the favorite-explanations-of-economic-phenomena department. There’s cred in the favorite-explanations-of-economic-phenomena department, right? I mean that’s a thing that exists?

The HAPs, my friends

Thomas Herndon is in my micro class you guys. I think that makes me famous.

The Herndon-Ash-Pollin paper refuses to fizzle and now has Mark Thoma attempting to explain why macroecon doesn’t work very well, with a focus on the difficulties in distinguishing actual causal relationships from historical patterns that by chance resemble causal relationships, problems with the ceteris paribus assumption, changing policy regimes, and one of the saddest sentences written about current macro:

I used to think that the accumulation of data along with ever improving empirical techniques would eventually allow us to answer important theoretical and policy questions. I haven’t completely lost faith, but it’s hard to be satisfied with our progress to date.

(sad Mark Thoma is sad)

Thoma can hardly be blamed for hoping that accumulating information would fix our macroeconomic problems. Here’s Kenneth Arrow in 1957:

“there is a cumulative growth of knowledge; as the number of observations gets larger, there is an increase in the efficiency with which we estimate our parameters. The information used this way may be termed cumulative information… The third kind of time is cumulative time, due to the growing knowledge of the laws of nature, not only in economics proper but also in the technology of production and presumably in consumer and worker psychology. A statistician may view cumulative time as increase in sample size… The concept of cumulative time has played little formal role in economics, though references to the growth of knowledge as the greatest source of wealth are to be found among the less orthodox economists.1 Recent empirical studies have laid great stress on the cumulation of information as the main source of economic growth. Of course, these studies refer to technological knowledge; I would like to believe that such cumulation can be as characteristic of economic as of physical knowledge… if I understand some scientific philosophers correctly, the accumulation of information and the passage of time are really the same thing” (Arrow 1957: 526-7)

It may be difficult to claim that we (as a species, as thinkers about the economy, as ____) don’t accumulate information with time. The link between accumulation of information and accumulation of knowledge, though, is less robust than would be convenient for academic work. Thoma’s concerns about causal mimics and causes apply just as much to the HAP as they did to RR. Because Thoma’s concerns address problems historical data rather than problems with particular economic theories or econometric techniques, there doesn’t seem to be much that can be done to change practice in a way that dodges the historical data critique. With this in mind, it’s difficult to come up with an environment in which asserting “P is true” shouldn’t be met with a healthy amount of skepticism. But the uproar about HAP suggests that we’re — with the exception of Reinhart and Rogoff — much more willing to feel certainly about “P is false.”

I’m not certain what’s left in the wreckage of RR because I’m not certain what it supplanted when it became gospel. With that said, the void it leaves should open space for a new and/or different positive model to guide some policy, at least in a world in which economic policy decisions are made according to academic reasoning rather than vague pandering about low taxes and decreased spending.

1 – Arrow meant Veblen, but he could easily have been foreshadowing or tipping his hat to human capital theories which were only beginning to circulate.