Intermediate Disturbance, NAIRU, and the Laffer Curve

Over the weekend, there was a lot of discussion of those ridiculous conservative faithtank graphs that were rerun in the Wall Street Journal. Several of my fellow ScienceBloglings have debunked the analysis that claims these data support the Laffer curve, although my favorite criticism is by Brad DeLong who points out that to prove something the editorial writers like (the Laffer curve), the Wall Street Journal editors use the the Norwegian data, and to weaken something they don't (increased corporate taxes lead to increased tax revenue), they remove the same data.

Is there any question how we wound up in Iraq? But I digress. I had sort of hoped that the Laffer curve was no longer being discussed seriously (although who takes the Wall Street Journal editorial page seriously...). From the archives, here are some thoughts on the Laffer curve:

There's a class of hypotheses out there which have always bothered me because they do not offer prospective predictions (they usually can fit past data). These hypotheses usually aren't wrong, but they have very little predictive power (as opposed to explanatory power). While I'm going to pick on conservative economics, I'm going to use the ecological intermediate disturbance hypothesis as my first example (and to show that it's not just economics that has daffy ideas).

Wikipedia has a very succinct description of the intermediate disturbance hypothesis:

The Intermediate Disturbance Hypothesis is an ecological hypothesis which proposes that that biodiversity is highest when disturbance is neither too rare nor too frequent. The notion that disturbance can increase biodiversity opposes the older idea that diversity is highest in undisturbed ecosystems. It was proposed by Joseph Connell in 1978, drawing in part on Henry Horn's 1975 paper.

To give an example, rocky ocean beaches ("rocky intertidal systems") that have heavy wave action have very few species of foliose algae (the floppy seaweed stuff) because only a few species can actually hang onto the rocks. Likewise, rocky intertidal systems that have very light wave action also have very few species of foliose algae because the environment is very benign (algae aren't being ripped off the rocks), so what matters is competitive ability: there will only be a few top competitors. Habitats where there is intermediate wave action have the most species because neither the 'clingers' nor the 'super-competitors' will be able to become dominant; the environment will vary enough such that one group will not be able to dominate and drive the other to extinction.

(note: I have greatly simplified the explanations for these patterns-perhaps embarasingly so. Decades of ecological research has revealed that the interactions are far more complex. For example, predation isn't even considered here.)

The intermediate disturbance hypothesis does seem to account for foliose algae biodiversity, but what's interesting is what happens with the crustose algae (the typically pinkish stuff that looks like someone covered a rock with several layers of paint). High wave action for the floppy seaweeds is a bonanza of biodiversity for the crustose algae. (In fairness, there probably is some environmental variable for crustose algae that yields the same pattern, but my marine ecology is rusty these days). The point is that while the intermediate disturbance hypothesis is verry good at explaining foliose algal patterns, it would be very hard a priori to predict which habitats have enough wave action to lower biodiversity. Yes, if you look at enough beaches and seaweeds you can figure it out, but that's not derived from the theory, but from repeated and sufficient empirical observation.

On to NAIRU. I've discussed NAIRU before, and I'm feeling lazy, so I'll quote myself (the linked post also discusses the moral implications of NAIRU, but that's not relevant here):

NAIRU is the Non-Accelerating Inflation Rate of Unemployment. NAIRU is also referred to as the "natural rate of unemployment." Essentially, once unemployment drops below a certain rate-and there is little theory to indicate just exactly what this rate is-inflation will rise. I've never really liked NAIRU because, during the Clinton era, it kept being revised downward as unemployment fell and inflation didn't soar. Nonetheless, it was essentially a cornerstone of Fed policy under Greenspan. As a result of this concept, the Fed would often attempt to 'slow' the economy through monetary policy (i.e., slow or even stall job creation).

There is probably some use of NAIRU. It doesn't seem unreasonable to think that as we approach real zero unemployment, wages could become too high. (an aside: on the other hand, too much unemployment and things obviously become very bad-we know this does happen). But, like the intermediate disturbance hypothesis, there really isn't a good predictive tool for this rate: is it 1%, 5%, and so on. During the Clinton era, estimates of NAIRU kept being lowered (six percent to three or four percent) as unemployment dropped and the economy didn't implode.

Now, the Laffer curve. Here's another good Wikipedia summary (italics mine):

The Laffer curve, popularized and promoted by economist Arthur Laffer and often used to justify tax cuts, is intended to show that government can maximize tax revenue by setting a tax rate at the peak of this curve and that raising tax rates further actually decreases revenue. The idea is clearest at both extremes of taxation--zero percent and one-hundred percent--where the government collects no revenue. At one extreme, a 0% tax rate means the government's revenue is, of course, zero. At the other, where there is a 100% tax rate, the government collects zero revenue because (in a "rational" economic model) taxpayers have no incentive to work or they avoid taxes, and the government collects 100% of nothing. Somewhere between 0% and 100%, therefore, lies a tax rate percentage that will maximize revenue.

The point at which the curve achieves its maximum will vary from one economy to the next and is subject to much theoretical speculation. Another contentious issue is whether a government should try to maximize its revenue in the first place.

First, a snarky aside. If you haven't been politically comatose since 1981, you'll realize that conservatives always argue that we're on the side of the curve where taxes should be lowered. It never occurs to conservatives that we might be on the other side of the curve, where taxes need to be raised. Funny how that works. Here too, we see the same problem: there is no good way to estimate where you are on this curve. Empirically, we can determine this. For example, we are running massive deficits, in large part, due to the Bush administration tax cuts (i.e., a decrease in revenue). That's too much tax cut.

To end this disparate, wandering post, I'll leave you with some advice: if you want to make policy, use a predictive hypothesis that can a priori provide estimates of where you actually are starting from and where you hopefully will be heading.