Government Efficiency: A Natural Experiment

...sequestrian offers a "natural" experiment to test views of government...

Microeconomists will thank Congress for providing a natural experiment to test competing views of government. One is that government is unnecessary, or at least inefficient in an allocative sense -- lots of money on things that aren't very useful (and perhaps not enough money on things that are). If so, then sequestration won't matter. Economists also employ a second concept of efficiency, which in layman's terms could be phrased as "doing the job in a lean manner." If the government is inefficient in this sense, then sequestration will not matter: fire 10% of our civil servants and the remainder should be able to easily pick up the slack.

Then come alternative views of the impact of fiscal policy on the macroeconomy, "the" multiplier. Now I highlight "the" because Principles textbooks and macroeconomists who live in other than a blackboard world have maintained since the 1920s that the impact of fiscal policy in a great recession is large, while that in the midst of an expansion is small or nil. There is no "the" there. We again have a natural experiment: if the multiplier is really zero or negative, as claimed by those who opposed ARRA (the 2009 stimulus package), then sequestration will do no harm to overall growth.

I hope the sequestration experiment will not undertaken long enough to actually provide the evidence we economists need. That's because my priors are that (i) the bulk of the non-defense spending of the government is productive, that (ii) our bureaucracy is surprisingly efficient, and that given our lingering Great Recession (iii) the multiplier is 1.5 or greater [that is, we get more than a dollar boost in GDP for a dollar of properly accounted stimulus]. So we will (ii) see disruptions that (i) will matter and (iii) that if they go on for long, will aggregate through lower expenditures to have a measurable impact on GDP.

So what of the natural experiment jargon? Well, identifying the direction of causation in economics is really hard. Fiscal measures are affected by a recession. Since taxe receipts are dependent on income, budget deficits explode. While expenditure changers also occur, some of those are a function of our growing economy and population, hence "endogenous". Furthermore, new policies are implemented only slowly; measuring "real" stimulus turns out to be hard. Statistical tests therefore seldom convince open-minded skeptics.

In contrast, sequestration will offer a sharp "exogenous" change with discrete timiing and clear numbers. Of course there are lots of other things affecting the economy, and so a one-time event won't be an unambiguous cause. If sequestration extends for months, however, we'll have another data point generated when it ends.

To reiterate: my priors are that sequestration is a big deal. So I hope that Congress (or more specifically, the House) sits down with the Administration in short order, and leaves us economists bereft of data.

...Mike Smitka...