Expressive Voting, Rational Irrationality, and the Size of Government

<a href="http://econfaculty.gmu.edu/bcaplan/e854/hw3.htm">A recent homework question asked:

Expressive voting and rational irrationality seem to predict that local governments will be more efficient than state governments, which will in turn will be more efficient than federal governments. Are these predictions true? Carefully explain your answer. Is there any reason why we should expect this relative ranking to be empirically small?

I'm not turning in my answer because I couldn't find cold numbers comparing spending patterns of different levels of government on similar projects. I also disagree with the premise of the question. I did, however, want to reflect a bit on the question:

1. Economies of scale and scope give an efficiency advantage to larger sizes of government.
   1. Note: Size generally but not necessarily corresponds with the level of government. For example, \"in 2012, the New York City Metropolitan Statistical Area generated a gross metropolitan product (GMP) of over US$1.33 trillion, while the Combined Statistical Area produced a GMP of over US$1.55 trillion, both ranking first nationally by a wide margin and behind the GDP of only eleven nations, respectively.\"
2. Exit/mobility costs are heterogeneous with the level of government. Obtaining residency in a new country can take a decade, obtaining residency in a new state merely requires a pile of paperwork, time, and fees. Moving cities is relatively cheap, although it's not absolutely cheap. Switching homes can cost thousands of dollars. Mobility limitations allow the level of government in question more bargaining power, so they would tend to make federal government less efficient in theory.
3. Special interests tend to attenuate the differences in the efficiency of spending at differing levels of government by lowering the maximum level of efficiency and thus reducing the possible variation in efficiency.
4. Perhaps most importantly, it's not at all clear that voters will behave instrumentally even at low levels of government.
   1. The theory goes that voters have an approximately 0% chance of swinging a national election, but where does the rounding threshold come in? In reality, expected value doesn't involve a rounding threshold. What matters is the expected value, not the % chance of swinging an election.
   2. Preserving the fallacious rounding technique, individuals still have approximately a 0% chance of swinging a state election. So there is no reason to expect an instrumental voting incentive.
   3. At the lowest level of government there may be some reason for a voter to expect to be able to swing a particular election with a non-zero after rounding probability. It does not follow that they will vote instrumentally. Here is where the rounding fallacy really kicks in. There is so much less to gain by winning a local election. Let's say my vote makes a difference and my buddy Jim wins a School Board position. Wtf does the school board actually do anyway? Jack shit. So they can fire the Superintendent? So what, the Superintendent doesn't do anything either. The School Board might usefully get a Superintendent fired if he does not perform his job as prescribed, but the prescriptions are set at the higher levels of government anyway. So the net effect is that the Board does nothing.
   4. From 3, it follows that while the odds of an individual swaying a local election may be many times better, they also have many times less benefit to gain from such engagement. As a proportional Return on Investment of time, etc, then, I'm not convinced instrumental rationality plays more at the lower or higher levels of government.

In conclusion, I differ on the seems of the question, that \"Expressive voting and rational irrationality seem to predict that local governments will be more efficient...\" Allowing that point, however, any differences will be attenuated by the net combined effect of economies of scope, economies of scale, mobility costs, and special interests.