Friday, January 9, 2009

Madison's Argument for Bigger Govt

One of my usual arguments when discussing systems of government, and incentive systems more generally, is the notion that smaller is better. A local government will be more in tune with the sentiments of the people, hence have better information compared to a government entrusted with a large populace. Better information will allow the government to more appropriately serve its constituents.

This argument faces some opposition in my current read, (American Creation by Joseph Ellis). Its a passage on James Madison's thoughts as he prepares for the Philadelphia convention of 1787.

So Madison reversed the conventional logic [apparently I am one of those conventionalists]. Small republics, like the states, were actually more vulnerable to factional squabbling and sectarian divisions than large republics, because of the larger scale of the enterprise vastly increased the number of competing factions, thereby producing "a greater variety of interests, of pursuits, of passions, which check each other...So an extensive Republic meliorates the administration of a small Republic." It was not just that a fully empowered national government was likely to attract a better class of statesmen more capable of resisting local pressures, though Madison believed that was true too. More fundamentally, an extended American republic, contrary to the prevailing wisdom, made space an asset rather than a liability. (p. 105-106)

I interpret this as nothing other than the age old wisdom of the law of large number, or the diversity principle, whereby increasing the quantity of data reduces expected variation. This is a nice point, but there is a fundamental difference to this view as compared to the conventional view I briefly outlined above. Madison is looking at the negative of government involvement. More specifically the use of government to exact concentrated gains from the unconcentrated whole.

Clearly, this is an excellent point. A very useful way to approach government action. On the other hand it is hardly the whole story. Politicians will always have an incentive to provide valuable services to thier constituents in order to promote re-election. (However, a valuable service must be easily traceable to its source in order for it to be effective. Also, as is all too apparent, politicians may make actions that appear valuable, but in reality are destructive.) The worse the politician's information and the less percentage any single constituent makes up of the whole, the less a citizen can expect his preferences to matter.

Admittedly, this is a big question and I have given an short answer. But, it may reveal a bit of the flavor of this balance/tradeoff.

Which is better, big or small?

No comments: