PolySciFi Blog

Friday, February 20, 2009


Sports Stadiums and the Stimulus Multiplier

I'm not a fan of demand-side stimulus efforts (in my best case scenario, you get inflation which might help with sticky prices / wages) so you can imagine my dislike for the stimulus bill, which seems to be best summarized as "throw a lot of money at pet Democrat projects and see if it helps the economy" and I've been looking for the best angle to convince my liberal friends that this is a *really* bad idea.

So I was happy to see this connection from Russ Roberts (both sets of facts I was well aware of, but I had not made the connection) as the first point is something that I think most liberals acknowledge so maybe the rest of the steps will follow.

1) Spending on construction for the Olympics / sports stadiums are net money losers for a local economy. (The Knoxville World Fair of 1982 is a notable exception to this rule as it netted $64.00 to go along with its wigsphere). Sometimes they're justified as a luxury good for the city (like a park), but they're not certaintly not money makers.

2) Further, the actual # of net jobs created by these construction projects has been shown to be 0. (see the cafe hayek post for a summary or this paper). Once you grant that there's not a net positive economic impact from the government building a new stadium, that's pretty easy to see as the government is just shifting funds around (robbing Peter to pay Paul, as such) and that's what the empirical surveys have shown.

3) So clearly, not all government spending has an associated multiplier that is greater than 1. Personally, I would expect that new spending with a multiplier > 1 to be the exception rather than the rule. Government funded research projects that transition to the private sector (ala the Internet and some drug research) and some infrastructure such as the interstate system (its multiplier is actually decreasing with time due to increasing maintenance costs) are the examples I can think of. But their payback is not immediate, i.e., doesn't satisfy the timeliness goal of a stimulus package.

4) While some infrastructure has a positive RoI, clearly lots do not (such as the stadium, but see also the bridges to nowhere in Japan and almost in Alaska).

5) With the default position that most govt spending will not have a multiplier effect > 1 (which generally follows the rule 90% of all new ideas are crap, though my personal assessment of the ratio for govt plans is higher), what's the likelihood that a hastily slapped together plan patched together from things that had not independently demonstrated their value enough to otherwise receive funding will have a multipler > 1?

Random asides:
The Keynesian idea that government spending has a multiplier > 1 is motivated by the assumption that money flowing through the government has a greater velocity than money flowing the private sector (enough to offset the implicit inefficiency that comes from x spending y's money on something that y likely would not have bought). When people saved money in mattresses, this was true.

But since people now save their money in banks which can and do loan out multiples of every dollar saved (fractional reserves) and government treasuries are 1-to-1 (1 dollar borrowed / spent comes from 1 dollar invested) the velocity argument actually cuts against government spending, and even more sharply when you consider the long appropriation cycles now used by the government (case in point see the stimulus bill allocations out beyond 2011).

Further, if you buy the theory that the problem with the economy is insufficient credit availability, what will be the effect when the government soaks up $1 trillion in investment capital? Even if it's financed from overseas, that's still investment capital not invested elsewhere which decreases the (world wide) aggregate credit supply.


This page is powered by Blogger. Isn't yours?