Saturday, February 18, 2012

To: John Cochrane, Re: Positional Externalites


Years ago, when your excellent book, Asset Pricing, first came out, I was a finance PhD student. I found some errors and sent you an email with them, and some feedback and suggestions I had. You replied, and were very gracious. So I'm hoping you might reply to this.

In your current blog post, you write:

Frank's article is hilarious in another way. Higher taxes are fine, he says, because more money won't make you feel better when everyone around you is wealthier too. Too much low-hanging fruit there, just go read it and have a laugh. Or shake your head in amazement. No, he's not joking.
With regard to Frank's contention that position/rank/prestige/context is a substantial part of how much utility the average person gets from goods, please give us the evidence why this is wrong. Please don't just say, ha ha, it's hilarious, it's too obvious. Because Frank's contention was published in one of economics' top journals, the American Economic Review, "Positional Externalities Cause Large and Preventable Welfare Losses." (2005).

And other authors have based top journal papers on the same contention. For example:

"Neighbors as Negatives: Relative Earnings and Well Being", by Erzo Luttmer, Quarterly Journal of Economics, August 2005.

"Diamonds Are a Government's Best Friend: Burden-Free Taxes on Goods Valued for Their Values", by Yew-Kwang Ng, American Economic Review, March 1987

Now, surely the wrongness of this idea can't be so laughably obvious if the editors of some of economics' top journals repeatedly published articles based on it. And, not to beat a dead horse, Nobel Prize winning economist Gary Becker wrote a paper based on this idea:

"Evolutionary Efficiency and Happiness", Journal of Political Economy, April, 2007 (with Louis Rayo)

Quoting Becker and Rayo:
For a long time, utility was assumed to depend only on the absolute level of an individual’s economic conditions. However, a large body of research now shows that the relative level of these conditions also plays a central role: an individual’s utility, whether defined in terms of decision making or hedonic experience, tends to be sharply influenced by his personal history and social environment. Examples include Markowitz (1952), Stigler and Becker (1977), Frank (1985), Constantinides (1990), Easterlin (1995), Clark and Oswald (1996), and Frederick and Loewenstein (1999).
-- pages 302-3.

So please John, don't treat Frank's general idea as laughably, obviously wrong. Becker's not an idiot. None of the very successful economists noted above are idiots. The editors of the AER, the QJE, and the JPE aren't idiots. So if positional/context/prestige externalites are insignificant factors in peoples' utility we're going to need your evidence for that.

For me personally, I'd be stunned if position/rank/context/prestige were insignificant or insubstantial factors in how much utility people get from clothes, cars, homes, etc., if the utility a person gets from a given house or car is not substantially different if it's in the 90th percentile or the 10th. I would be at a complete loss to explain all that I've seen, read, and experienced, day by day, in 48 years in this world. It would be hard for me to think of any alternative hypothesis that would fit it. But I'm very open to your evidence, and very curious to hear it.


Richard Serlin