Thursday, January 08, 2009
Austrian Business Cycle Theory
When I was in charge of capital allocations at KeyCorp, I spoke with many of the business line managers, and was impressed by the fact that all of them had rather sterling track records, especially in the last recession. But I later figured this was all survivorship bias: the losers in the last recession lost their jobs. Thus, each thought they had some special alpha, special asset class, impervious to the mistakes of those who caused the big losses in the past. For a 45 year old who has never really screwed up it's hard not to think this is the case, as opposed to merely the blind selection process of capitalism. This error is the fundamental genesis of business cycles, in my opinion.
In grad school, while learning macro, I would spend nights reading Austrian Business cycle theory, including von Mises's Human Action (900+ pages!), and lots from Hayek (see picture, I don't think they are related but she's a lot prettier). I was intrigued by the idea that business cycles were caused by a misallocation of resources, as opposed to merely 'too much' investment. That is, say you have 10% of the country's resources in internet development, but discover the demand only wants 5%. All is fine as long as you don't care about profits, look at sales/price ratios, but then eventually people get tired of not making money, someone says 'the emperor has no clothes! There will never be profits', and everyone stops investing in these areas. The transition from the old to new regime is only possible via firm failures and involuntary unemployment, because people don't switch to new jobs until their old ones are gone, forcibly.
My only beef with the Austrians is that they emphasized the genesis of this missallocation via money creation, especially the fiat money creation of central banks and how this causes the interest rate to be 'too low', causing overinvestment in capital. This again gets into a straight aggregate overinvestment story, and that isn't very empirically robust. I think to the extent there is overinvestment, it isn't total investment nearly as much as in the wrong sectors. Today, that sector is housing, and finance. The key is finding some metrics one can apply to these subcategories to see that it is overinvested, and if there is some consistent tipping point in such a metric. The problem I see is that for commercial real estate in 1990, or residential real estate in 2008, there were real profits and cash flow immediately prior, based on people buying assets that were overvalued. In contrast, in the tech bubble, there were no profits immediately prior to the correction. So cash-flow in the sector would not work. You could say, use cashflow plus some metric of collateral overvaluation, but it is not obvious to construct a model of this, as even Shiller's 2006 revision of Irrational Exuberance had a rather wishy-washy forecast of housing prices (could go up, could go down) even though in hindsight it appears we were at the peak of an unsustainable housing bubble. I'm sure commercial real estate had a similar issue in 1990.
But, I think that idea, of a missallocation, is still the best explanation of recessions, and thus suggest a laissez-faire approach because until these guys lose their jobs, and firms fail, you won't get the appropriate (dare I say optimal) reallocation of resources. Looking at the current housing bubble, and all the strange incentives and errors by investors, legislators, regulators, etc., the missallocation's genesis is not unicausal, but one key is that it generally is in an industry that did well in the prior recession. Thus, just as commercial real estate CDO's had very low relative loss rates relative to other CDO's subsequent to 1990, I suspect that buying residential real estate CDO's now is probably one of the better investments around if you are in that space. Beware of the outperformers, as failure is endogenous to any business plan because success breeds overconfidence via the overattribution of their success to alpha versus random luck, though as always luck, skill, and effort are relevant.