Sunday, 11 January 2009

Rigour? In a blog?

Brad DeLong has an impressively clear (although, or perhaps because, a bit technical) piece on risk appetite and fundamental values of assets. I feel I should respond to it, but it's such a tour de force of clear and explicit assumptions, rigorous logic and thunderous conclusions that I feel I barely can.

I particularly loved this statement for pure bloody-minded literalism (that's a good thing, by the way):
The fundamental values of asset prices are the money-metric values that the costate variables associated with the commodities would have in some reasonable utilitarian central-planning social-welfare-maximization exercise under reasonable utilitarian preferences.
However, here is what comes to mind:
  • Given that asset prices rarely approach the values Brad ascribes to them, could this be because there is competition for the limited amount of savings capital available in the world? If I, as a productively investing business, want to get access to some of it, I probably need to pay a lot more than 2% (I have anecdotally heard figures as high as 19% for small business loans recently, and that's for a loan 75% guaranteed by the UK government; perversely, I also saw a firm quote of 11.49% for an unsecured loan). Then again, I can definitely make a more than 2% return so I wouldn't mind paying more.
  • Is it true that default risk is or should be minimal? One argument quoted (by David Smith of EconomicsUK) against a proposed £50 billion state guarantee of business lending is that the Treasury estimates losses of up to £12 billion on it.
  • While long-run productivity growth should indeed be the guide for average investment returns, a dynamic economy is likely to produce lots of businesses which increase productivity by 5, 10 or 40% a year, and others which stagnate, shrink or fail. More and more investments in the context of a 'knowledge economy' are intangible assets which are mostly written off in the latter scenarios, and thus the successful businesses need to generate much higher returns to compensate. This implies both demand for, and supply of, higher-yielding assets. And again, a high default risk.
Against all these points, Brad is probably talking about average returns for (say) equities across the whole class, rather than expected return of any individual asset. If half of today's equities will be worthless in 10 years, then they each need to have an overt return much higher than the 2.3% target, in order to average out.

(An interesting aside: this argument indicates that equities should have lower annual returns in the short term than the long, because the long-term return has to take into account a higher risk that the company won't survive that long. However, equity investors, especially in the angel and VC markets, always insist on high early returns, to minimise their risk and force the company to reveal/prove information that reduces uncertainty.)

But perhaps more fundamentally, I suspect the information discount is higher than he believes it should be. Risk is easier to arbitrage, or average, away than information is; a big information gap not reflected in his analysis is knowledge about future behaviour. This makes it hard for people to optimise consumption to their ideal time preference - you might die next month before cashing in that T-Bill, or Ben might inflate it away. It also makes competition between assets difficult, and that may lead to a rational undervaluation.

These are just some initial thoughts, late at night and without the level of self-confidence and technical muscle that is bulging from Brad DeLong's posting. He does remind me of Krugman in that way. I wonder whether the Keynesians are simply more confident, have more testosterone, or just more confidence in the system to work right, than the less interventionist Mankiw-Cowen-Kling school.

Thought-provoking, anyway.

No comments: