In Which I Disagree with Paul Krugman and Brad DeLong
This is a rare event. Both Brad DeLong and Paul Krugman argue that a recent argument by Kevin Warsh and Michael Spence makes no theoretical sense. I disagree. I am sure Warsh and Spence are wrong, but I can make a simple theoretical argument for their conclusion.
The argument is that QE causes low non residential fixed capital investment. DeLong and Krugman noted that their is no anomaly to explain — non residential fixed capital investment is what one would expect given the sluggish growth of other components of GDP. In other words, the explanation of why it hasn’t bounced back this recovery is low housing investment and government purchases of goods and services. I am convinced by the evidence Brad presented (click the damn link).
But I am not convinced by his theoretical argument (and not just because it is a theoretical argument)
Krugman agrees with DeLong
QE reduces volatility in financial markets by making some of the risk tolerance that was otherwise soaked up bearing duration risk free to bear other kinds of risk. That is what it is supposed to do. With more risk tolerance available, more risky real activities will be undertaken
“Brad DeLong spends what may be too much time on the latest; it’s an argument that doesn’t make any sense, deployed to explain something that isn’t happening”
I think it is easy to make the argument (although I am do not find it convincing).
My comment on Krugman below (I am less polite to Brad — you can see my comment over at his blog if you click the link).
I think it is possible to write down a simple (not even obviously forced to reach the desired conclusion) economic model in which QE causes low nonresidential fixed investment. Here I define QE as purchases of 10 year Treasury notes & 30 year Treasury bonds (not RMBS).
The argument is that the duration risk in long term Treasuries is negatively correlated with the risk in fixed capital. I think this is clearly true. The risk of long term Treasuries is that future short term rates will be high. This can be because of high inflation or because the FED considers high real rates required to cool off an overheated economy. Both of these are correlated with high returns on fixed capital (someone somewhere keeps arguing that what the economy needs is higher inflation).
This means that a higher price for long term treasuries should make fixed capital less attractive — the cost of insuring against the risk in fixed capital is greater.
It is important that the assets bought via QE are nominal assets with extremely low default risk. The risk in corporate junk bonds is mainly the risk of default. Their returns are similar to the returns on fixed capital (and on stock). If the Fed were to (illegally) buy corporate default risk, it would increase incentives for corporations to invest.
None of this reasoning applies to Fed purchases of mortgage bonds. Also it has no influence on my actual beliefs (it is economic theory).