Next, conduct a thought experiment. What happens if there is an increase in the aggregate stock of liquid assets, say because the Treasury issues more debt? This will in general reduce liquidity premia on all assets, including money and short term debt. But we’re in a liquidity trap, and the rates of return on money and short-term government debt are both minus the rate of inflation. Since the liquidity payoffs on money and short-term government debt have gone down, in order to induce asset-holders to hold the money and the short-term government debt, the rates of return on money and short-term government debt must go up. That is, the inflation rate must go down. Going in the other direction, a reduction in the aggregate stock of liquid assets makes the inflation rate go up.
Do read the rest for the full picture. I understand how this works if one postulates a strict equality between (risk-adjusted?) rates of return on currency, T-Bills, and other assets in the economy. But given current T-Bill and inflation rates, that is going to mean negative real rates of return on a variety of other physical assets, which I take to be at variance with some rather overwhelming data to the contrary, including stock market returns, internal hurdle rates at corporations, investment trends, measured profits, asset prices, and so on. I also fully endorse Scott Sumner’s point that short-term asset price and exchange rate reactions to QE pretty clearly show it is expansionary, although by how much is arguably not clear.
The most likely alternative, in my view, is that some kind of funny asset market segmentation is going on, and that risk premia and collateral demands for T-bills are high in funny ways, so that (risk-adjusted?) rates of return are not so strictly negative all over. What that means concretely is that if the Treasury issues more debt and liquidity premia on debt fall, there is not an offsetting shift in the rate of price inflation to restore the previous set of relationships across asset returns; rather some of the incidence of the quantity change falls on T-Bill returns themselves. I have yet to see that well-modeled and do not have a preferred approach of my own which will in some simple way preserve a zero profit condition. (We also may need to talk about knife edge conditions and whether these rate of return equalities must be truly and exactly equal for the model to work, besides how well do we understand the substitutability of money and T-Bills anyway?) Williamson may damn me for that non-foundational approach, but I see that as the least indefensible answer to this question given our current state of knowledge.
I like many of Williamson points in macro, but in general I prefer to put the empirical evidence in the driver’s seat more than he does. He does write:
What’s the qualification [to my argument]? There are various short-run effects of monetary policy that could come into play. However, I think it’s fair to argue that any of those short run effects have played themselves out in the financial crisis and its aftermath, and now we’re looking at the effects I’ve described.
That’s the right question, but it’s not nearly a close enough engagement with the evidence, which does pretty clearly show some short-run effects are still operating, even if those effects are diminishing with time. The evidence also does not clearly indicate that the long-run effects (e.g., check out the term structure of interest rates) have to move in the direction Williamson is indicating.
It’s still a puzzle exactly what is going on, but we need to be very careful about how we use the overly seductive argument from elimination. When in doubt, reread David Hume.
Addendum: I do get Scott’s and Yichuan Wang’s argument about the hot potato effect for monetary injections, but I am not sure this can handle the case of an increase in the supply of Treasury securities. Go back to my earlier remark about how little we understand money-T-Bill substitutability.
And here is more from Williamson.