George Selgin has some doubts:
Both Market Monetarists and Wicksell himself differ from the Austrians in paying little if any attention to the direct bearing of short-run interest rate changes on real activity, and especially real activity in asset markets. The Market Monetarists consider interest rate movements an unimportant sideshow without significant knock-on effects, and therefore a distraction from what really matters; Wicksell considered them only as a harbinger of changes in spending. Missing in both perspectives is any attention to the way in which interest rate movements redirect demand from certain markets to others. Monetary policy innovations can, in other words, involve both (nominal) “income” and “substitution” effects.
They needn’t do so, of course. “Helicopter” money approximates the case in which monetary innovations boost nominal wealth, and thereby stimulate spending and nominal income, without necessarily involving any particular relative-price changes. The short-run/long-run AS-AD framework tells us all, or almost all, of what we need to know about the potential real consequences of helicopter money drops; and deviations of nominal income from trend supply a reliable indicator of the degree to which monetary policy has been either excessively easy or excessively tight.
But open-market purchases aren’t helicopter drops. Instead (as Wicksell took for granted) they initially involve increases in the relative price of the securities being purchased, and corresponding reductions in market-clearing (but not in underlying “natural”) interest rates. Lower interest rates in turn encourage a re-orientation of spending toward investment, and especially toward those prospective investment projects whose present values increase most in response to a marginal reduction in the cost of funds. This “substitution” effect of easy money–an effect that depends on real interest rate movements rather than on changes in aggregate spending per se–is the key to unsustainable asset price movements, where “unsustainable” indicates a movement than can only go on while interest rates remain unnaturally low. It is possible, at least in principle, to conceive of a monetary policy that gives rise to large substitution effects–that is, to a substantial increase in the perceived present value of particular investments–while having only a modest ultimate effect on the growth rate of nominal final income. The narrower the initial credit channel through which excess liquidity is injected into the economy before spreading out to the rest of the economy–the further removed we are from helicopter drops–the greater the likely importance of relative-price and substitution effects.
The possibility of substitution effects stemming from “unnatural” (monetary-innovation based) interest rate movements supplies reason for taking even modest innovations to NGDP growth, and upward innovations especially, seriously. The possibility suggests that such deviations are likely to be associated with disproportionate deviations of total spending—that is, of spending on both final and intermediate goods—from its own trend. In so far as money supply innovations tend to drive interest rates either below or above their natural levels, increases in the growth rate of NGDP and other nominal income measures may understate the extent to which monetary policy is excessively easy or excessively tight (and are likely to continue to understate the laxness of monetary policy while a boom persists), because the amplitude of short-run deviations of total spending from trend will be greater than that of nominal income, and because velocity and money multiplier declines that typically accompany the bursting of asset bubbles will suppress the acceleration in nominal income growth that might otherwise be observed once substitution effects have worn off.
I look at things a bit differently:
1. I think NGDP tells us pretty much all we need to know about whether money is too easy or too tight.
2. I don’t think that the method by which money is injected is very important, at least during normal times when interest rates are positive. That’s because the injections are so tiny that the direct effect on the market of whatever’s being purchased is small. Consider the Fed’s options in 2003. It could have aimed for a 5% NGDP growth rate over the following three years. It actually produced a growth rate of over 6%. How would interest rates have been different if the policy generated an expected 5% NGDP growth rate? Hard to say. The slightly tighter money would have resulted in slightly higher nominal interest rates. The slightly lower expected NGDP growth would have resulted in slightly lower nominal rates. The net effect? I can’t say, I’m not even sure whether rates would have been higher or lower with slightly slower NGDP growth. They probably would have bought slightly fewer T-securities, but the difference would have been trivial in the vast T-securities market. The effects would have been essentially identical if they had purchased copper or oil with those extra purchases. Just to be clear, I’m not saying a permanent switch from T-bonds to oil would have no impact, I’m saying that if the differential in base demand between 5% NGDP growth and 6.5% NGDP growth had been diverted from T-securities purchases to oil purchases it would have made almost no difference. Money matters because it’s money, the medium of account, not because of what is bought as it’s injected in the economy. Even under the gold standard, when newly injected money purchased gold, short terms rates used to fall after monetary injections. In my view the liquidity effect on short term rates occurs because prices are sticky, not because of what the Fed happens to purchase.
3. We know that 6.5% NGDP growth would not, ceteris paribus, produce a housing bubble. Indeed most other economic expansions saw faster NGDP growth, without a housing bubble. So the argument rests on the proposition that the lower interest rates resulting from easy money produced the housing boom. That’s possible, but the difference in the level of longer term real interest rates between an expected 5% and 6.5% NGDP growth is likely to be very small, regardless of what the Fed buys. So I simply don’t see how it comes close to explaining the housing boom. The current price of homes is closely linked to what people think homes will be worth in 5 or 10 years. And I don’t see how a year or two of easy money right now would have much effect on the expected value of homes 5 or 10 years out, unless it led to expectations of very fast NGDP growth. But it didn’t. Of course the weakness of my argument is that I have no rational explanation for the housing bubble. I’ve discussed bad regulation and tighter zoning and rapid Hispanic immigration to the 4 subprime states, but I don’t really think those explanations are adequate. It’s a mystery to me. Had real housing prices been strongly affected by monetary policy in other periods of US history I’d been very sympathetic to this argument.
I do understand why commenters who are closer to the Austrian tradition don’t find my analysis persuasive. I don’t have an explanation for the bubble and they think they do. That’s an advantage to the other side. But until I’m convinced that it’s a general theory that can explain other money–bubble linkages in other periods of history, I’m not willing to sign on.
(For example, in the 1920s real interest rates were not low and NGDP growth was not high. And yet we still had a big stock “bubble” followed by a severe slump.)
Marcus Nunes also comments on George Selgin.