Today (Thursday, November 18) was my first day at the National Tax Association’s 103rd Annual Conference on Taxation, being held in Chicago. (Being posted a day later, however, due to web access issues at the execrably out-of-the-way Hyatt Regency McCormick Place in, but not really of, Chicago.)
Best part of the day was the lunch talk by Narayana Kocherlakota, the President of the Federal Reserve Bank of Minneapolis, concerning the current state of the play in monetary policy. The talk had more intellectual content than one typically finds in luncheon talks, even at the NTA (a neighbor at lunch made the same observation with opposite spin, calling it “very abstract”).
Just to compare it to a couple of past “highlights” at NTA lunches, I well remember a session at which I got into a heated colloquy with the speaker, Ed Lazear (while he was working for the Bush Administration, but before he joined the Tax Reform Panel), because I had been irritated by his passionate encomium to how long-term fiscally responsible his Administration was. And a few years before that, Glenn Hubbard (also while working for the Bush Administration) explained that the Bush tax cuts cost zero, not based on supply-side arguments, but on the view that every single dollar of tax cuts necessarily reduced government outlays by a dollar. But that may not have been at lunch.
Kocherlakota, by contrast, while discussing issues related to his current job, has a position in which he is not comparably required to be a political spokesman. He also clearly has an academic temperament, as one would expect from his work (I’m most familiar with that in “new dynamic public finance,” which I discuss here).
The talk left me feeling better about Kocherlakota’s recognition that it’s desperately important for the Fed, if at all possible, to address sustained unemployment. It left me feeling worse, however about the chances that the Fed’s current policy will actually help.
The basic theme was as follows: What should the Fed do when it wants to stimulate the economy, given the inadequate and slowing recovery, and can’t lower interest rates because they’re already effectively zero? He discussed two main things: the current policy of quantitative easing (QE), a.k.a. asset purchasing; and the potential interchangeability between fiscal policy and monetary policy. By this he meant not just the truism that either one can be stimulative, but rather that one can in principle exactly replicate or simulate the other. This second part of the talk was the one that got him plaudits from me and demerits from some others for being abstract – it was more of a thought experiment than anything with likely short-term policy ramifications.
OK, I’ve been reading a bit here and there about QE, but I must admit, as it’s outside of my real field of knowledge and I have plenty of other things to keep me busy, my understanding of it has been a bit vague. The talk was nicely helpful – unfortunately, in convincing me that it is highly unlikely to be very effective, even leaving aside the question of whether the scale would be too small even if it were in principle a sharp tool.
The ongoing QE ruckus concerns the Fed’s buying $600 billion of U.S. government bonds in simple market transactions. Republicans have been denouncing this – I believe, not in good faith given how 2012 avowedly trumps all other considerations in their thinking. The likes of Krugman say it’s much too small. But a starting point would be to ask how it is supposed to work in stimulating the economy.
One can understand how printing money potentially works. People feel richer in the short run and demand goes up (eliminating slack before inflation becomes a problem). And one can understand directly lowering interest rates, so consumers and businesses are more inclined to borrow for current spending. But why would Fed purchases of Treasury bonds be stimulative?
One path to QE’s being stimulative, which Kocherlakota rightly discounted, is that in practice it involves the Fed’s effectively giving the banks that sell it the bonds $600 billion more capacity to lend. In other words, increase the money supply because they can now make more loans. Only, the banks apparently have at present $1 trillion in unused capacity to lend, even with near-zero interest rates. So raising the unused capacity to $1.6 trillion is not going to accomplish much.
He instead proposed two alternative mechanisms for QE to work, in each case by replicating the policy act of lowering the interest rate despite its being zero. First, he said, the bond purchase is a credible statement that the Fed actually expects and believes and plans for interest rates to be very low for a very long time. So it lowers expected future interest rates that businesses are taking into account today.
This appeared to me to be a “skin in the game” type of argument. Suppose Warren Buffett purchased a vast quantity of U.S. government bonds. This might credibly signal that he believes U.S. interest rates will be low for a long time. If he were somehow the person who gets to decide what interest rates will be, it would be evidence that he expects to keep them low, plus he’d now have an extra reason for doing so (since the bonds he had purchased would lose value if interest rates went up).
But how does this to apply to the Fed, a governmental and hence nonprofit agency? To what extent are the Fed governors’ true incentives (which presumably are largely reputational) either illuminated or affected by its buying bonds? I think all QE really does in this dimension is show that they’re trying to shout. It’s a way of trying to pound the table and say: We really plan to keep interest rates low for a long time. But I’m not sure how much the act adds to the statement – which itself faces a pre-commitment problem, since what the Fed wants to do in the future might not be the same as what it wants people today to think it will do then.
In sum, surely the statement helps, but it’s unclear how much, and also unclear how much QE strengthens it. When a Fed governor admits they’re basically jawboning, it’s time to get nervous about their capacity to do anything effective even if they place a very high priority on addressing unemployment.
Route 2 that Kocherlakota identified by which QE can replicate lowering the interest rate is that, once the Fed has acted, the public (or rather those in the capital markets – including, of course, foreign governments and the like) now holds $600 billion less in U.S. government bonds than it did before. So the “public” is now collectively less exposed to U.S. government bond risk than otherwise, and requires less of a risk premium to hold bonds. This means that interest rates can now be lower (and expected to be lower) than in the absence of QE, continuing into the future. This rationale, unfortunately, strikes me as likely to be a bit trivial as an empirical matter.
In the more analytically, though less practically, interesting part of his talk, Kocherlakota discussed how fiscal policy can in principle replicate monetary policy. He said: Suppose the Fed lowers the interest rate by 100 basis points. Now the amount of consumption you can buy in a year by saving has dropped by 1%. Thus, suppose there is no inflation but the interest rate is zero. So, by saving for a year you are choosing between 1 widget now and 1 widget in a year. If the Fed could make the interest rate negative 1 percent, it could force you to choose between 1.01 widgets now and 1 widget in a year, inducing you to shift consumption forward to the present. But it can’t achieve this due to the zero lower bound. (After all, if you live in a safe neighborhood your mattress offers a 0% rather than negative return.)
Aha, he said. But suppose the legislature enacts a 1% consumption tax, to take effect in a year. And suppose it simultaneously enacts a 1% reduction in the wage tax, also to take effect in a year. Suppose further (though he was not explicit about this) that we assume, quite reasonably, that people in general can inter-temporally shift consumption but not labor supply. He also added in a temporary investment tax credit for purchases this year, but let’s ignore that to keep the example simple.
The exercise is roughly budget-neutral, and also keeps the labor versus leisure tradeoff the same as previously starting next year (since the price increase from the consumption tax is offset by the after-tax wage increase from the wage tax cut). But it means that you are choosing between 1.01 widgets now, before the consumption tax takes effect, and 1 widget next year (since with the 1% consumption tax increase that’s all you’ll get for the price of 1.01 widgets today). So it effectively replicates monetary policy while generally (if not quite perfectly) keeping other fiscal policy variables, such as the budget situation and incentive effects from next year forward, the same.
I thought it was a neat little analytical exercise. The fact that, in the real world of crude and imperfect instruments of fiscal policy, it would never hold quite exactly (as some questioners emphasized) is important as well, but doesn’t defeat the analytical point.
It’s also encouraging that at least one of the Fed governors is evidently so interested in wrestling intellectually with the problem of what the government can do when an interest rate cut is the preferred policy but unavailable. Discouraging as well, however, in that it doesn’t seem he’s come up with anything that we can anticipate actually being done and having a significant effect. Plus, it was almost as if he was saying: It's the legislature's fault, not mine. Even if they wanted to stick to replicating monetary policy, rather than running a more expansionary fiscal policy, they have the power that we at the Fed lack to escape the zero bound on interest rate cuts.
In sum, while this is hardly a new point, if QE is the best they have and all they are likely to do – and even that is attracting hysterical pushback from people who may not want the economy to get better before 2012 – we’d better fasten our seatbelts, or perhaps tighten our waist belts, because there are more bad times ahead.