First of all, I believe you should think of the Fed as simply part of the federal government when it comes to the financial side of its interventions.
I would distinguish between conventional monetary policy which sets the interest rate and this kind of financial intervention of buying what appear to be undervalued private securities. Issuing what appear to be overvalued public securities and trading them for undervalued private securities, at least under some conditions and some models, is the right thing to do. In my mind, it doesn’t make a big difference whether it’s done by the Federal Reserve, the Treasury or some other federal agency.


・・・there are two branches to the exit strategy: There’s paying interest on reserves, and there’s reducing reserves back to more normal levels. They’re both completely safe, so it’s a nonissue. The Fed itself is just not a danger. It is run by people who know exactly what to do. And we have 100 percent confidence they will do it. It’s not something I worry about.


・・・even though the Fed has driven the interest rate that it controls to zero, it hasn’t had that much effect on reducing borrowing costs to individuals and businesses. The result is it hasn’t transmitted the stimulus to where stimulus is needed, namely, private spending.
I mainly look at, as kind of a thought experiment, how much of a decline in activity occurs when that kind of a friction develops. When private borrowing rates rise and public borrowing rates fall, the difference between them is the amount of friction. I show that that’s a potent source of trouble. I haven’t tried to align it with history prior to the current crisis. That’s an interesting question, but data on historical events aren’t always so easy, so that lies ahead.
・・・Most of the undervalued assets that the Fed has bought have been mortgage related. It’s been kind of an obsession with trying to solve these problems as they arise in home building, but home building is only part of the story. The collapse in other types of investment spending has been equally large. There would be a case for expanding that type of policy to other seemingly undervalued instruments.

That would presumably result in the same pattern you’ve seen in mortgages. That policy has been successful—differentially successful in depressing mortgage rates as opposed to bond rates or other areas.


It’s a danger whenever you have guaranteed financial institutions that have gotten into a very low capital situation. They’ve suffered asset value declines, they’ve become extremely leveraged and they have this very asymmetric payoff to the owner: If they go under, it’s the government’s problem; if they recover, it’s the owner’s benefit. That asymmetry, which is the so-called moral hazard problem, is just a huge issue.
And yet, while we have a lot of institutions in that setting today, we don’t see many of them doing things that Akerlof and Romer described, such as paying themselves very large dividends. It’s been difficult to get them to cut the dividends, but they have not paid out very large dividends or concealed dividends.
So it looks like we’ve been somewhat successful in preventing the worst kind of stealing, but the asymmetry is still potentially a big issue. There are way too many bank failures that should not have occurred and especially should not have cost the taxpayers as much as they did.


A lot of people look to the example of Citibank. Citibank’s long-term debt has been selling at a considerable discount, which is a sign that the market knows that there’s an issue. So instead of doing what we have done, which is give guarantees of short-term debt with government investments, the alternative that the Squam Lake people are thinking of, and I’ve been thinking of too, is to somehow convert Citibank’s long-term debt into equity, which is the same thing that the market is in effect doing. That would eliminate the danger then that the bank couldn’t meet its obligations, in a way that is less burdensome to the taxpayer.


First of all, you have to take it apart, as I do in that paper, and ask how much of it went directly into government purchases, which is fairly small, or would stimulate state and local purchases, which was also fairly small.
A lot of it was providing income supplements, and there you get into the question of whether the people receiving the supplements increased their spending or not. That’s a whole other issue; I’m not commenting on that issue. That’s a very difficult question to answer.


To go on to the other part of your question, had there been an increase in government purchases that was successfully achieved, how much would that have increased GDP? The answer I got was around a factor of 1.7, which is at the high end of the range of what most economists were talking about.
I only reached that by thinking very carefully and reading a lot of recent commentary on this question of the implications of having a zero fed funds rate. That turns out to be very important. Others have found that to be true.
So I think that the people who looked at the evidence of what the multiplier is in normal times and said it’s maybe 0.8 or 1.0 (which I would agree with) kind of missed the point. There was a lot of, I think, inappropriate criticism.
Valerie Ramey, in contrast, has focused not on the immediate policy question but raised the scientific question about the long-run multiplier. Her numbers are ones that I respect and agree with. They’re more in the 0.9 range.
But on the issue of multipliers during periods of zero interest rates, because we didn’t have any changes in government purchases during this one time when we’ve reached the zero interest point, we don’t have any good empirical evidence. What we need is a time when interest rates are zero and there’s a big increase in government purchases. That just hasn’t happened.
So we have no way to know through pure practice; we have to use models. The models are very clear that it makes a big difference when we’re at the zero interest rate limit. The normal configuration is that you get this fiscal expansion—the government buys more, but that triggers sort of an automatic response from monetary policy to lean against it. If you shut that down by having interest rates stay at zero, you’ll get a bigger effect. That’s what this literature says and it’s quite a big difference.
Valerie Rameyは、対照的に、現下の政策の問題ではなく、長期的な乗数という科学的な観点からの問題を提起しました。彼女の出した数字については尊重しますし、同意もします。それは0.9付近の値です。


*2:cf. 小生の2/14エントリ

*3:cf. サムナーの異論