The new DeLong and Summers paper looks impressive enough. I don’t buy their argument for fiscal stimulus, but I am willing to listen respectfully to the hysteresis argument (which seems central to their self-financing stimulus argument.) Obviously Brad DeLong would not listen respectfully to a conservative arguing tax cuts are self-financing due to long run output effects. That would be “voodoo economics.” So the following points should be viewed in this context; I’m not about to trash their paper, which seems excellent. I’m about to trash the political culture that leads to this sort of paper. Here’s D&S:
This paper focuses on policy choices in a deeply depressed demand constrained economy in which present output and spending are well below their potential level. We presume for the moment that monetary policy is constrained by the zero lower bound, and that the central bank is unable or unwilling to provide additional stimulus through quantitative easing or other means—an assumption we discuss further in Section V.
Now stop right there. I want to throw a brick at the screen when I read “unable or unwilling.” Here’s how things are in the real world. No fiat money central bank ever tried to inflate and failed. It will never happen. It’s really, really easy to debase a fiat currency. The Fed can engage in 5% NGDP targeting, level targeting, if it so chooses. You create an NGDP futures market and buy assets until NGDP futures are right on target. How many assets would you have to buy? Under NGDP level targeting the monetary base would be less than 1/2 half its current level. The Fed doesn’t want to do this, or any of the million other methods that would work. So their entire paper boils down to what is the fiscal multiplier if the Fed chooses to “do the wrong thing.”
So let’s start over. The Fed is unwilling to provide enough monetary stimulus. OK, now what is the point of this paper? Is this to train our future econ PhD students? Are we trying to teach them the optimal policy regime? Obviously not. The optimal regime relies on monetary policy to steer the nominal economy, and fiscal policy to fix other problems. So we are going to defend the model how? A blueprint for failed states? For banana republics? Fair enough, but ask yourself the following question: In a failed state, which is more incompetent branch of government; the central bank or the legislature?
Yes, the Fed is bad. But Congress is downright ugly. Deep down most economists are technocrats. They see the central bank as being the best and the brightest, the guys who are above politics, who will “do the right thing.” And how do economists view our Congress? The terms ‘stupid’ and ‘incompetent’ don’t even come close to describing the disdain. So are we supposed to change our textbooks in such a way that the fiscal multiplier is no longer zero under an inflation targeting regime (as the new Keynesians had taught us for several decades?) And on what basis? Because the Fed might be so incompetent that we need Congress to rescue the economy? In what world does that policy regime actually work? If you have a culture that has its act together, such as Sweden or Australia, the central bank will do the right thing. If not, then all hope is lost.
[Check the previous post if you have any doubts about which institution is more incompetent.]
And it’s even worse than that. DeLong and Summers seem to make the common mistake of assuming that as long as the monetary authority is not doing its job, we can assume the fiscal multiplier will be positive, as fiscal stimulus won’t be offset by monetary tightening. But that’s not right. It’s not enough for the monetary authority to be incompetent; they must be incompetent in a very special way. Consider the follow two examples.
1. After a financial crisis the central bank is completely passive, allowing severe deflation to set in.
2. After a financial crisis the central bank is slightly thrown off course, and allows inflation to fall 1% below target, but no lower.
In case 1, the fiscal multiplier might well be every bit as high as the textbooks suggest. But in case 2 the multiplier might well be zero. If the Fed does just enough QE to keep inflation in the 1% to 2% range, but no more, then any extra fiscal stimulus will be offset by less QE. I’m not saying that exactly describes the current situation, but surely it better describes recent Fed policy that case 1.
In fairness, in section V DeLong and Summers try to make a more sophisticated argument against the reliance on monetary stimulus:
Perhaps, though, as Mankiw and Weinzerl (2011) suggest, arguments for temporary fiscal expansion are even better arguments for expansionary monetary policy. Here too we are skeptical. While a much richer model would be necessary to fully address the issue, it seems to us that if fiscal policy is self financing it will be de-sirable to use as an instrument once it is recognized that (i) with uncertainty about multipliers diversification among policy instruments is appropriate as suggested by Brainard (1967), (ii) expansionary monetary policies carry with them costs not represented in standard models (including distortions in the composition of investment, impacts on the health of the financial sector, and impacts on the distribution of income), and (iii) the historically-clear tendency of low interest rate environments to give rise to asset market bubbles.
This is a very widely held attitude, but gets things completely backwards. D&S implicitly make the very common mistake of assuming current Fed policy is expansionary, but not expansionary enough; and then assuming that the monetary policy needed to generate adequate NGDP growth must be much more of the same, and must produce even bigger distortions than this policy. (Yes, that’s reading between the lines, but I think most readers would share my reading.)
Exactly the opposite is true. As Milton Friedman pointed out, ultra-low rates are a sign that money has been very tight (driving NGDP far below trend.) I’d add that a bloated monetary base reflects the same forces. The base is 23% of GDP in Japan and 18% of GDP in the US because our NGDP growth has been really low in recent years. It’s only 4% of GDP in Australia because they have a much higher trend NGDP growth rate (roughly 7%) and hence much higher nominal interest rates, and hence a much higher opportunity cost of holding ERs. If you want an economy free of the “distortions” caused by low rates and the Fed buying up lots of assets–then set a more expansionary monetary policy target.
As far as asset bubbles are concerned, they don’t hurt at all when NGDP growth is maintained (1987), hurt slightly when NGDP growth falls slightly (2000), but they seem to hurt a lot when it isn’t maintained (1929, 2008.) Actually, the pain of the latter two cases was caused by the falling NGDP, not the bursting bubble, but to the average person it looks like the bubble did it.
If we are going to teach our students how to “do the right thing,” we need to start by eliminating “depression economics” from the curriculum. Start with the fact that what Krugman calls “depression economics” should actually be called “expected depression economics.” That’s because fiscal stimulus acts with a lag that is just as long as monetary stimulus. (Here and here I argued monetary lags are surprisingly short.) So it doesn’t matter where the economy is right now, but rather where it will be in 12 months. Monetary policy should always be set in such a way as to produce on-target expected NGDP growth. That’s the Lars Svensson principle. If you do that, there’s no room for fiscal stimulus, even if the economy is currently depressed. With a central bank that targets expected NGDP growth along a 5% growth path, you are in a classical world. Spending has opportunity costs. Unemployment compensation discourages work. Saving boosts investment. Protectionism is destructive. And so on. That’s the policy we should be teaching our grad students. The optimal monetary policy. Not a policy mix that only has a prayer of making sense in countries where the central bank is even more stupid and corrupt than the Congress. As far as I can tell, those countries don’t exist.
PS. Just to reiterate, this post is a not a comment on the core of D&S. I’m not qualified to judge their model, but it looks fine to me. I just don’t buy the assumption that motivates the entire exercise.
PPS. Marcus Nunes also has a post on the DeLong & Summer paper.