On the Negative Information Revealed by Marvin Goodfriend’s “I Don’t Teach IS-LM”
The smart and snarky Sam Bell wants to taunt me into rising to his bait by twittering https://twitter.com/sam_a_bell/status/872116967070732288 a quote from likely Fed nominee Marvin Goodfriend: “I don’t teach IS-LM”. He succeeds. Here is the quote:
TOM KEENE: But, Marvin, with, you know, basic IS-LM and theory and all that stuff you teach in Economics 101, aren’t we going to see a dampening of GDP if we see a restrictive Fed?
MARVIN GOODFRIEND: By the way, I don’t teach IS-LM. But what I would say is this…
And here is the tape:
March 23, 2012: https://www.youtube.com/watch?v=emvSYwUnWyI&ab_channel=Bloomberg
Let me start by analyzing “I don’t teach IS-LM”. And let me preface this by saying that Marvin Goodfried is a very sharp and honest economist. But I believe that whenever anybody says “I don’t teach IS-LM” they are one of:
- Making completely implausible and wrong claims about how the economy works.
- Being lazy and/or stupid.
- Declaring a tribal affiliation to a particular Carnegie-Mellon tradition of macroeconomic analysis that the late Rudi Dornbusch described to me and others as “Jim Tobin with original errors”, and that I think has shed a lot more heat than light on real issues.
Let us start with (1), and let us start with Irving Fisher’s monetarism: the quantity of money demanded in the economy is given by the equation:
Md = PY/V
where M is the quantity of money demanded, P is the price level, Y is the level of production, and V is the velocity of money—the value of transactions that having $1 in the bank or in cash as money can support, in the sense of manufacturing the needed trust so that the transactions will go through.
If you believe that that velocity of money is fixed by the institutions of the banking system and the technology supporting transactions, then you do not have to teach IS-LM. You have reached a full stop, and have the monetarist conclusion that the total nominal spending in the economy—prices times quantities produced—is equal to a constant times the economy’s money stock, with the constant of proportionality chaining slowly over time as the institutions of the banking system and the technology supporting transactions slowly changes.
That is meaning (1) of “I don’t teach IS-LM”: I do not need to teach it because it is not important in determining how much spending there is the economy. That is implausible and wrong. Here is the graph of velocity since 1960—the thing that is supposed to be on a smooth and steady time trend if “I don’t teach IS-LM” is a sensible thing to say:
Even before the 1990s any model assuming an unproblematic relationship between the money stock and total spending was badly awry, although not as badly awry as it has been since.
Now let’s move on to (2)—lazy and/or stupid. The graph above tells you that if you want to forecast—or even retrospectively explain—the relationship between the money stock and the level of spending, you need a model of what the determinants of the fluctuations of velocity we see are. If we draw a graph with the level of spending on the horizontal axis and some sufficient statistics for the determinants of velocity on the vertical axis, the path traced out by our equation:
Md = PY/V
is conventionally called “the LM curve”. But you then need to know where on the LM curve the economy will be—you need another curve. And that other curve is conventionally called “the IS curve”.
To claim that you do not teach IS-LM is to implicitly claim that you do not need to figure out where on the LM curve the economy will be. That is something it is only possible to say if you are being lazy, or stupid.
The third meaning of “I don’t teach IS-LM” is that it is a CMU-school tribal indentification marker, and has no purpose beyond that—no intellectual purpose.
So, yes, the fact that Marvin Goodfriend would go on Tom Keene’s surveillance and say “I don’t teach IS-LM” makes me think a good deal less of him. I do, however, interpret that claim as a declaration of tribal allegiance to CMU-school macro. I do not interpret it as a claim that you don’t need a model of the determinants of fluctuations in velocity. I do not interpret it as a claim that there are no fluctuations in velocity large enough to worry about.
What worries me more, however, is what comes next:
GOODFRIEND: There is no way that this recovery can proceed with any degree of confidence unless the Fed makes sure that inflation does not move up. So I think the risks are exactly reversed from the way the Fed chairman discusses this. He has to make the public understand that any whiff of doubt about the Fed’s ability and willingness to stabilize inflation is going to put a crimp into the public’s willingness to take positions and commitments over the next two or three years that would produce genuine growth. And so I would just take it, and turn it on its head, and not put the question as you did to me, but reverse it.
The risks of allowing any latitude in inflation expectations to build dup, or any doubt about the Fed’s willingness to do what it takes to keep inflation down, is to me the most likely risk in preventing this recovery from getting any traction…
Do notice that Marvin Goodfriend is, here, thinking in terms of an IS-LM model. When he says “any whiff of doubt about the Fed’s ability and willingness to stabilize inflation is going to put a crimp into the public’s willingness to take positions and commitments… is to me the most likely risk in preventing this recovery from getting any traction…”, he is saying: “any whiff of doubt about the Fed’s ability and willingness to keep inflation low will cause a large leftward shift in the IS curve that will prevent this recovery from getting any traction…” He does not do more than gesture at an expectational mechanism for this leftward shift in the IS curve that he wants the Federal Reserve to take action to head off. But it is what he fears.
And, of course, Goodfriend was wrong: a continuation of Bernanke’s extraordinary easing policies was not going raise “any whiff of doubt about the Fed’s ability and willingness to stabilize inflation”.
Here we have a market-based measure of inflation expectations—the ten-year breakeven inflation rate since 2010: that inflation rate over the forthcoming ten years that would, at each date, have made investments in conventional Ten-Year U.S. Treasury bonds and investments in Ten-Year Inflation-Protected Securities (TIPS) equally profitable. The vertical blue line marks March 23, 2012: the date of Marvin Goodfriend’s interview. The point that Marvin was hammering home again and again on March 23, 2012 was that the Federal Reserve needed to rapidly start shrinking its balance sheet and raising interest rates lest inflation expectations break out to the upside.
The Federal Reserve ignored Marvin Goodfriend.
And Marvin Goodfriend was wrong. The shift to a tighter, more restrictive policy he demanded then was not necessary to prevent an upside breakout of inflation expectations.
In fact, the Federal Reserve’s persistent problem since has been that expectations of—and actual outcomes for—inflation have been well below rather than above the Federal Reserve’s targets.
I would very much like to hear Marvin Goodfriend explain why he misjudged the situation in the spring of 2012, and how he has updated his view of the economy and of optimal monetary policy since.