Tyler Cowen pointed me to a excellent interview with John Cochrane. As usual, I agree with Cochrane on most things. He makes excellent points about finance, regulation, health care, etc.
But not money. Here are a few examples:
I’ve been searching all my professional life for a theory of inflation that is both coherent and applies to the modern economy. That might sound like a surprising statement, especially from someone at Chicago, home of MV=PY. But although MV=PY is a coherent theory, it doesn’t make sense in our economy today. We no longer have to hold an inventory of some special asset — money — to make transactions. I use credit cards. We pretty much live in an electronic barter economy, exchanging interest-paying book entries, held in quantities that are trillions of dollars greater than needed to make transactions. The gold standard is a coherent theory too, but it doesn’t apply today either.
Several problems here. MV=PY is not a coherent theory, because it’s not a theory at all. It is a definition of velocity. V is the ratio of NGDP to money. That’s all it is, a definition. Now let’s cut Cochrane some slack and assume he meant something more like a monetary theory of prices. Thus I might argue that the Fed can target inflation at 2% by offsetting shifts in V and Y. I take it that Cochrane is opposed to that view, and has an alternative fiscal theory of the price level.
But even in that case the argument is very weak. Unless I’m seriously mistaken Cochrane does believe the gold market theory of the price level worked under the gold standard. But his objection to MV=PY (that some people use currency substitutes such as credit cards) was equally true under the gold standard. During the early 1900s, gold was actually used even less frequently than cash is used today. So if the gold market model worked during the gold standard period, despite gold substitutes, why don’t simple MV=PY (i.e. monetary) models work in an economy where currency is the medium of account?
Despite the existence of credit cards, currency demand remains very strong, and indeed is rising. By adjusting the supply of base money the Fed can control its value, at least when not at the zero bound. Even at the zero bound they can do so, but it’s a bit trickier. (I believe Cochrane supports my futures targeting approach.)
I think we’ve left the point that we can blame generic “demand” deficiencies, after all these years of stagnation. The idea that everything is fundamentally fine with the U.S. economy, except that negative 2 percent real interest rates on short-term Treasuries are choking the supply of credit, seems pretty farfetched to me. This is starting to look like “supply”: a permanent reduction in output and, more troubling, in our long-run growth rate.
In one sense he’s right. Growth is slow despite a falling unemployment rate. We are in the Great Stagnation. But surely the fall in unemployment from 10% to 7% suggests that there was some demand deficiency when it was 10%. I think even Cochrane would agree on that point, as he suggests we’ve “left the point” where demand deficiencies could be blamed. But have we? Isn’t it very likely that unemployment will continue falling even as inflation stays below 2%? That’s what most forecasters believe. If we had futures markets for unemployment I’m pretty sure they would indicate that this is also what the market believes. The demand shortfall is smaller than before, probably smaller than many Keynesians believe, but it’s still there.
Later in the interview Cochrane criticizes European policies that pay prime age able-bodied males not to work. And yet America adopted just such policies in 2008, and only ended them over the weekend. There are good arguments both ways on extended UI, but surely everyone would agree that the best policy is to eliminate the need for extended UI, by having unemployment return to its natural rate. And if you think we are there now, does that mean you disagree with the consensus view that unemployment will continue falling over the next year or two, even as inflation stays low?
So I don’t think the theory suggests QE can have a big effect. What about the evidence? Most of it comes from announcement effects. Even there, it’s pretty weak: a 15-or-so basis point change in interest rates in return for a pledge to buy trillions in Treasuries.
. . .
If the Fed announces more QE or delayed tapering of QE and bond prices rise on that announcement, is that because QE itself is moving the markets? Or is it because bond investors think, “Wow, the Fed is scared, so it will keep interest rates low for a lot longer than we expected”? Without a solid economic reason to believe QE on its own has much of an effect, the latter interpretation seems more likely.
This is Paul Krugman’s view as well. Two problems here. First, the response of nominal interest rates tells us nothing about how expansionary QE will be, as the liquidity and income/inflation effects go in opposite directions. More importantly, the Fed can only hold rates lower for longer by a more expansionary monetary policy down the road. So all Cochrane is actually saying is that QE might work, but only because a part of it is expected to be permanent. Krugman makes the same argument. To me that would be like saying:
“Yes, I suppose you could argue that pushing the gas pedal to the floor of the car makes it go forward, but only because the other foot is not simultaneously depressing the brake.”
Yup, that’s why cars move. And monetary stimulus works, if it works at all, only if it is not expected to be withdrawn in the near future. That’s even true when rates are positive.
Long-term growth is like a garden. You have to weed a garden; you don’t just pile on fertilizer — stimulus — when it’s full of weeds. So let’s count up the weeds.
I really like John Cochrane. For his sake I hope Krugman doesn’t see this comment.
PS. People asked for a link to my posts over at Econlog. I added it to the right column, under “Sites I Visit.” So far I have only one post. I expect to add a second one tomorrow. Because Econlog is shared with three other bloggers, I will post less frequently there than here.
Update: Mike Freimuth asked a good question:
I’m confused by this:
“the Fed can only hold rates lower for longer by a more expansionary monetary policy down the road”
Since low rates often indicate tight monetary policy, how is it that the Fed couldn’t keep them lower for longer with more contractionary monetary policy down the road. If they tried to shrink NGDP by 10%/year, wouldn’t interest rates be very low for a long time? I think I’m missing something important…
And I responded:
Mike, Yes, my mistake. I was reacting to Cochrane’s comment about Fed signaling. So if the Fed is actually signaling an intent to have a higher future price level than investors currently believe (which seems to be the implication of his comment) then it would also involve a higher long run money supply. But yes, low rates can also be achieved via a tight money policy, a la Japan. I interpreted Cochrane as saying that it would be a signal of both lower future short term rates and a central bank intention to push the price level higher, but I should have made that clear. I.e. the central bank would wait longer to exit the zero bound, because they would wait until NGDP and prices rose by more than previously expected.