John Cochrane has an interesting paper on the current economic crisis. I agree with some but not all of his analysis. Naturally I will focus on areas where I disagree. Here’s how he opens the paper:
I offer an interpretation of the macroeconomic events in the great recession of 2008-2009, and the subsequent outlook, focused on the fiscal stance of the U. S. government and its link to potential inflation. What happened? How did policies work? Are we headed for inflation or deflation? Will the Fed be able to follow an “exit strategy?” Will large government deficits lead to inflation? If so, what will that event look like?
I base the analysis on two equilibrium conditions, some form of which hold in almost every model of money and inflation: the valuation equation for nominal government debt and a money demand equation
We are both University of Chicago economists (although I suppose there is a slight difference between someone who teaches there and someone who merely took some classes.) So I am surprised that he would try to answer question like “are we headed for inflation” with a mathematical model. Why not just check out TIPS spreads? (Later he explains why, a point I’ll return to.)
Next Cochrane explains what he thinks caused the crisis:
Why did a financial crisis lead to such a big recession? We understand how a surge in money demand, if not accommodated by the Fed, can lead to a decline in output. I argue that we saw something similar — a “flight to quality,” a surge in the demand for all government debt and away from goods, services and private debt. In the fiscal context of (1), this event corresponds to a decrease in the discount rate for government debt.
Many of the Government’s policies can be understood as ways to accommodate this demand, which a conventional swap of money for government debt does not address. This story is in contrast to “lending channel” or “financial frictions” stories for the recession, essentially falls in aggregate supply.
As you know, I agree that this is mostly an AD problem, not a recession caused by financial system distress. I should add that as each day goes by the recession is less due to AD and more due to labor market rigidities. But AD is still a big factor. Were we differ is that he puts relatively more weight on fiscal policy:
Last, but perhaps most important: Will a fiscal inflation come with a boom or stagflation?
I argue that the fiscal valuation equation acts as the “anchor” for monetary policy, or the “expectation” that shifts the Phillips curve. A fiscal inflation is therefore likely to lead to the same stagflationary effects as any loss of “anchoring.”
Inflation can be explained from a fiscal perspective, but I see monetary policy as the dog and fiscal policy as the tail (except in places like Zimbabwe.) The Fed determines inflation and the fiscal authorities adjust their policies to accommodate the Fed’s preferred inflation path. Cochrane goes on to argue that future expected fiscal policy is the key:
Our most pressing question is, how might debt and deficits translate into inflation? Equation (3) gives an unusual answer and a warning: Expected futuredeficits St+j cause inflation today. Inflation need not wait for large deficits to materialize, for large debt to GDP ratios to occur, for monetization of debt or for explicit seigniorage. As soon as people figure out that there will be inflation in the future, they try to get rid of money and government debt now.
Yes, future expected inflation determines current inflation, a point I keep emphasizing here. But why the term ‘unusual answer?’ I thought this was now the standard (Woodfordian) view. Current inflation and AD are driven by future expected inflation and AD, and hence future expected fiscal and monetary policy. I suppose Cochrane is thinking of all those silly complaints that Obama wasn’t responsible for the economic deterioration in early 2009 because the stimulus money hadn’t been spent yet. Certainly the smarter sort of new Keynesians should have known better than to make that argument. (On the other hand some of the smarter Keynesians may occasionally let ideology cloud their judgment.)
Then Cochrane presents some evidence on the relationship between deficits and inflation:
One might well ask, “What surpluses?” as the U.S. has reported continual deficits for a long time. However, equation (3) refers to primary surpluses, i.e. net of interest expense. Figure 1 presents a simple estimate of the primary surplus, taken from the NIPA accounts, and expressed as a percentage of GDP. In fact, positive primary surpluses are not rare. From the end of the second world war until the early 1970s, the US typically ran primary surpluses, and paid off much of the WWII debt in that way. 1973 and especially 1975 were years of really bad primary deficits, on the tail of a downward trend, and suggestively coinciding with the outbreak of inflation. The “Reagan deficits” of the early 1980s don’t show up much, especially controlling for the natural business cycle correlation, because much of those deficits consisted of very high interest payments on a stock of outstanding debt. The return to surpluses in the late 80s and the strong surpluses of the 1990s are familiar, and suggestively correlated with the end of inflation. Our current situation resembles a cliff, motivating some of the concerns of this paper.
Regarding the last sentence; when the model says inflation and the markets say disinflation, a UC professor should go with the markets. In addition, I just don’t see the correlations he refers to (check graph on page 5.) The rate of inflation picked up in the 1960s, when we were still running primary surpluses. What caused that upsurge in inflation? I say monetary policy. And why did inflation slow sharply in the early 1980s? I don’t see his model giving any explanation. Maybe I missed something. I’m not saying there is no story you could tell. Maybe in 1982 people began anticipating the Clinton administration surpluses. But that doesn’t seem likely.
Then Cochrane applies his model to the current crisis:
The first issue is, why was there such a large fall in output? For once in macroeconomics we actually know exactly what the shock was — there was a “run” in the shadow banking system (See for example Gorton and Metrick, 2009b, or Duffie, 2010). But how did this shock propagate to such a large recession?
We have long understood that a sharp precautionary increase in money demand, if not met by money supply, would lead to a decline in aggregate demand. With price-stickiness or dispersed information, a decline in aggregate demand can express itself as a decline in real output rather than a decline in the price level. This is in essence Friedman and Schwartz’s explanation for the great depression. However, this story cannot credibly apply to the 2008-2009 recession. The Federal Reserve flooded the country with money (reserves). There is no evidence for a flight to money at the expense of government bonds.
Two p0ints. I agree that there was an increase in the demand for both money and T-bonds. But I disagree about the Fed flooding the economy with money (and later I’ll argue that Cochrane also disagrees with this assertion.) Interest bearing reserves are not “money,” at least not the non-interest-bearing asset that Friedman and Schwartz had in mind. Nevertheless, the non-interest-bearing base did rise by quite a bit in Japan–so I won’t belabor this point.
As the financial crisis took off in the third week of September 2008, the Federal reserve swiftly cut the Federal Funds target to a range between 0 and 25 bp, and signaled it would leave interest rates there for a long time (Figure 4). Inflation declined, never turned to deflation so real rates on these assets remained near zero.
Expected inflation did turn negative, and the real interest rate on 5 year TIPS rose from 0.5% to 4.25% between July and November 2008. But I agree that real rates don’t provide a reliable indicator, so let’s skip over this issue.
Excess reserves rose from $6b to $800b. While it’s hard to disprove anything in economics, it certainly seems an uphill battle to argue that the recession resulted from a failure by the Fed to accommodate shifts in money demand.
We will see about that.
The combination of near-zero government rates and reserves paying interest, means that the distinction between government bonds and money (reserves) was a third-order issue for financial institutions, especially compared to the very high interest rates, lack of collateralizability, and illiquidity of any instrument that carried a whiff of credit risk. If they wanted more of either, they wanted more of both.
. . .
The Fed has also started paying interest on reserves. Reserves that pay interest are government debt. By creating such reserves the Fed can rapidly expand the supply of shortterm, floating rate debt, without needing any cooperation from the Treasury or a rise in the Congressional debt limit. It also can execute massive open-market operations at the stroke of a pen. With a trillion dollars of excess reserves, changing the interest on reserves from 0 to the overnight rate is exactly the same thing as a trillion-dollar open-market operation.
. . .
The fact that reserves now pay interest dramatically changes our interpretation of the data. Reserves that pay market interest are debt, not money.
Exactly. These reserves are debt, not money. So in fact this case does not provide much evidence against the monetarist view. Cochrane continues:
This is not how the Fed thinks about its policy actions, at least as I interpret Fed statements. The first stage, trading private for government debt without increasing money, was, to the Fed, a way to support private credit markets without the inflationary effect that increasing M might have had. The Fed wanted to stimulate in a noninflationary way, an idea beyond my simple analysis.
Yes, how the Fed came to the conclusion that it was possible to boost AD without boosting inflation is also something far beyond my simple analysis. I am as puzzled as Cochrane.
Many critics objected that fiscal stimulus won’t stimulate in time, because the spending will come too late, after the recession is over. This reflects the standard analysis, enshrined in undergraduate textbooks since the 1970s, that fiscal policy, affects “demand” as it is spent. Equation (14) suggests the opposite conclusion. In order to get stimulus (inflation) now, future deficits (St+1 for large pi) are just as effective as current deficits, and possibly more so.
Again, this just floors me. Cochrane may be right, but I am stunned that the standard view of fiscal policy doesn’t take expectations into account. If there are any grad students out there, please ”say it ain’t so.”
But then Cochrane seems to assume modern monetarism is equally ignorant of expectations:
But in our framework, it’s hard to see how quantitative easing can have any effect. The Fed can increase reserves M and decrease B, but nobody cares if it does so. Agents are happy to trade perfect substitutes at will. Velocity V will simply absorbs any further changes. The argument must rest on the idea that V is fixed, but why should the relative demand for perfect substitutes be fixed? (With interest on reserves, the same logic applies even at nonzero interest rates, and one would expect the argument to hold as an approximation at small positive rates.)
Cochrane is comparing an outdated monetarist model where current monetary policy drives current and future AD, with a sophisticated fiscal model where current AD is driven by future expected deficits. But that’s not fair. We have known for a long time that it is future expected monetary policy that drives current AD. The fact that money and T-bills are now almost perfect substitutes does not in any way inhibit the Fed from targeting the price level (unless you assume that the liquidity trap will last forever.) Believe me, when T-bill rates get up to 2% or 3%, then non-interest-bearing reserves will not be considered perfect substitutes for T-bills, demand for excess reserves will fall almost to zero, and the Fed will have its usual ability to control the price level path that comes from its position of being the monopoly producer of the stuff we carry around in our wallets. I don’t plan on carrying T-bills to go shopping at Wal-Mart. The quickest way to get out of the liquidity trap is to target a much higher NGDP growth path than what is currently expected. That will dramatically lower the demand for base money. (And of course they can also pay negative interest on bank reserves.)
What about a “helicopter drop?” Wouldn’t this increase money M and inflate? A helicopter drop is at heart a fiscal operation. To implement a drop in the U. S., the Treasury would borrow money, issuing more debt. It would spend the money as a government transfer. Then the Federal Reserve would buy the debt, so that the money supply increased. A real drop of real cash from real helicopters would be recorded as a transfer payment, a fiscal operation. Conversely, even a helicopter drop would not be “stimulative” if everyone knew that the money would be soaked up the next day in higher taxes, or by the Fed, i.e. by future taxes.
Thus, Milton Friedman’s helicopters have nothing really to do with money. They are instead a brilliant device to dramatically communicate that this cash does not correspond to higher future fiscal surpluses; that this money will be left out in public hands as in a currency reform. To be effective, a monetary expansion at near zero rates must be accompanied by a non-Ricardian fiscal expansion as well. People must understand that the new debt or money does not just correspond to higher future surpluses.
I’m glad this came up, as I was about to do a post on the widely misunderstood “helicopter drop.” These money drops don’t do anything that an open market purchase can’t do, unless it changes fiscal policy expectations. An OMP at the zero bound will be highly inflationary if it is expected to be permanent, and not at all inflationary if it is expected to be temporary. And exactly the same is true of the helicopter drop. There is a strong effect if expected to be permanent, and almost no effect if expected to be temporary.
Cochrane does not explicitly forecast higher inflation. But his model suggests that it is likely to be a problem, as the graph on page 5 shows a fiscal situation that is far worse than what we saw in the 1970s. Here he first discusses three reasons why the Fed doesn’t fear inflation, and then hints that he sees worrisome signs of inflation:
The first asserts that “inflation expectations have been relatively stable” and points to a graph (figure 2) of actual inflation. The second (under “prices”) summarizes median survey data, excusing a jump in short-term expectations by energy prices and pointing to more stable long-term expectations. The third inferred expectations from Treasury vs. TIP yields, again arguing that “short-term” expectations might have risen but “long-term” expectations had not changed much. In evaluating the latter, we should remember that neither surveys nor long-term yields gave any warning of inflation in the 1970s nor disinflation in the 1980s. These are the only mention of expectations or documentation of the FOMC and Chairman’s assertions in the document. Occasionally, sophisticated Fed statements allude to the New-Keynesian idea that expectations are anchored by a belief that the Fed will respond quickly to inflation, though not why people should have such a belief. The volume of popular press coverage of deficits and inflation — clearly about expected future inflation — and even the ads for gold on cable TV suggest at least a more widespread concern about inflation than has been present for some time.
I think he is grasping for straws here. If you want a relatively direct take on inflation expectations you look at TIPS spreads and CPI futures markets. Real gold prices are distorted by massive Asian demand.
I share many of Cochrane’s criticisms of Keynesianism. I like his discussion of the need for a dynamic Laffer curve, and his contempt for “slack” models of inflation. And I agree that a lack of AD, not financial turmoil, explains the severe phase of the recession. But I disagree on a few key points:
- Because money and T-securities are not close substitutes during normal times, the Fed can control the price level in the long run.
- This means that open market purchases that are permanent do affect the price level. They affect future expected prices and hence current prices.
- The Fed can create expectations that open market purchases are permanent if it sets an explicit target for the price level or NGDP.
- Monetary policy is the dog in America, fiscal policy is the tail. This may reverse in the future, but . . .
- Markets aren’t afraid of fiscal inflation. They saw what happened in Japan. TIPS markets are much better inflation indicators than gold prices.
BTW, this post was inspired by another post called “Is John Cochrane Sumnerian Now?“ I vaguely recall that he had a more finance-oriented view of the recession when we debated last year. But that is probably wishful thinking on my part. My memory may be faulty, and even if it isn’t I doubt I had any effect on his views.