Stein jumps to conclusions

Tyler Cowen is very impressed with a new paper by Jeremy Stein, of the Board of Governors.  I see why, as I was also very intrigued by this finding:

In our paper, Sam and I begin by documenting the following fact about the working of conventional monetary policy: Changes in the stance of policy have surprisingly strong effects on very distant forward real interest rates. Concretely, over a sample period from 1999 to 2012, a 100 basis point increase in the 2-year nominal yield on FOMC announcement day–which we take as a proxy for a change in the expected path of the federal funds rate over the following several quarters–is associated with a 42 basis point increase in the 10-year forward overnight real rate, extracted from the yield curve for Treasury inflation-protected securities (TIPS).

On the one hand, this finding is at odds with standard New Keynesian macro models, in which the central bank’s ability to influence real variables stems from goods prices that are sticky in nominal terms. In such models, a change in monetary policy should have no effect on forward real rates at a horizon longer than that over which all prices can adjust, and it seems implausible that this horizon could be on the order of 10 years. On the other hand, the result suggests that monetary policy may have more kick than is implied by the standard model, precisely because long-term real rates are the ones that are most likely to matter for a variety of investment decisions.

And I agree that Stein has some very interesting and plausible explanations for this phenomenon.  But the final sentence in the quotation is unwarranted; it doesn’t tell us much of anything about the potency of monetary policy.  A perfect example of why it doesn’t occurred in December 2007, when the Fed adopted a more contractionary policy than markets expected.  Both the two-year and the 10 year Treasury bond yield fell on the announcement. Stocks also declined sharply. I don’t have data on the real bond yields but I presume those fell as well.

You might want to reread what I just wrote. Bond yields fell on a more contractionary than expected policy. In contrast in January 2001, and again in September 2007, long term bond yields rose on more expansionary than expected policy. Bond yields often move in a “perverse” direction. Obviously, the more contractionary than expected policy announcement of December 2007 was not made even more contractionary by the decline in long-term interest rates. Stein made the mistake of jumping to the conclusion that two-year bond yields move in a predictable fashion after monetary announcements. He assumed that an easier than expected monetary policy stance would make two-year bond yields fall. That often does occur, but not always.

You might wonder if the three unusual cases that I just cited are really all that important. It turns out they are. In those three cases the movement in stock prices was far more dramatic than usual. Most Fed announcements are relatively predictable and the stock market doesn’t respond very strongly. But in those three cases the implied swing in stock prices (after the announcement) was in the 5 to 10% range over a mere quarter point differential in the Fed funds target. That’s huge. So movements in long-term bond yields actually undermined monetary policy during exactly those times when it was most powerful (as indicated by the stock market.)

What’s the take-away from all this? It’s not that monetary policy is unimportant, just the opposite. Monetary policy is much more important than Stein realizes. But it’s not important because of the effect it has on interest rates, they are not an important part of the monetary transmission mechanism. If economists focused more on stock price movements and less on interest rate changes they would notice that monetary policy is extremely important even at the zero bound. Monetary policy is very important because it affects the expected path of nominal GDP.  An expansionary monetary policy might raise or lower long-term bond yields, but it will always increase expected NGDP growth.

You might wonder how monetary policy can be so important during periods where it doesn’t even move interest rates in the “correct” direction. The answer is that important monetary policy decisions are decisions that impact the expected future path of the monetary base relative to the demand for base money, and hence the future path of nominal GDP.

Some people tell me “Sumner, come on, the markets don’t care about the monetary base.”  That’s right, I don’t care either.  But they care a lot about M*V.  So do I.  And the Fed controls M*V by controlling M.

And as my previous post showed, M is the base.



12 Responses to “Stein jumps to conclusions”

  1. Gravatar of foosion foosion
    1. October 2013 at 04:30

    >>the Fed controls M*V by controlling M>>

    This appears to me to be the major point of contention between some of the monetary policy camps. One side says the Fed can’t control GDP because increases in M are offset by declines in V, while the other side says the Fed can control M*V if it tries hard enough.

  2. Gravatar of ssumner ssumner
    1. October 2013 at 04:43

    foosion. I really can’t see how there is a debate any longer. The idea that a fiat money central bank would be unable to debase its currency Zimbabwe style is patently absurd. Especially after all the contrary evidence that we have seen over the past few years. Even Krugman is now praising the BOJ for their stimulative monetary policy. As far as I’m concerned the debate is over. The real debate is over what sort of inflation/NGDP target is optimal.

  3. Gravatar of honeyoak honeyoak
    1. October 2013 at 04:58

    I have serious Micro-Economic disagreements with this paper. It seems to claim (or at least strongly hint) that these agents are not profit maximizing conditional on a given level of risk. There is no serious framework of why they would increase the riskiness of their portfolios and how this works out in aggregate.

  4. Gravatar of foosion foosion
    1. October 2013 at 05:07

    Scott, many people predicted Zimbabwe if the Fed expanded M to the extent it already has. I read Krugman as saying can’t hurt, may do some good.

  5. Gravatar of jknarr jknarr
    1. October 2013 at 07:26

    Im not sure that I’ve ever understood the fixation on demand deposits and M1 as being more important than the base. Demand deposits are a derivative of base money.

    I can swap bills for notes without any funds changing hands, but this does not make either into currency. I can swap currency for demand deposits, but demand deposits remain only a promise to deliver currency… on demand.

    I can swap around these DD promises for goods like an old “real bill”, but this does not make it any less of a promise to deliver base money.

    And I am certain that the fed would find it curious that the base does not matter – the base is how they control short rates, with a nice statistical relationship here:

    Take the data, and you can reverse it into a how-much-base-does-the-bill-market-demand for a given level of NGDP. They did screw up in 2007 by not providing sufficient base earlier.

    Also, I suspect that the recent increase in demand deposits assets is the flip side to Treasury liability growth.

  6. Gravatar of Roger Sparks Roger Sparks
    1. October 2013 at 08:24

    Current economic theory does not seem to take into account my observation that ALL assets have a rate of depreciation. My observation is that the expected rate of money depreciation is a primary driver in economic decisions of all kinds.

    Surprisingly, interest rates do not reflect the expected depreciation. In my opinion, this is because interest rates are controlled by the Federal Reserve. Lacking control, the only choice for investors is to choose the least evil of investments, expecting that nearly all investments will loose money in the long run due to inflation.

    Following my line of thinking, investors today find holding treasuries the least of investment evils. The most risky of all would be to invest in new plant, employing people to build products that can only be purchased so long as the Federal Reserve continues the present money supply unbalances.

  7. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    1. October 2013 at 08:48

    James Hamilton’s audience could use a visit, Scott;

  8. Gravatar of dlr dlr
    1. October 2013 at 09:14

    Stein made the mistake of jumping to the conclusion that two-year bond yields move in a predictable fashion after monetary announcements.

    I think you’re right that this is the fulcrum issue. These newer event papers are slightly improved. They don’t just focus on an actual FF surprise to determine what is monetary policy news, they try to look at the surprise in the path of future rates. But this is a bridge not far enough. What matters is a the conditional path of future rates, the entire reaction function. Unfortunately you can’t use the actual path change to observe this, because that only reflects a particular equilibrium path between the policy instrument and instrument-influenced economic conditions. The news is in the entire matrix, but the matrix is harder to see.

  9. Gravatar of Scott Sumner Scott Sumner
    1. October 2013 at 09:59

    foosion. The response of inflation to the recent increases in the base is NOT, and I repeat NOT, a test of whether the Fed is able to control NGDP. The way to control NGDP is to make changes in the expected future path of the base, relative to base demand. The Fed has not even ATTEMPTED this.

    MMs predicted that QE would produce very little inflation.

    Roger, If the Fed were holding interest rates below equilibrium (it isn’t) that would be GOOD for investors. They could borrow cheaply and invest in stocks.

    dlr, For pre-2008 data he could look at the actual fed funds target relative to the predicted fed funds target (from futures markets). Post 2008 I’d look at stocks.

  10. Gravatar of W. Peden W. Peden
    1. October 2013 at 10:07

    Stock markets crash disastrously on Federal government shutdown-

    – proving once and for all that it’s all about fiscal policy in a liquidity trap/ZLB scenario.

  11. Gravatar of Scott Sumner Scott Sumner
    1. October 2013 at 10:10

    Patrick, He’s halfway there, but forgets that changes in the expected future path of interest rates mean changes in the expected future path of the money supply. But that’s always true.

  12. Gravatar of flow5 flow5
    2. October 2013 at 11:14

    The base is required reserves (delimited by transaction deposit classifications 30 days prior), – but the base is determined by the individual commercial bankers as the Fed accomodates all requests for free gratis reserves at its policy rate.

    The base is morphing into liquidity reserves (prudential clearing balances like the unregulated Euro-dollar banking system).

    The Financial Services Regulatory Relief Act of 2006 provides authority for “the Federal Reserve to implement monetary policy without the need for required reserve balances”…”In these circumstances, the Board–as authorized by the act–could consider reducing or even eliminating reserve requirements, thereby reducing a regulatory burden for all depository institutions.”

    The Fed’s research staff thinks that this act would make our bankers more competitive internationally. Such thinking will destroy the system.

Leave a Reply