The premises don’t follow from the conclusion

Consider the following claims:

1. The Fed has tightened policy over the past few years.

2. The likely Fed rate cut in July shows that they made a mistake in raising rates last year.

3. The fact that unemployment has fallen to lower levels than anticipated provides further evidence that the Fed erred in raising rates.

4. The Fed should do a full 50 basis point cut in July, not a measly 25 basis point cut.

I believe the fourth claim (the “conclusion”) is true, but I don’t get to this place via the same path as many “doves”. I do not accept the first three claims (the “premises”).

1. Monetary policy has actually eased since Trump was elected, as evidenced by a significant speed-up in NGDP growth.

2. A rate cut in July does not show that the previous rate increases were mistaken (although they may have been mistaken for other reasons.) Interest rates should go up and down with changes in the equilibrium rate, and there is abundant evidence that the equilibrium interest rate did indeed rise in 2017-18, and that this rate is now falling. The Fed should probably not be targeting interest rates at all, but if they insist on doing so then it’s appropriate to raise and lower the interest rate target along with changes in the equilibrium rate.

3. The 100 basis point fall in the unemployment rate over the past couple of years is not evidence that money was too tight a year or two ago. Unemployment fell in the late 1960s, and no one in their right mind would say money was too tight in the late 60s.

Some argue against any rate cut on the grounds that the economy is booming, and in the past the Fed did not generally cut rates when the economy was strong. Perhaps we don’t need to ease policy right now, but even in that case I’d advocate an interest rate cut, because the Fed needs to cut rates just to keep the stance of monetary policy at its current level.

But I’d go even further. The fact that inflation has been mostly running below 2%, and is expected to continue running below 2%, suggests the need for the Fed to “get ahead of the curve” with a greater than expected monetary easing. That’s why I favor a 50 basis point cut.

BTW, I don’t favor inflation targeting, but if the Fed is going to adopt a symmetrical 2% PCE inflation target, then they need to produce that inflation rate, on average.

PS. Don’t say, “The 1960s were different, because today there is no inflation problem”. That’s my point. The unemployment rate isn’t the number to look at; it’s the low inflation today that tells us that money has been a bit too tight.



19 Responses to “The premises don’t follow from the conclusion”

  1. Gravatar of Michael Sandifer Michael Sandifer
    18. July 2019 at 09:14

    I’d say monetary policy’s been too tight over the past two years, because unemployment has continued to fall and the inflation rate has mostly been below the Fed’s target, using their preferred metric.

    I do agree that there’ve been brief periods over the last few years in which policy has loosened, but it was still in the context of policy being too tight. I think real GDP potential is higher than you do, even in the long run.

  2. Gravatar of Garrett M Garrett M
    18. July 2019 at 09:52

    One problem I’ve found with employing the “ahead of the curve” argument is it’s easy for someone to dismissively respond “so you’re saying they need to cut now so they can raise later?”

  3. Gravatar of Paul Paul
    18. July 2019 at 12:31

    I don’t agree with measuring the stance of monetary policy based on its supposed results (the inflation rate, price level or NGDP), because it then becomes tautological. “Easy money causes inflation, and when inflation happens money must be easy.”

  4. Gravatar of Paul Paul
    18. July 2019 at 12:35

    That’s one area where the old monetarists and quantity theorists were clearer. They believed easy money led to a higher nominal money stock, which then, through an effect on portfolio allocations, drove the price level and spending higher.

    If NGDP = monetary policy, and anything that affects NGDP must be monetary policy, there is no coherent transmission mechanism.

  5. Gravatar of Don Geddis Don Geddis
    18. July 2019 at 15:09

    @Paul: There are many many trivial things that “affect” NGDP. That doesn’t make them monetary policy. Monetary policy is about changes in the stock of money.

    The point is that the central bank’s monopoly control over the money supply gives them an infinitely powerful bazooka. It doesn’t really matter what other people are doing by throwing pebbles. The power of the central bank can overwhelm all other influences. MV=PQ, so there is always some setting for M that will result in any price level you wish for P, regardless of what minor shifts might be happening to V or Q. If inflation isn’t where the central bank wanted it to be, then the central bank simply made a bad decision and chose the wrong value for M.

  6. Gravatar of Benjamin Cole Benjamin Cole
    18. July 2019 at 15:43

    I think this post is correct.

    That said, recently every time a central bank has tightened, it has later proved to have been a questionable or premature move.

    Sure, we do not want to repeat the monetary excesses of the late 1960s or 1970s— evidently, still the defining moment for many American macroeconomists.

    But times change and realities change. Perhaps not theoretically but in fact we live in a less inflation-prone time.

    The Federal Reserve should err on the side of growth.

  7. Gravatar of Michael Sandifer Michael Sandifer
    18. July 2019 at 16:35

    Garrett M,

    You said, ‘One problem I’ve found with employing the “ahead of the curve” argument is it’s easy for someone to dismissively respond “so you’re saying they need to cut now so they can raise later?”’

    This is easy to counter. Short nominal rates should fall now, as long rates rise, the latter signalling higher expected NGDP growth. Then, as the short neutral rate rises, short nominal rates can rise too.

  8. Gravatar of Benjamin Cole Benjamin Cole
    18. July 2019 at 17:28

    Add on: Ray Dalio is raising reasonable concerns that monetary policy is losing its oomph.

    Some will posit that monetary-policy impotence is impossible, the Fed has only to keep expanding QE until the wheels gain traction. Maybe so.

    But there may be political obstacles that prevent the Fed from doing enough QE. Suppose the Fed has to buy back $8 trillion of Treasuries to purchase results? If this seems far-fetched, please see Japan. Or Europe. Yes, only the two other major developed free economies of the world. Oh, that.

    Thus, there is a good argument the Fed should “get ahead o the curve,” and err constantly on the side of avoiding the Japan-Europe scenario. Another recession, and the Fed will not have the firepower it needs, due to constraints on the level of bond-buying that it feels it can pursue. And that assumes QE works. With globalized capital markets, one could reasonably wonder if the bond-buying of a lone central bank has much effect on long-term rates. There are chronic capital gluts anyway.

    Ben Bernanke has suggested helicopter drops as a central bank option for Japan. So it is not nutty to ponder such an option, unless Bernanke is nutty. (I wear Bernanke’s tin-foil hat on this one). The ECB has deployed about $1 trillion in TLTROs (created money loaned to banks specifically to be re-lent to industry and consumers).

    These tools—money-financed fiscal programs and TLTRO—are worth exploring.

  9. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    18. July 2019 at 17:51

    I’m still trying to figure out what Scott’s joke is, with his title.

  10. Gravatar of Paul Paul
    18. July 2019 at 18:28

    Don, MV=PQ doesn’t help your case unless you impose the restriction that V is stable (as you do with the statement it is only subject to “minor shifts”).

    V = NGDP/M
    M = NGDP/V

    Therefore, anything that affects NGDP must necessarily affect M and/or V, since they are defined in terms of NGDP! No matter what monetary aggregate you’re using, you assume that aggregate is the one that matters for NGDP, and that it “circulates” a given number of times to create NGDP.

    The central bank only has control over the monetary base. As I said before, although it can create as much HPM as it wants, none of that is guaranteed to lead to shifts in the stock of currency or the broader monetary aggregates, which are determined by the profitability of lending and a whole array of economic processes.

  11. Gravatar of ssumner ssumner
    18. July 2019 at 18:35

    Garrett, Actually, in a sense they do need to cut rates so that they can raise them later.

    Paul, Then use expected NGDP growth.

  12. Gravatar of Derrick Derrick
    19. July 2019 at 05:53

    Uh oh Scott, shots fired via Bloomberg:

  13. Gravatar of Paul Paul
    19. July 2019 at 06:31

    Here’s the issue, Scott. If nothing happens to change the economic behaviors that affect NGDP, why would anyone expect NGDP in the future to be higher? The Fed isn’t a NGDP “factory” that can just force more spending anytime it wishes. It needs to depend on response functions to its actions.

  14. Gravatar of Tim Tim
    19. July 2019 at 08:31

    Scott, Can you elaborate on the abundant evidence that the equilibrium rate rose in 17-18 and is now lower? It seems like there are too many factors affecting growth and inflation to be able to extract a signal about changes in the equilibrium rate, particularly over relatively short periods. Assuming that changes in market rates imply a change in the equilibrium rate seems to be circular, and takes policy makers into Bernanke’s hall of mirrors.

  15. Gravatar of ssumner ssumner
    19. July 2019 at 12:42

    Paul, You said:

    “The Fed isn’t a NGDP “factory” that can just force more spending anytime it wishes.”

    Actually, it is.

    Tim, You need to look at both interest rates, rate expectations and inflation expectations. During much of this period, both NGDP and inflation expectations were rising, even as rates were rising. That’s pretty strong evidence that the equilibrium rate was rising.

  16. Gravatar of P Burgos P Burgos
    19. July 2019 at 16:14

    Is hyperinflation a case where a central bank isn’t actually an NGDP factory? Also, wasn’t Emi Nakamura just awarded the John Bates Clark medal in part by showing that there are cases in which central banks do not and cannot function as NGDP factories, or have I not understood her work?

  17. Gravatar of ssumner ssumner
    20. July 2019 at 04:00

    Burgos, Which Nakamura paper are you referring to?

    To be a NGDP factory, the central bank needs to be given the power to make appropriate changes in the supply and demand for base money. In countries with hyperinflation they may be forced to monetize debt. The Fed is a NGDP factory.

    However “factory” is a bad metaphor, as different monetary policies do not represent different levels of real output.

  18. Gravatar of Paul Paul
    20. July 2019 at 16:32

    Scott, even if we accept that the central bank would just need control over the S&D for base money to drive NGDP (arguable, given that it represents only 5-10% of the broad money supply), it isn’t certain that it has it in the way you envision it. What if the demand for base money is proportional to the supply, such that the velocity of circulation just falls upon any increase in the stock? This would pretty well describe the drops in interest rates we usually see associated with open market purchases.

  19. Gravatar of Monetary policy: Better to expect too much than too little – Econlib Monetary policy: Better to expect too much than too little - Econlib
    7. January 2020 at 03:15

    […] few months back I argued that it was mistake to use a falling unemployment rate as evidence that monetary policy has been […]

Leave a Reply