Five, Ten or Even Thirty Years

Neel Kashkari is clearly a very smart guy. He’s been an invaluable voice for sanity at the Fed these past few years. Doesn’t he see that something has gone very wrong here?


When people borrow money to buy a house, or businesses take out a loan to build a new factory, they don’t really care about overnight interest rates. They care about what interest rates will be for the term of their loan: 5, 10 or even 30 years. [*] Similarly, when banks make loans to households and businesses, they also try to assess where interest rates will be over the length of the loan when they set the terms. Hence, expectations about future interest rates are enormously important to the economy. When the Fed wants to stimulate more economic activity, we do that by trying to lower the expected future path of interest rates. When we want to tap the brakes, we try to raise the expected future path of interest rates.

Here’s the problem: recession don’t last 5, 10 or even 30 years. Per the NBER, they last a year to 18 months.

Mainstream theory says we have a long run dictated by supply side — technology, demographics, etc. On average the output gap is zero, or at least, it’s at a stable level. On top of this are demand disturbances or shocks — changes in desired spending — which produce the businesses cycle, alternating periods of high unemployment and normal growth or rising inflation. The job of monetary policy is to smooth out these short-term fluctuations in demand; it absolves itself of responsibility for the longer-run growth path.

If policy responding to demand shortfalls lasting a year or two, how is that supposed to work if policy shifts have to be maintained for 5, 10 or even 30 years to be effective?

If the Fed is faced with rising inflation in 2005, is it supposed to respond by committing itself to keeping rates high even in 2010, when the economy is sliding into depression? Does anyone think that would have been a good idea? (I seriously doubt Kashkari thinks so.) And if it had made such a commitment in 2005, would anyone have worried about breaking it in 2010? But if the Fed can’t or shouldn’t make such a commitment, how is this vision of monetary policy supposed to work?

If monetary policy is only effective when sustained for 5, 10 or even 30 years, then monetary policy is not a suitable tool for managing the business cycle. Milton Friedman pointed this out long ago: It is impossible for countercyclical monetary policy to work unless the lags with which it takes effect are decidedly shorter than the frequency of the shocks it is supposed to respond to. The best you can do, in his view, is maintain a stable money supply growth that will ensure stable inflation over the long run.

Meanwhile, conventional monetary policy rules like the Taylor rule are defined based on current macroeconomic conditions — today’s inflation rate, today’s output gap, today’s unemployment rate. There’s no term in there for commitments the Fed made at some point in the past. The Taylor rule seems to describe the past two or three decades of monetary policy pretty well. Now, Kashkari says the Fed can influence real activity only insofar as it is setting policy over the next 5, 10 or even 30 year’s based on today’s conditions. But what the Fed actually will do, if the Taylor rule continues to be a reasonable guide, is to set monetary policy based on conditions over the next 5, 10 or even 30 years. So if Kashkari takes his argument seriously, he must believe that monetary policy as it is currently practiced is not effective at all.

In the abstract, we can imagine some kind of rule that sets policy today as some kind of weighted average of commitments made over the past 5, 10 or even 30 years. Of course neither economic theory nor official statements describe policy this way. Still, in the abstract, we can imagine it. But in practice? FOMC members come and go; Kashkari is there now, he’ll be gone next year. The chair and most members are appointed by presidents, who also come and go, sometimes in unpredictable ways. Whatever Kashkari thinks is the appropriate policy for 2022, 2027 or 2047, it’s highly unlikely he’ll be there to carry it out. Suppose that in 2019 Fed chair Kevin Warsh  looks at the state of the economy and says, “I think the most appropriate policy rate today is 4 percent”. Is it remotely plausible that that sentence continues “… but my predecessor made a commitment to keep rates low so I will vote for 2 percent instead”? If that’s the reed the Fed’s power over real activity rests on it, it’s an exceedingly thin one. Even leaving aside changes of personnel, the Fed has no institutional capacity to make commitments about future policy. Future FOMC members will make their choices based on their own preferred models of the economy plus the data on the state of the economy at the time. If monetary policy only works through expectations of policy 5, 10 or even 30 years from now, then monetary policy just doesn’t work.

There are a few ways you can respond to this.

One is to accept Kashkari’s premise — monetary policy is only effective if sustained over many years — and follow it to its logical conclusion: monetary policy is not useful for stabilizing demand over the business cycle. Two possible next steps: Friedman’s, which concludes that stabilizing demand at business cycle frequencies is not a realistic goal for policy, and the central bank should focus on the long-term price level; and Abba Lerner’s, which concludes that business cycles should be dealt with by fiscal policy instead.

The second response is to start from the fact — actual or assumed — that monetary policy is effective at smoothing out the business cycle, in which case Kashkari’s premise must be wrong. Evidently the effect of monetary policy on activity today does not depend on beliefs about what policy will be 5, 10 or even 30 years from now. This is not a hard case to make. We just have to remember that there is not “an” interest rate, but lots of different credit markets, with rationing as well as prices, with different institutions making different loans to different borrowers. Policy is effective because it targets some particular financial bottleneck. Perhaps stocks or inventories are typically financed short-term and changes in their financing conditions are also disproportionately likely to affect real activity; perhaps mortgage rates, for institutional reasons, are more closely linked to the policy rate than you would expect from “rational” lenders; perhaps banks become more careful in their lending standards as the policy rate rises. One way or another, these stories depend on widespread liquidity constraints and the lack of arbitrage between key markets. Generations of central bankers have told stories like these to explain the effectiveness of monetary policy. Remember Ben Bernanke? His article Inside the Black Box is a classic of this genre, and its starting point is precisely the inadequacy of Kashkari’s interest rate story to explain how monetary policy actually works. Somehow or other policy has to affect the volume of lending on a short timeframe than it can be expected to move long rates. Going back a bit further, the Fed’s leading economist of the 1950s, Richard Roosa, was vey clear that neither the direct effect of Fed policy shifts on longer rates, nor of interest rates on real activity, could be relied on. What mattered rather was the Fed’s ability to change the willingness of banks to make loans. This was the “availability doctrine” that guided monetary policy in the postwar years. [2] If you think monetary policy is generally an effective tool to moderate business cycles, you have to believe something like this.

Response three is to accept Kashkari’s premise, yet also to believe that monetary works. This means you need to adjust your view of what policy is supposed to be doing. Policy that has to be sustained for 5, 10 or even 30 years to be effective, is no good for responding to demand shortfalls that last only a year or two at most. It looks better if you think that demand may be lacking for longer periods, or indefinitely. If shifts in demand are permanent, it’s not such a problem that to be effective policy shifts must also be permanent, or close to it. And the inability to make commitments is less of a problem in this case; now if demand is weak today, theres a good chance it will be weak in 5, 10 or even 30 years too; so policy will be persistent even if it’s only based on current conditions. Obviously this is inconsistent with an idea that aggregate demand inevitably gravitates toward aggregate supply, but that’s ok. It might indeed be the case that demand deficiencies an persist indefinitely, requiring an indefinite maintenance of lower rates. There’s a good case that something like this response was Keynes’ view. [3] But while this idea isn’t crazy, it’s certainly not how central banks normally describe what they’re doing. And Kashkari’s post doesn’t present itself as a radical reformulation of monetary policy’s goals, or mention secular stagnation or anything like that.

I don’t know which if any of these responses Kashkari would agree with. I suppose it’s possible he sincerely believes that policy is only effective when sustained for 5, 10 or even 30 years, and simply hasn’t noticed that this is inconsistent with a mission of stabilizing demand over business cycles that turn much more quickly. Given what I’ve read of his I feel this is unlikely. It also seems unlikely that he really thinks you can understand monetary policy while abstracting from banks, finance, credit and, well, money — that you can think of it purely in terms of an intertemporal “interest rate,” goods today vs goods tomorrow, which the central bank can somehow set despite controlling neither preferences nor production possibilities. My guess: When he goes to make concrete policy, it’s on the basis of some version of my response two, an awareness that policy operates through the concrete financial structures that theory abstracts from. And my guess is he wrote this post the way he did because he thought the audience he’s writing for would be more comfortable with a discussion of the expectations of abstract agents, than with a discussion of the concrete financial structures through which monetary policy is transmitted. It doesn’t hurt that the former is much simpler.

Who knows, I’m not a mind reader. But it doesn’t really matter. Whether the most progressive member of the FOMC has forgotten everything his predecessors knew about the transmission of monetary policy, or whether he merely assumes his audience has, the implications are about the same. “The Fed sets the interest rate” is not the right starting point for thinking about monetary policy manages aggregate demand.


[1] This is a weird statement, and seems clearly wrong. My wife and I just bought a house, and I can assure you we were not thinking at all about what interest rates would be many years from now. Why would we? — our monthly payments are fixed in the contract, regardless of what happens to rates down the road. Allowing the buyer to not care about future interest rates is pretty much the whole point of the 30-year fixed rate mortgage. Now it is true that we did care, a little, about interest rates next year (not in 5 years). But this was in the opposite way that Kashkari suggests — today’s low rates are more of an inducement to buy precisely if they will not be sustained, i.e. if they are not informative about future rates.

I think what may be going on here is a slippage between long rates — which the borrower does care about — and expected short rates over the length of the loan. In any case we can let it go because Kashkari’s argument does work in principle for lenders.

[2] Thanks to Nathan Tankus for pointing this article out to me.

[3] Leijonhufvud as usual puts it best:

Keynes looked forward to an indefinite period of, at best, unrelenting deflationary pressure and painted it in colors not many shades brighter than the gloomy hues of the stagnationist picture. But these stagnationist fears were based on propositions that must be stated in terms of time-derivatives. Modern economies, he believed, were such that, at a full employment rate of investment, the marginal efficiency of capital would always tend to fall more rapidly than the long rate of interest. … When he states that the long rate “may fluctuate for decades about a level which is chronically too high” one should … see this in the historical context of the “obstinate maintenance of misguided monetary policies” of which he steadily complained.

12 thoughts on “Five, Ten or Even Thirty Years”

  1. I think you have already mentioned (in a previous post) the logical outcome of such long-term commitments by the central bank. If stabilizing the current business cycle requires lowering long-term interest rates through committing to low short-term rates then each business cycle will be accompanied with even lower interest rates until we eventually reach the ZLB.

    This is roughly the story of the federal funds rate since 1980.

    1. Yes, I was thinking of making just this point in this post, but decided it was overstuffed enough already.

      This was Kalecki’s view, incidentally.

  2. I am just an amateur but I tend to see monetary policy working the way Kashkari does, primarily via expectations of private lenders and investors. It also works via the concrete mechanisms of setting credit prices which help demonstrate the Fed’s intentions. As Kashkari points out, the Fed’s commitment is to keep inflation and prices stable over the long run (and unemployment low). This is what investors and lenders should expect if the Fed is doing its job.

    The problem is that the Fed has not been doing its job. It said it wanted a quick recovery but didn’t do enough (either via concrete measures or communications) to bring about a quick recovery.

    It could have had it had wanted to which is where I differ from the fiscalists who oppose monetary policy and say it doesn’t work.

    I agree with fiscalists that fiscal policy does a much better job at providing concrete mechanisms to induce recovery. Fed officials were supportive of Obama’s stimulus bill.

    The Fed is making vague guesses at what the market needs to hear and see in order to make its long range commitments a reality.* The markets are making vague guesses on whether the Fed is going to keep its commitments to price stability and growth or whether its flaking out and delivering another crappy recovery. And maybe the markets aren’t looking primarily at the Fed’s action but at more immediate indicators like market demand, etc.

    Any businesses or banks that bet the Fed would deliver a rapid recovery bet wrongly. Doesn’t mean that monetary policy doesn’t work.

    *To me the Fed appears overly paranoid about inflation. They treat the inflation target as a ceiling, never to overshoot lest we accidentally slip into hyperinflation. And then some people argue somehow that inflation is hard to create. Both can’t be true.

    1. Policy might work the way he says, but in that case it really isn’t suitable for dealing with business cycles.

  3. It could be that the Fed is worried about asset bubbles even though they don’t explicitly say so. So they run the economy in low gear, grinding away at potential GDP via hysteresis. Ever since the Reagan neoliberal era of deregulation and rising inequality the economy hits dangerous asset bubbles before hitting inflation troubles.

    Monetary policy offset the loss of jobs because of the China shock, but the resulting housing bubble led to an epic financial crisis. Stimulative fiscal policy would avoid the dangers of asset bubbles in a deregulated economy.

  4. Haven’t you answered the question without noticing? You don’t care about the policy rate over the other 29 years of your mortgage, but you sure as hell cared about the rate on the day of signing.

    The flow of new credit into the economy has a diverse term structure. That structure is influenced, today, by the policy rate. As a rule, borrowers pay more for longer terms because they get rid of the liquidity risk associated with each event of refinancing (Keynes called this lender’s risk). But the higher today’s policy rate, usually, the higher the rest of the term structure for loans contracted today, along some yield curve.

    Sometimes, the central bank tries to alter the term structure explicitly eg Operation Twist/TLTRO. And sometimes, the government changes the institutional set-up so as to get more long lending.

    Have a look at the papers in this blog post:

    1. Sure, the policy rate today influence should the long rate today. The question is how it does so. The mechanism Kashkari describes – the expectations hypothesis — if correct, suggests that monetary policy cannot really be used at business cycle frequencies. Of course other links are possible.

      It is true that if we accept a strong form of the expectations hypothesis the statement that I care about long rates today is the sam as saying I care about short rates in the future. But the expectations hypothesis is exactly what’s in question here.

  5. A really vague tought:

    It seems to me that the argument revolves about the decisions of the lenders, so the problem is whether the changes in the interest rate change the choices of the lender.

    However, since I’ve an underconsumptionist view of the economy, I tend to think that the problem is to make people spend more, so I would expect the interest rate policy to influence the choices of the borrowers, if it has to have any effect at all.
    The borrowers aren’t really that interested in the rate 30 years from now, as you and others have pointed out.

    Does this change something? Is it possible that in some situations changes in the interest rate policy affect borrowers, and in other cases lenders? (like in, in slumps a fall in the interest rate influences borrowers, but in booms it influences lenders?)

    1. It would have to be a two step process. Borrowers look at rates today, and lenders set rates today based on their believes about future rates.

  6. The larger point is that many thoughtful central bankers (Roosa, Bernanke, etc) have believed that monetary policy does not work via interest rates in general moving with the policy rate, but through some other channel.

Comments are closed.