Part of the Fed's transparency schtick is that, barring surprise shocks, it intends to only change rates at its pre-scheduled meetings (http://www.federalreserve.gov/fomc/#calendars) .

The futures market is pricing in move probabilities factoring in probable forward dates.

I'm surprised this is an issue.

Central banks try to avoid embarassing policy reversals, which means they minimise the number of times they reverse the direction of interest rate changes. Hence they make many small interest rates changes in the same direction, lagging the data flow, rather than moving in interest rates in discreet jumps. The result of this lag-the-data behaviour is that interest rate changes are predictable to the futures market, making the yield curve an excellent predictor of policy.

Look, I think the thinking might (I'm only guessing) go like this. There are two main calls running in parallel at the moment: that the recovery is slow but continuing, or that a double-dip is on the way (of course there are more possibilities but let's keep it simple for now). The probabilities just shifted in favour of the second with the quarterly drop in the stock markets, but both outcomes are still in the running. In the first case, dropping the rate isn't necessary, and in the second a whole shock package is supposedly being prepared (see Krugman piece on this blog last week). In either event a small change in the open market rate isn't likely to make too much dent in the basket of decision items facing both companies and consumers in the short run right now.

Of course the markets are only reflecting what many imagine might happen, whereas the central banker is playing poker, and trying to influence outcomes. This may explain the thinking. Whether it's right or not, only time will tell.

Brad, it all depends on how historically accurate your estimator is at estimating future value. In engineering, you'd use the best mathematical model you could come up with to predict ahead. The ability to predict ahead is limited by the error in your knowledge of the entire current state, assumptions or simplifications in your model, and unpredictable, unmodelable inputs to the model. If the futures market (realize, I have very little idea what this is or how well it can predict) has historically been able to predict the Federal Funds rate four months in advance with high accuracy (say, ten percent error in the predicted value, or 1 point if you prefer) then I'd say you have a point.

But then again, if there's no causal model relating pertinent variables (I have no idea what those might be in this case) and the market just happened to be right nearly every time, I'd say letting the market determine the funds rate is a mistake. And is really correct to reduce the rates in pace with market expectation? Does market expectation have any real meaning? I'm not contradicting you; I really don't know and am interested in how this works. To me, this is not unlike saying "let's manipulate corn prices in pace with the futures". Of course, in this case it's the Fed that sets the rates, not really market forces.

And if, in the past, the futures market was allowed to determine the Federal Funds rate, then it was a self-fulfilling prophecy. In other words, your model was an accurate predictor because there was a strong feedforward path in the process that helped drive the state to the predicted value.

Which has more impact on the economy: market set borrowing rates or the announced Fed Funds rate?

fwiw, the 2year treasury yield is still trading below 1.75(1.69)

I happen to agree with Brad on the policy concerns, but what this really points out is that graduate school maths for economics is focused on control-engineering optimisation techniques which tend to make the assumption that one is dealing with an ergodic system; an assumption usually good enough for control engineers, but one which is quite unwarranted in economics.

Systems with non-Hamiltonian dynamics (including some surprisingly simple systems of equations; viz, those described by Zeeman & Thom's "Catastrophe Theory") can have the characteristic that the desired control response function is discontinuous.

There is no request in this post for some bright button in their second year of economics graduate school to pop up and say that "Chaos Theory hasn't amounted to much in economics", but I doubt that this means it won't happen ...

(If there are two of these, my apology. First one didn't seem to go.)

I agree that big discontinuities in policy variables lead to inefficiency. So do Fed officials. That is why they telegraph policy changes when they have a strong idea what policy changes are coming. That can give the impression that the Fed follows the market, but except for a reluctance to gratuitously shock markets, the Fed decides, hints, and then validates market pricing after markets take the hint, rather than allowing markets to dictate policy. Fed officials also seem to think that big swings in other variables, like output, employment and inflation are bad, and are willing to sacrifice continuity in policy variables for the sake of smoothing the path of other, more important variables.

There are a couple of other misconceptions in the comments. It is not a given that central bankers are embarassed by having to reverse themselves on policy. Greenspan has been pretty open in saying that overshooting is preferable to undershooting, since correcting a policy overshoot is pretty easy after the fact. What with lags and all, overshooting doesn't really have much impact on the important dependent variables if corrected quickly enough. Volker was clearly willing to overshoot, though I just don't remember whether he engaged in much short term poicy reversal.

As to the assertion that the yield curve is an excellent forecaster of Fed policy, it certainly hasn't been this year. Neither the Treasury curve nor short term interest rate futures (Fed funds and eurodollars) pointed in the right direction in the spring of this year, when they were advertising a rate hike in June followed by more into year end. I recall something like 3% Fed funds implied in both markets for December at that time. Now, Brad tells us, it's more like 1.33%. In the interim, the Fed hasn't adjusted rates at all. As noted above, Fed officials do telegraph policy changes when they can. That has made market prices a better forecaster of policy one meeting ahead, because that is about how far Fed officials are able to see, sometimes.

Brad's problem is not solved. Market participants do, in fact, price in policy changes several months hence that Fed officials do not plan to make. Sometimes they are right. Sometimes, like earlier this year, they are wrong. The engineering view is an interesting one, but of course, engineering is not the only discipline that argues for minimizing shocks to the system. Farmers who want to optimize weight gain for hogs can tell you the same thing, without all the fancy jargon. There are just a lot of math jocks who have dabbled in both engineering and economics, and can use jargon from one to talk about the other.

Suppose that the Fed board think the economy is on the mend while investors think the economy will continue to grow weakly if at all. The stock and bond markets seem to be pricing in another quarter of weak growth. This is a different kind of recession and bear market than has been the norm since 1945. Neither the economy nor the stock market made significant gains with the series of 11 Fed rate reductions.

We may be in poorer shape than the Fed realizes. I hope not, but I am having trouble understanding why interest rates are so low if a strong growth period can be viewed as in the offing.

Paul Krugman

"I don't need to tell you about the stock market. Economic indicators strongly suggest that the economy is either sliding into a double-dip, "W-shaped" recession — bet you thought I was talking about the guy in the White House — or close enough as makes no difference. Bond markets are clearly predicting that the Fed will have to cut interest rates again. What if the Fed, like the Bank of Japan, goes all the way to zero and finds that it still hasn't turned the economy around?

"Not many people realize that in some ways Japanese economic policy responded quite effectively to a sustained slump. It's easy to make fun of the country's enormous spending on public works — all those bridges to nowhere in particular, highways with no traffic, and so on. Without question enormous sums have been wasted. But it's also clear that all that spending pumped money into the economy, preventing what might otherwise have been a full-fledged depression."

Wow. I didn't know the Fed Funds curve was a Rorschach Test.

I'm afraid I don't understand this argument.

If you want policy to be maximally effective in the face of rational expectations, you have to have uncertainty about what the Fed will do, right? That would seem to imply that large, sudden changes would have a larger effect than the same changes applied gradually, because it would be harder for the market to predict them.

What am I missing?

Perhaps I took one too many applied mathematics-engineering courses. But I was taught that in a properly-managed system one should never expect to make rapid changes in the system's control variables ...

Apropos of nothing, as folk wisdom goes, this isn't the least bit memorable. It's fortunate Prof. DeLong has some distance between now and geezerhood to come up with something pithier.

>>What am I missing?

Well, in a rational expectations environment, why would you want policy to have any effect on the market at all? How would a rational expectations economy get into a situation in which there was any need for the surprise action you're describing?

>> Well, in a rational expectations environment, why would you want policy to have any effect on the market at all? How would a rational expectations economy get into a situation in which there was any need for the surprise action you're describing?

That makes a lot of sense. I was thinking of the possibility that there might be a rational bubble in progress, which a policy maker might want to pop early. I guess that's fairly unlikely to happen in practice, though.

If George Bush shoots himself, the economy will rebound and then interest rates will have to rise.

The dow jones industrial average should not be a barameter for how well the economy is recovering. The bubble did not burst for no reason. The dow is in another short term bubble which financial institutions will ignore just to make some commision. The upside bias of the dow is more prevalant than ever,so, be cautious.

The dow jones industrial average should not be a barameter for how well the economy is recovering. The bubble did not burst for no reason. The dow is in another short term bubble which financial institutions will ignore just to make some commision. The upside bias of the dow is more prevalant than ever,so, be cautious.