Tuesday, January 29, 2008

When does monetary policy become ineffective?

Mark Thoma* leaves a succinct comment on my previous post:
Where we differ is the point at which monetary policy loses its effectiveness - I think that happens way before i-rates hit zero.
It’s an interesting point, because it is a position that many economists (including Keynes himself) seem to have held over the years, but one which, as far as I can tell, has never made much sense.**

I should be more specific: It may make sense if you measure monetary policy in certain ways, but not if you measure monetary policy in the way that is reasonable given how today’s central banks set policy. One might be (but in my opinion shouldn’t be) inclined to measure monetary policy in terms of the volume of open market operations, or some similar measure. In that case, it is quite true that a volume of operations that was effective when the interest rate was 5% is no longer likely to be effective when the interest rate is 1%. And certainly the people responsible for conducting those operations do need to be concerned with the volume. But for us, as economists and such, who can and should view monetary policy with some degree of abstraction, it makes little sense to concern ourselves with the volume of such operations. The transaction costs associated with open market operations are tiny (and not proportional to volume anyhow); the market for Treasury bills to be purchased is vast and quite liquid; the absolute size of an open market operation is of little importance, except inasmuch as it affects other variables, such as interest rates. Moreover, the same argument applies to other “quantitative” measures of monetary policy, such as changes in bank reserves and changes in monetary aggregates.

Since today’s central banks (and the Fed in particular) generally define their policy stance in terms of an interest rate, the reasonable way to measure that stance is in terms of an interest rate. Now one might argue (but again, I don’t think one should) that a proportional change in the interest rate that was effective when the interest rate was high will no longer be effective when the interest rate is low. For example, if the interest rate is 8% and you lop off one fourth of it, making the interest rate 6%, that could be quite an effective policy move; but if the interest rate is 1% and you lop off one fourth of that, making the interest rate 0.75%, that is not likely to be very effective at all by comparison. But central banks don’t measure interest rates proportionally, they measure them in…usually 25 basis point increments. And a 75 basis point cut by the Fed, for example, is considered a big move whether the interest rate starts at 8% or at 3%. The sensible question, it seems to me, is whether the effect of a given cut – defined in basis points – will be diminished when interest rates are already low.

If anything, I would argue, the exact opposite should be true. Monetary policy works largely by affecting the discounted value of expected returns on capital assets. When the Fed cuts interest rates, all other things being equal, stocks are worth more, houses are worth more, factories are worth more, machines are worth more, contemplated investment projects are worth more, and so on. The more the value of an asset rises relative to the cost of producing it, the more it becomes profitable to employ people in producing that asset. And theory says this effect should get stronger the lower are interest rates to begin with.

To see the point, consider a world where the risk premium is not an issue and where the Fed sets long-term interest rates. And just to make it clear, consider the extreme case where “long-term” means perpetual. In that case, the value of an asset that produces a fixed stream of returns equals the value of the periodic return divided by the interest rate. Thus if the Fed were to reduce the interest rate from 5% to 4%, it would increase the value of such an asset by 20%. But if the Fed were to reduce the interest rate from 1% to 0%, it would increase the value of the asset by…well, you do the math. In the enterprise of producing an infinitely valuable asset, it is of course profitable to employ as many people as you possibly can, at whatever wage they might require. In the real world, where the Fed controls only short-term rates, and where there is a risk premium associated with most assets, the effect is not so dramatic, but the difference in the effectiveness of policy when interest rates are low vs. high should certainly be in the same direction.

If, therefore, we may define “monetary policy” as the manipulation of an interest rate by a central bank, then we should expect that monetary policy gains more effectiveness the closer the interest rate comes to zero. And indeed, technically, there is no point at which monetary policy, thus defined, “loses its effectiveness.” There is, of course, a point at which additional stimulative monetary policy becomes impossible to practice, namely, the precise point when the interest rate reaches zero.***



* As far as I know, this is the real Mark Thoma – by which I mean the one that writes Economist’s View – not someone else of the same name. As an aside, though, it occurs to me that there is no shared authentication between Blogger and Typepad, so one doesn’t really know such things for sure. If I wanted to, I could probably post comments on other people’s blogs while pretending to be Brad DeLong or Barry Ritholtz. Or one of them could pretend to be me – though it's hard to think of any reason why they might want to.

** Please do not fear, Gentle Reader, that I have entertained for even a brief moment the abominable heresy that St. Maynard may have held a view that was in any way unreasonable. (Indeed, at the very thought, I must ask you to excuse me while I make the sign of the Keynesian cross over my chest.) Rather, I merely posit that there are certain inherent difficulties in communication between the Truly Awakened and ordinary sentient beings such as we. Interpreting the words of our lord**** in accordance with the mere shadows that form our limited experience, it is we who may have fallen into error. I’m personally intrigued by an alternative exegesis preached to me once by radical political economist Stephen Marglin, who suggests that Keynes was referring not to a lack of effectiveness per se but to the political difficulties in implementing a very low interest rate policy in an economy where the rentier class is loath to give up the income it receives in the form of interest.

***Strictly speaking, this is not quite true. Interest rates on some Treasury bills went below zero in 1938. Apparently, there are some people out there who like their Treasury bills so much that they won’t give them up even if you offer a premium to redemption value.

****I trust that the lower case L keeps me safe from the charge of blasphemy. Surely the lord I have in mind was indeed ours. To whom, after all, could Keynes belong***** if not to the Keynesians.

*****Actually, I have nothing to say down here, but I got so fascinated with the idea of nested footnotes that I decided to push the concept one level further.

Labels: , , , ,

5 Comments:

Anonymous Anonymous said...

Two points.

(i) Surely when the Fed cuts rates in response to event X1, the present discounted value of future income flows takes into account both the current rate (set in response to X1) and possible future rate cuts (in response to X2, X3, X4...); so I think there is an extremely good case to be made that most of the real-economy effect of cutting the rate that last 25 basis points has been front-loaded onto previous rate cuts.

(The option value of additional rate cuts, one presumes, is even higher than this implies because of the correlation between not cutting the rate and signs of recovery.)

(ii) Isn't the canonical explanation is expectations about future changes in the price level? Krugman has a snappy little essay which starts from, I take it, the view that you are advocating, and shows that it's unworkable.

Thu Jan 31, 02:14:00 AM EST  
Anonymous Anonymous said...

Your view that the rate cuts should be evaluated on the basis of their size relative to the initial rate suggests that their power should expand exponentially as the rates approach zero. This suggests that the Fed doesn't just have power, but very, very, very, *very* large power. A cut from (say) 1% to 0.5% should have an incredibly large effect on the economy.

Is that consistent with what happens?

Fri Feb 01, 10:50:00 AM EST  
Blogger knzn said...

The power should expand exponentially as the perpetual discount rate for the private sector approaches zero. In practice, there is a risk premium for the private sector, and there is the expectation that rates will eventually rise again, which affects the perpetual rate, so if the Fed's interest rate is "i", then the perpetual rate "r" is something like

r = a*i + (1-a)*R + p

where p is the risk premium,
R is the long-run expected interest rate, and
a is some positive constant less than 1.

So r never gets very close to zero, even when i is zero. But it is nonetheless true that, the closer i gets to zero, the closer r gets to zero. Thus, as i goes down, monetary policy becomes more effective but never approaches the extreme levels of effectiveness that would apply if r were near zero.

Fri Feb 01, 12:21:00 PM EST  
Blogger Unknown said...

طرز تهیه نان پونجیک سنتی با سیب گردو دارچین
طرز تهیه پیراشکی پنیری ترکیه ای و استفاده از سرشیر گوسفندی
کاهش درد قاعدگی با دمنوش گیاهی خانگی

Wed Nov 08, 06:34:00 PM EST  
Blogger yanmaneee said...

birkin bag
cheap jordans
kobe basketball shoes
kawhi leonard shoes
jordan shoes
moncler jackets
kd shoes
golden goose
stone island sale
bape outlet

Mon Mar 15, 06:32:00 AM EDT  

Post a Comment

<< Home