Yes this is what I'm talking about
I think I am right here and Krugman & Rogoff are wrong! It's not "an inflation cure" or "two or three years of inflation". What is needed is a step-change shift in the price level. These are two distinct concepts - we want a change in P with constant dP/dT, not a change in dP/dT. This has the effect of a one-off windfall redistribution from nominal creditors to nominal debtors, a one-off increase in the ratio of collateral to credit. If anticipated, it would encourage an investment boom, as owners of nominal assets attempted to exchange them for real assets.
The reason why I'm insisting on this is not just anal retention, by the way - it's that several of the world's most important central banks have it written into their constitutions that they need to control inflation or target price stability. I think a good lawyer could make a case that a decision to target stability, but stability at a higher level would be consistent with the letter if not the spirit of these laws. Also, if you're going to change the constitution of the ECB, the aims of the Fed or the contract between HMT and the Bank of England to accomodate a more expansionary policy, making the change in terms of a step shift in the price level rather than a change in the inflation target means you only have to do it once.
Can you have a step change in the price level without inflation? I mean that other than trivially, obviously you would need inflation for one second or something but is that feasible?
ReplyDeleteYou couldn't just do the opposite of a currency reform (such as old francs to new francs) and a nought on, as that would change debts as well, wouldn't it?
I think I've seen a French proposal for this - an "administered" uprating of prices and wages. It's pretty French, but then a lot of French things are also pretty.
ReplyDeleteI'm not convinced. Looking at the nightmare scenario Krugman constructs in that model, don't we have a problem as soon as Period 2 has become Period 1? The new price level is now in force but the ageing population still wants to save for a future when capacity will be lower. But if I'm right that means Krugman's model isn't consistent with rational expectations. Agents must know that a one-shot increase in the price-level isn't enough.
ReplyDeleteHowever I'm not pretending that my opinion is worth anything. I hate this stuff. My hope is that when the Paradox of the Unexpected Hanging is properly understood, it will be shown that rational expectations is a crock.
Well, if that will be realized, something good may happen.
ReplyDeleteNow that I've read the paper again I see that my last comment is wrong. Krugman assumes constant y=y* from Period 2 onwards. So a once-off jump in the level of P is enough.
ReplyDeletePresumably if we assume instead that y is going to fall forever at a constant rate, a negative real interest rate is needed in all periods, not just the first one. If the price-level refuses to fall in the first period, then only a credible promise of constant future inflation will do the trick.
ah, good point Kevin. So in the Japanese case, you did need a credible constant inflation. But surely the US credit crunch is a lot more like a situation where in the long term dY/dT is positive ...
ReplyDeleteIs this the New Augustinian economics?
ReplyDeleteIt actually looks like you are in violent agreement with K&R. Rogoff suggests that a step or perhaps accessibility ramp change in price level would be a good thing. Krugman agrees but points out that it would be hard for a central bank to achieve without making a larger commitment.
Certainly it looks like it would be hard for a central bank to persuade enough people that change was coming to make it happen without also convincing people that it wasn't going to continue to happen.
It would also be a very dysfunctional institution that believed that, say, a 20% step change preceded and followed by very low levels of inflation was more reasonably considered stable than an extra 2% for ten years.
It would also be a very dysfunctional institution that believed that, say, a 20% step change preceded and followed by very low levels of inflation was more reasonably considered stable than an extra 2% for ten years.
ReplyDeleteNot really. Step changes are totally orthodox for central banks managing the external value of a currency.
we want a change in P with constant dP/dT, not a change in dP/dT
ReplyDeleteIn other words, we "want" a mathematical impossibility.
The existence of cruise control rather destroys that assertion.
ReplyDeleteThe existence of cruise control rather destroys that assertion.
ReplyDeleteEr, no.
Really? Change in a variable (X) at a constant velocity (dX/dt)? Not ring a bell?
ReplyDeleteAnonymous, you're wrong here, clearly. I don't want to resurrect the stopped clock theorem discussion (and nobody else does, I'm sure), but the definition of a step jump is that X changes but dY/dX has the same value before and after the point that X changes, and does not exist at the point where X changes. If there's a devaluation under a fixed rate regime, for example.
ReplyDeleteAs you say, it doesn't exist at the point where X changes. Therefore at this point it is not equal to the value it had before the step change. Therefore it does not stay the same across the transition. Therefore what you propose is impossible.
ReplyDeleteUnless, of course, you want to refine your statement to include measure theory... except you don't understand that, and in fact make this ignorance the basis of pseudo-informed conundrums such as the stopped clock "theorem".
That should be Y. I was distracted by Richard J's baffling cruise control-inspired contribution.
ReplyDeleteLook up "level targeting" in Google Scholar. (Yes, I agree with you.)
ReplyDeleteOr TheMoneyIllusion blog.
Or this speech:
http://www.federalreserve.gov/boarddocs/speeches/2003/20030531/default.htm
"What I have in mind is that the Bank of Japan would announce its intention to restore the price level . . . to the value it would have reached if, instead of the deflation of the past five years, a moderate inflation of, say, 1 percent per year had occurred."
It would be nice if someone like this was running a central bank right now, wouldn't it?
Unless, of course, you want to refine your statement to include measure theory
ReplyDeleteActually I only need to refine it with the words "before and after the step change". TFH though.
Um, in case it wasn't clear, my comment above was directed at the original post, not the cruise control discussion.
ReplyDelete(Of course, in practice a "step change" would mean 2 or 3, or at least 1 or 2 years of higher inflation. But for some reason Krugman does seem to frame the option as permanent 3 or 4% inflation, rather than simply catch-up.)
Sorry, the hazards of reading a comment too quickly while waiting to go down the pub on a Friday evening. Trivially true that a discontinuity in a function causes problems in ya derivatives; that said P is an approximation of a very discontinous social reality anyway, so mathematical formalism is likely to be a moot point.
ReplyDeleteSaying that a step change in inflation is not price stability is not saying that central banks never put up with step changes. It's like saying that a devaluation is not a strong pound policy.
ReplyDeleteIs a step change desirable only because it might sneak through a loophole in central bank rules or are you claiming that it would in fact be better than adopting a higher equilibrium or promotoing higher than usual inflation for a period of two or three years? If you think it would actually be better do you also think it is possible? I think that Krugman is maintaining that it would be hard to get any inflation without at least creating the impression that inflation will continue to be higher.
I can't think of a good way of making instant inflation let alone one that would also be plausibly be a one off, that wouldn't be very disruptive and that would also provide the right kind of inflation.
It's like asking for a one time strengthening of weak currency more than allowing a devaluation. The only thing you could just let happen at the moment would be rampant deflation.
Practically speaking, since prices are sticky, a "step change" would mean higher inflation for at least a year or two. So I think that distinction is pedantic.
ReplyDeleteThe question is really whether you want to bring the inflation rate back down again afterwards.
Krugman seems to suggest that the higher inflation target would be permanent. But as the links I posted indicate, there are a lot of respectable academics/central bankers who argue for "level targeting" i.e. catching up for past under/overshooting of the target.
I took this to be the implication of what dd was advocating (although maybe he wants a bit more than catchup), and wanted to point out that yes, smart people have thought of this before, and there is a better name for it than "step change".
Declan is right that this is related to the level targetting debate (which I was around for one of the iterations of), but it's not quite the same thing as I still want there to be an unchanged inflation target - just at a higher level.
ReplyDeleteIn a friction-free universe, I'd just announce a new price level via currency reform (overstamp all the notes with an extra zero or something) and keep the inflation target constant (note to anonymous measure-theory knobbers - the inflation target would still exist and be exactly the same, even at the precise point where the price level discontinuity happened).
In real life, there would have to be a period of adjustment and repricing, but note that unlike Krugman & Rogoff, my proposal has a built-in stabiliser; we know when the tap is going to be turned off by looking at the CPI index (grossed up for however long it takes, at my inflation target). So from day one, everyone knows that the money-printing exercise has an institutional shut-off built into it. This is the same reason why, as Kevin correctly points out, the step-change target wouldn't be much use for a Japan-like situation. I'm basically using it in place of a windfall tax on bondholders, redistributed to debtors.
I think that this built in shut-off is the attractive property of a price level shift target, compared to an X-years-of-inflation target. If, during the adjustment period, we get an exogenous upward or downward shock to the price level, my policy reacts better to that than K&R's (in a situation where you want to use monetary policy. That's the main reason for proposing it, although I do think it would facilitate some sneaky legal manouvreing around central bank constitutions, in the spirit of FDR.
Bernanke in Declan's link is certainly using level targetting to slip an inflationary policy past regulatory pre-commitments.
ReplyDeleteI see the attractions of the shutoff if you are worried about inflation running away but if the target looks like more than a catch-up it will still mess with expectations and if you are having trouble achieving inflation in the first place the target might undermine your efforts.
This might be good in the UK but we already ahve inflation above target and the significant change would look a lot like a convenient default. Which it would be but I don't know how much the target would stave off the negative consequences of that.
dd - Sorry to imply you hadn't heard of it, but I still have trouble working out the differences between level targeting, K&R, and your proposal. Yours is different from level targeting because you will return to a conventional inflation target with no memory once you achieve the 1-off level rise? Or it is level targeting, but you are shifting the level as well?
ReplyDeleteWhen you say "unlike Krugman & Rogoff", what specific proposal are you referring to? I can only get very vague suggestions out of the 2 columns you linked to, not enough to say how they would handle missing their SR target (although Krugman in other places talks about a permanent 3 or 4% target). If the distinction is just about getting it past the lawyers, fair enough, but you seem to be implying more by saying they are "wrong". I agree that a memory-less inflation target of say 4% inflation for 2 years and then 2% would be clearly inferior to a level target of 2% inflation starting from 4% above the current price level. But I don't think K&R have been that specific.
I agree with Bunbury that a level shift would be a lot more problematic than catch up, both in terms of constitutions and anchoring expectations. But in the US situation that K&R are discussing, catch up might be enough.
Bunbury - I don't understand "if you are having trouble achieving inflation in the first place the target might undermine your efforts." Isn't setting a target with memory a good way of committing to inflation (no-one thinks the liquidity trap will last forever, which is what K's model is built on)? The Japanese then, and the Fed now, did not even make it clear that they wanted much more than zero inflation.
For a company to raise prices is a gamble. If companies are confident that prices are going to keep rising they can have confidence that prices will catch up but if the end is in sight they will be tempted to keep prices low to steal market share from rivals. The price shift is a coordination problem and the limit makes it an endgame.
ReplyDeleteOnce there is a target the required inflation must still be achieved. In the US very loose monetary policy and other stimuli have not even achieved the target. If there are further measures that could be more effective but are not being used because they present too large a risk of high inflation in the LNG term the target might help but I'm not sure what they might be.
I wouldn't say the US has "very loose monetary policy". They were initially slow to cut rates, and started paying interest on reserves in the middle of the crisis. If low interest rates = loose money, then there was loose money in the Great Depression.
ReplyDeleteThe setting of a target is the crucial first step in achieving it. At the moment it is not even clear that the Fed thinks inflation is too low! The very fact of setting a target should raise expected inflation and therefore cut the real interest rate. As for other tools, unlike everyone else the Fed probably can't devalue, but ending interest (or charging penalties) on reserves and more QE is possible - Bernanke has repeatedly outlined more "nuclear" options, but doesn't think (or can't convince the FOMC) they are needed yet.
Anyway I'm just an amateur on this - http://www.themoneyillusion.com/?p=6937 and the FAQs say it better than I do.
Embarassingly, it took a post on Matthew Yglesias' blog to make me realise that Declan is right and there is absolutely no difference between what I'm talking about here and normal level targetting. I'm now struggling to remember why I thought there was.
ReplyDeleteIf you say so! I thought it sounded different for one of the reasons above, but couldn't work out exactly which.
ReplyDeleteAnyway, it is still a good idea. "Average" or "over the cycle" inflation targets should have room for it.