« The eurozone pain is mainly in Spain | Main | 2011* »


Feed You can follow this conversation by subscribing to the comment feed for this post.

I'm confused. What makes you think (1) is the case?

Jason: Well, if I can reason that I might be tempted to try to get people to expect higher inflation, maybe some of the people who are telling people that inflation will be high might have followed the same reasoning as me. Probably not, but I can't rule it out.

I think that is how the Fed sees it. They fear negative talk would be self-fulfilling so they talk about inflation and how they are prepared to fight it, disclaiming any need to do more. How much of what they say do they mean? The question is how much is in the control of expectations and how much facts. Can expectations reverse facts? Can facts reverse expectations? When are they self-fulfilling and when self-defeating? Facts are present and expectations future, but when they disperse so far, how believable can expectations be?

That is an amusing fancy, but once again you are asking us to consider seriously a proposal that will not bear the weight of too much examination.

Expectations have their place, but they also have their limits. If you will allow me the same latitude to make sweeping generalizations that you occasionally grant yourself, I would characterize the difference between "micro" and "macro" by saying that macro obliges us to consider the finiteness of things. We may not make convenient assumptions such as unlimited buyers, sellers, or money, and must account for the interaction between things as a given change propagates throughout the system - in other words, we deal with general rather than partial equilibria.

The relevance to the present discussion is that you have attributed to expectations the propensity to hold cash without considering the alternatives. As in, the alternatives available if one is not to hold cash. What asset would you buy in lieu of cash? You have already ruled out business reinvestment. Gold, a traditional safe asset, is now behaving like a risk asset - check out its correlations. Real estate is another traditional haven, but in the US it is now oppressed by a frightening degree of selling overhang. One might buy non-perishable commodities on the theory that they will retain their intrinsic value - but so many have already acted on this notion that commodity prices are now running well ahead of final demand. Cash seems no more than the least-bad alternative, rather than a positive expression of expectations.


Good point about the drawbacks to framing monetary policy with a time-path of interest rates. But most such central banks spend significant resources trying to figure out what inflation rates people expect (with surveys, yield curve models, inflation-indexed security yields, international interest rate differentials, and just plain reading the press and talking to people.) And my understanding is that they try to react systematically to changes in expected inflation (the Taylor princple gets talked about a lot by these guys.)

So my question is whether you think there is much possible impact (assuming that you could change expectations, which is a whole different can of worms.)

I think it's really fear of having to fall on your sword after thirty years of fighting inflation since Volcker's day. It's fear of admitting that you were wrong.

At best it means having to stop that policy and embrace what you said you wouldn't do. Inflation targeting was supposed to continue indefinitely.

What's the difference between switching from fighting inflation to fighting deflation in the name of an inflation target and not fighting inflation at all? Nothing except credibility, and since the central bank would have to reverse itself it would have low credibility.

It seems so much of the business and financial world just can't get their heads around deflation.

Nice post. When I was doing a recent post criticizing Raghu Rajan (who opposes easier money) I thought about the paradox of my position. He opposes holding rates at 0% until 2012. So do I. I want the Fed to raise their inflation target high enough to get a rapid recovery, in which case central banks following the Taylor rule will WANT TO raise rates before 2012.

I like your third point.

So you are saying that with a screwed up perceived transmission mechanism, the Fed should consider outsourcing monetary policy to Miley Cyrus (or anyone else whose communications about future inflation might influence the public even when people aren't sure how the prognosticator can go about effectuating their predictions). If Paul Krugman can compete with the Fed in conducting monetary policy, it's a good sign they haven't done too well figuring out how to use their monopoly power.

Hopefully this leads to more on the unfinished discussion between you and Adam P. on an alternative real-bill framing of policy. I remember the conversation getting caught up in whether the Fed should target the price of GE stock. What I don't understand is why that possibility has to include a target for the instrument. This is probably naive, but why can't the Fed simply undertake inflation targeting and use real bill operations as an instrument to get there (to avoid losing inflation targeting credibility at the zero bound), without have any particular "target" for the instrument at all? Every week the Fed would buy or sell some proportionate (based on market value outstanding, so as not to screw with signals about spreads) amount of real and nominal assets (stocks, bonds, TIPS, treasuries) without any target for interest rates, money supply or asset prices. The only target is the inflation target and the evidence of the transmission mechanism if needed would just show up on the Fed's balance sheet every week.

It is interesting that during the depression, even after policy had been turned around, after inflation had been created and growth returned, private debt continued to be liquidated for the rest of the decade. Was that a lack of belief, a failure of confidence, faltering expectations, or fear itself? Has the institutional credibility of the Fed been so damaged that whatever they do or say, only building one's personal ark is acceptable policy and cries of fire are noted but flood takes precedence in the public imagination even if the threat recedes.

Scare-mongers? What's scary is being asked to swallow dubious propositions and seeing lots of conformists gulping them down.

The answer is not to call them 'scare-mongers'. The answer is plainly admit that there is some serious long-term risks with the Fed's balance sheet. An honest confrontation of that fact would do more to restore credibility by instilling the notion that positive action is necessary for restrained future inflation rather than the present sentiment that we're in the mud and the wheels are spinning.

"Or we, who were stupid enough to frame monetary policy in such a way to make their doing that the right thing to do?"

Honesty is the right thing to do. Then money goes back to be neutral, and we all blissfully go back to being employed.

I find simultaneous satisfaction of your assumptions 2 & 3 very implausible. Monetary authorities are people.

Lord: for a given time-path of nominal interest rates, an awful lot of what happens is due to expectations of inflation and real growth. See my answer to Simon van Norden below. But yes, central bankers want to "talk up" the economy at present, but they can't be overly optimistic, since if they are wrong it would harm their credibility.

Phil: given the low nominal interest rates on short safe government debt, plus the large quantity of cash being held, plus low real investment, all the signs point to people believing that cash and government bonds are good relative to real investment. Higher expected inflation would make cash and bonds less attractive; higher expected real growth would make real investment more attractive.

Simon: when we are not at the lower bound, as long as the central bank can change the expected time path of nominal interest rates more quickly than expectations of inflation and real growth can change, the Taylor Principle seems to be able to do the job, and keep us close to the inflation target. The credibility of the inflation target helps satisfy this condition, of course. But when nominal interest rates hit zero, and an inflation target (not that the US really had one) loses credibility, this condition is not satisfied.

As to the magnitude of the effect of expectations: I would say it's big. A 1 percentage point increase in expected inflation should be equivalent to a 1 percentage point drop in nominal interest rates. Then there's expected real growth on top of that. Plus any Fisherian debt-deflation effects.

Determinant: The Bank of Canada has always (well, for the last 10-15 years) made a big deal about the inflation target being symmetric. Inflation 1% below target is as bad as inflation 1% above target. That's not true for the ECB. Not sure about the Fed. That may be part of the problem.

Scott: thanks! I know what you mean about the difficulty of expressing one's views when the "dominant narrative"(? Oooh!) makes them sound nonsensical. "I want the Bank of Canada to loosen monetary policy by doing things that would raise nominal interest rates". What?!

dlr: your "naive" question is a good one. I'm not sure about the answer. It might be that what the Fed does today matters little, except insofar as what it does today affects expectations about what it will be doing for the next few months or years. And some sort of announced target path for the "instruments" helps influence those expectations.

Lord: It might be in part just that the fall in the price level had caused real debt to rise, and people were slowly bringing their debt levels back to a long run equilibrium. (I'm talking about both debtors and creditors of course, since there can't be one without the other).

Jon: "scare-mongers" was a bit tongue in cheek. But if expectations are stuck in one equilibrium, anyone who tries to get people to believe in the second equilibrium will sound a bit crazy. And, *given existing expectations* those people *are* totally wrong, of course.

Manuel: you lost me on that one. People make mistakes, and get things wrong. Especially difficult things, like how best to frame monetary policy. It's "stupid" only in hindsight.

Jon, the problem is that there's no actual evidence to support your claims about a looming inflation problem. You're going to have to tell some kind of coherent story about how the US is going to go from it's current state of near deflation, high unemployment, and huge capacity underutilization to rampant inflation. And while you're at it, you probably want to explain why it is the Fed couldn't contain inflation if it did appear.

You have a point here Nick. In the same sense, people who keep mentioning of deflation can also be called ‘scaremongers’ because the are scaring people from buying and investing, and may even be encouraging more people to default. This leads to them also being right.

So, in a nutshell: Krugman and DeLong are personally killing the recovery, right? ;)

Nick, one day you'll have to explain how monetary policy should be "framed" and how it is different from framing as a time path for the policy rate.

In particular, you have said on other posts that policy now should be for a sufficiently large permanent increase in the money supply. How is that different than promising a time path for interest rates.

What I mean is, there will be some value for the interbank interest rate right? So, suppose the CB increases the money supply by some amount dM and promises that this monetary injection is permanent. Further assume that everyone believes this addition to the money supply is permanent. there will be some path for the short interest rate that the permanently higher money supply implies right?

Now suppose instead the CB promises exactly the path for interest rates that is implied by the permanent increase in the money supply of dM. Is that different?

Or do you mean the CB should conduct monetary policy some other way, not by promising a time path for the policy interest rate and not by promising a time path for the money supply?


The "fire in Noah's flood" analogy is a false one. It implies that inflation and deflation are at opposite poles. While this may hold true in most cases, it doesn't, demonstrably, in all.

Imagine a country with high (10% of gdp) but not growing fiscal deficits, so that fiscal stimulus is nil. In this country, the only reason why nominal rates are low is because of deflation expectations. This allows the country to finance its deficit at little cost to the budget. Now assume that deflation expectations worsen. Actors will expect the deficit to worsen, and the central bank to finance a portion of it. The worse the deflation expectations, the higher the expected deficit, the higher the proportion of central bank financing. So far, I don't think anyone would argue with the scenario, but here's where it gets tricky. At some point, market actors would begin to hedge against the tail risk that consistent, permanent increases in the money supply would raise the price level. Let's assume hedging involves storing a portion of wealth in an alternative currency. That currency would rise in value, creating a signal of future inflation, which would result in more hedging. Eventually this hedging would spread to real assets and commodities, which would also rise in price. Here is where we part ways. You would argue this is exactly the intended effect, and that higher nominal prices would reduce real wages and unemployment, and therefore boost output. I would argue that the hedging might raise nominal prices with zero impact on output, because the whole reason for hedging is the actors' pessimism about real growth prospects (which create high deficits).

What I described above is basically what happens in Latin countries with chronically low real growth prospects and high and permanent fiscal deficits financed by central banks. In this case, one should race over to the dock and scream "fire" the more the water level rises. In fact, this is just what happens to the sovereign bonds (domestic currency) of these countries: the risk and term premiums for these bonds rises as the economy worsens, exactly because there is higher risk that the Central Bank will purchase the majority of these bonds.

It seems that when one party to a debate is arguing over an expected result, and the other over tail risk, they end up talking past each other. I witnessed this over and over when I tried to convince other market actors that a leveling off of house prices might lead to unforeseen loan losses. I put Krugman's "Fire in Noah's Flood" comment in the category of "pretending tail risk doesn't exist".

David: In the situation you describe, when people started buying real assets and stopped hoarding cash, how can it be that output continues to drop and the price level goes to the moon? Are they buying foreign commodities like oil and storing them offshore? I'm not understanding the mechanics of what you're proposing.


To give you an example, the Brazilian steel industry was quite large and operated at far below capacity during the late 80's and early 90's. Real domestic steel output had cratered despite high inflation and rising (in local currency) steel prices. The financial part of the economy made most of their profits either speculating in inflation-related trades or facilitating their customer's inflation hedging activity (including capital flight). The result was an economy with rising prices, a chronic "output gap", and a parasitic but highly modernized and efficient financial system.

Yes, I know, "the U.S. is not Brazil". The point of the analogy is not to say that they are "equal", but that there is an alternative to the prevalent view that higher prices boost output, that growth causes inflation, and that deflation and inflation risk cannot coexist.

Ok, I think I see what you're getting at. Thanks.

"a parasitic but highly modernized and efficient financial system"

Well that could never happen here ... Oh wait.

It seems that the conclusion to draw from this is that we need to create a monetary system where the correct Game Theory move for commentators interested in the health of the system is to throw Dove, rather than the current one which encourages you to throw Hawk (to analogize lying and aggressing, which often seems correct to me.)

I'm not clear on why the framing matters. If you are (as either a firm or a household) choosing how to invest your money, you care about the expected path of interest rates, whether or not you perceive them as being the Fed's instrument. The Fed could be managing the monetary base, but their policy will have some implications about the behavior of interest rates, and that's what you care about. And your expectations about inflation will still make the same difference in how you interpret the path of interest rates that you expect to result from the Fed's management of the monetary base. And if you're the Fed, you will still raise interest rates (by tightening the monetary base) when you expect inflation and lower them (by loosening the monetary base) when you expect deflation. And it will still be ineffective for you to purchase T-bills when the nominal interest rate is zero. And you will still be uncomfortable acquiring unconventional assets. And if you're an economist who wants to get out of a depression, you will still want to convince firms/investors that there is going to be inflation and convince the central bank that there is not going to be inflation.

David, I'm not understanding your argument/example. What happens to real wages in the scenario you describe? If they fall, why do firms not respond to this incentive by increasing output? If they don't fall, why not? That is, why do nominal wages rise?

I can see this happening perhaps because of labor unions that insist on stable real wages, but then it's more of an insider/outsider story (and not very relevant to the US) rather than one about the government. In that case, the large output gap is an illusion because it assumes a competitive labor market, which doesn't in fact exist.


A good question as it gets to the real effects of higher inflation. Latin countries suffered from chronically low investment (and private credit) to gdp rates. There were many reasons for this (corruption, lack of property rights, etc). However, the main reason was the difficulty in estimating real returns to projects when inflation was high and variable. The large output gap was an illusion because productive capacity literally wasted away from lack of even maintenance investment. I toured the massive Cosipa steel works in Brazil just before it was modernized. The plant was nominally operating at half of capacity, but capacity could not be achieved without substantial catch-up investment.

Inflation created uncertainty in many ways. First, the Central Bank produced stop-go growth in its attempts to spark the economy and later restrain it. This "careening" around made it difficult to invest in projects with long paybacks. Second, I have tried to model real discounted cash flows in high inflation environments, and I can tell you, the range of NPV's (hyper-sensitive to the relative inflation rates of input costs, labor and product pricing) is so broad as to render a mean estimate meaningless. Third, most industrial companies in Brazil found it more profitable to engage in inflation-related speculation than in long term investments. During the 90's many companies made more money from the former than from core operations.

Median real wages fell in Latin America during much of the high-inflation period, and income inequality increased. The rich were not as affected because they could shield their wealth by hedging (capital flight); and union and government workers managed to negotiate flat real wages. The rest -- the bulk of the population -- had their incomes screwed by inflation and received nothing in return (such as more employment).

One big difference between Latin America and the U.S.: Most Latins fully expected the government to screw them with inflation, and they spent a great deal of time thinking about how to defend themselves. This led the government to create more and more inflation to produce less and less money illusion. The end result of this process was bouts of hyperinflation followed by abandonments of the currency (i.e., the peso, nuevo peso, austral, nuevo austral, etc.) and austerity, only to begin the inflation cycle anew.

Adam: "What I mean is, there will be some value for the interbank interest rate right? So, suppose the CB increases the money supply by some amount dM and promises that this monetary injection is permanent. Further assume that everyone believes this addition to the money supply is permanent. there will be some path for the short interest rate that the permanently higher money supply implies right?

Now suppose instead the CB promises exactly the path for interest rates that is implied by the permanent increase in the money supply of dM. Is that different?"

Those are good questions. Here's my attempt at an answer:

If the Bank announced (credibly) a (say) permanent increase in the money supply, that would lead to a *rise* in (at least the earlier part of) the time-path of nominal interest rates.

That's what's so weird. Looked at one way, the very same path of nominal interest rates is a loosening of monetary policy; looked at the other way it's a tightening of monetary policy (compared to the existing expected path). It's all in the expectations.

A commenter on Scott Sumner's blog had a good metaphor. Putting exactly the same set of cards on the table could be "winning" at one game and "losing" at another. It's not the physical facts; it's how we interpret them.

David: It depends on the short run AS curve. The rise in domestic prices signifies that domestic AD did increase. With a vertical AS, no effect on real output. With an upward-sloping AS real output should increase too. Yes, I am assuming that output and employment are too low in the US right now at least partly due to deficient AD. If that weren't the case, I wouldn't be tempted to cry "Fire!"

Andy: see my answer to Adam above.

It's ambiguous what the effect on nominal interest rates would be if the Fed announced an increase in the money supply. One the one hand, the quantity of money demanded at any given nominal interest rate would go down because of anticipated inflation. On the other hand, the quantity of money supplied would go up. It's not clear to me that the first effect would predominate. If prices are sufficiently sticky and there is sufficient slack in the economy, the anticipated inflation might be quite small, and if the demand for money is sufficiently elastic, the path of interest rates might go down. (US interest rates initially went down, I believe, across the maturity spectrum, when the US went off the gold standard in 1971 -- though perhaps this was because of price controls.)

But I'll grant you, you can probably come up with a not-altogether-unreasonable set of parameters for which the same path of nominal interest rates would correspond to a tightening under an interest rate targeting regime and a loosening under a money targeting regime. I'll have to think about that.

I still don't see how this would stop you from wanting to cry "Fire!" in a liquidity trap (unless you think a liquidity trap is impossible under a money growth rule, but I would take quite a bit of convincing on that point). If the market has an unlimited appetite for the money the CB is printing, the only way only way to reduce the cost of capital is by raising inflation expectations.

So are you saying the ends justify the means?

Other problems with this include:

1. You're forgetting that if Krugman and Delong were being listened to, then there would be widespread advocacy for bigger monetary and fiscal stimulus which would improve inflation expectations. So you don't need to tell them to SHUT UP to get what you want.

2. You're not in control of those who shout FIRE! They could easily change their mind, or go further and advocate mindless deflation (I've definitely seen some Austrians say we need deflation after the last century of inflation!).

3. You're not in control of the response to those shouting FIRE! so how can you be sure it will be what you want it to be?

4. Why then aren't you shouting FIRE! yourself then?

My overarching point is that if you want the outcome of the marketplace of ideas to be X, you can't sit back and pray to Flying Spaghetti Monster that X will win the debate. You have to get stuck in. And that is what Krugman and Delong are doing. Without lying as the FIRE! could be, or using very deterministic methods of understanding how social change is brought about.

"If the Bank announced (credibly) a (say) permanent increase in the money supply, that would lead to a *rise* in (at least the earlier part of) the time-path of nominal interest rates. "

Empirically is that true? The usual story is that an increase in the money supply first lowers short term rates (the liquidity effect) and only later raises it (the inflation or Fisher effect).

After all, aren't interest rate cuts implemented by increasing the money supply?

Furthermore, I don't really think there is a difference here. The rise in interest rates you are talking about doesn't actually happen if the inflation doesn't actually happen right? And if the inflation does show up then a CB that is sticking to its reaction function raises the short rate so again it looks like the outcome is identical for the path of both prices and interest rates.

Recall, as we've discussed before that you need the entire CB reaction function to pin down the equilibrium. What the CB promises to do in off equilibrium states is as important to pinning down the equilibrium paths as what it actually does and, most importantly, this fact is equally true of a CB that controls the money supply as it is of a CB that controls the short interest rate. The promise of permanence of the money injection is a particular type of reaction function that presumably has an equivelent representation with respect to interest rates.

I guess that brings me to the point, saying this:

"That's what's so weird. Looked at one way, the very same path of nominal interest rates is a loosening of monetary policy; looked at the other way it's a tightening of monetary policy (compared to the existing expected path). It's all in the expectations."

says nothing at all. And the same for the card game analogy. If you want to make a case here you need to somehow prove that there exist money supply reaction functions that can't be represented as interest rate reaction functions and lead to different equilibrium outcomes. You're nowhere near that.

Adam: Just thinking out loud:

Suppose we have a model where there are two types of central banker. Type one bankers hold the money supply constant. Type 2 bankers go Zimbabwean-wild. A new central banker has just been appointed. People don't know his type. The type is revealed when the first money supply statistics are published. What happens to nominal interest rates when the type is revealed by the money supply statistics? Probably a positive relation between M and i.

Take the same structural model, but with only one type of banker, who is conservative but a bit random, and sets Ms=Mbar+e, where e is white noise. Probably a negative relation between M and i.

One way of describing the difference between those two models is that it's not what the central bank does, but what it means. How the action is interpreted. How it is "framed"?

I don't think this example is helping you here. The difference in this example is in the reaction functions, in particular the off-equilibrium behaviour. Your example just illustrates my point about equilibrium prices and interest rates depending on the entire reaction function.

In order to make this about the framing of monetary policy it would have to be the case that I can't re-state your examnple entirely about central bankers who set i not M. I think I can:

In the first part type one bankers adjust the interest rate in such a way to keep the money supply constant, surely that's possible. Type 2 bankers hyper-inflate (by holding i too low and not following the Taylor principle). Type is revealed by choice of i. So, if i is set to low and inflation picks up then i is raised but by less than the increase of inflation (violates the Taylor principle). Since i is raised by less than the increase in inflation it must be the case that M was increased. Positive relationship between M and i.

In the second case just say the banker sets i as whatever is needed to result in Ms=Mbar+e. Then, as usual, raising requires lowering M or equivalently lowering M requires allowing i to rise. Negative relationship between M and i.

In this example what you would need to argue to show that your "social construction" argument makes any sense would be that choosing i to keep M constant is somehow impossible. This is an argument you might be able to make but I don't think you'd get a pervasive result that would hold more generally than this example and really have policy implications. But that is what you have been claiming, that the policy implications are important.


(I'm responding to 440am and sidestepping Nick's Type I/II example)

Can't Nick simply argue that even if all money supply reaction functions might be theoretically represented by equivalent interest rate reaction functions, the communication of the former might have a different effect on (messy) expectations than the latter?

A typical zero-bound suggestion is to change expectations; in particular, expectations about future Fed policy. But we don't really know much about how to reliably but prudently change expectations -- so isn't any suggestion that basically says "I have a better way of changing expectations" fair game? What if the Fed was promising to keep rates low for an extended period, but doing so in an ancient language very few people understood. Can't Nick argue that they ought to switch to English (even if both ultimately reflected the same reaction function)? Here he seems to be saying that a time path for interest rates might be much less intuitive (I agree: does the liquidity effect or the Fisher effect dominate, and over what time period, and how does that correspond to simple interest rate target/promise, what do other people think about those things?) than simply finding a credible way to convince people that M will increase permanently -- even if there is a time path for interest rates that is equivalent.

Yes, that would be a reasonable argument. I would then have three responses.

The first is that he would not have shown that the social construction is in any sense the problem here. I say that because if you work through any basic model with a fully specified intertemporal maximization problem for consumers and/or firms then you see that the reason that a permanent money injection works where a temporary one doesn't is exactly because the permanent injection raises the expected future long run price level. So, if the reason that a permanent injection works is to raise the expected long run price level, and agents understand all this, then is the "social construction" the problem or the fact that the Fed won't just set a target for future prices that is high enough.

For the "social construction" thing to really be relevant Nick would have to show somehow that people in the economy have a rather specific form of irrationality, they understand that the monetary injection needs to be permanent but somehow don't realize it means a higher future inflation or accept the higher inflation in this case but not if the policy is implemented by the interest rate?

The second response is to ask if there is any evidence that in fact this is the problem, is there evidence that the problem is that when the Fed says "low for longer" they really mean the money supply increase is permanent? I don't think so. In fact, the guys crying fire are exactly saying that if the increase in base money that has already happened were to be permanent then we'll eventually have plenty of inflation and the Fed is denying this by effectively saying we can always take it back pretty much in an instant. Given that, is it really any easier to convince people that the increase in M is permanent or the is Fed basically trying not to do that? And is there some reason to believe that if the Fed did just promise a higher future price level then that would work equally well however we frame monetary policy, path for i or path for M?

The third response is just that if you read the original "social construction" post it never seemed to me that what you're suggesting is actually what Nick was saying there. He seemed to me to be saying "you can still be expansionary at the zero bound but the Fed is completely unaware of this because they think that once the policy rate hits zero they can do nothing more". That's just a silly argument if that's what he was saying. First of all the Fed did do more, they continued to expand the balance sheet. Secondly, the Fed was clear they did not want to set a higher price level target (or apparently any at all) so the problem was not their lack of awareness of what could be done, it was their choice not to do it.

Corrction to last sentence of the second last paragraph, it should be:

"And is there some reason to believe that if the Fed did just promise a higher future price level then that would *NOT* work equally well however we frame monetary policy, path for i or path for M?"

I just reread the original "social construction" post and I think you can delete my third response, Nick is not saying that (it is something Scott Sumner used to say when I first started reading blogs about 18 months ago so that's probably how I got it in mind).

Here's my take on why this recession is not in any sense a failure of monetary policy:



I took a look at the DeLong thing, you were certainly right he was calculating the permanent drop in investment. However, long (infinite) horizon growth models with convex adjustment costs have a "turnpike" property that says investment spends basically all its time near its steady state level so convergence to steady state is very slow. This is due to the adjustment costs, which are being written with respect to the change in capital stock not investment (so dis-investing through depreciation is just as costly as littteraly dis-investing).

I need to think more about it but DeLong may be fine in his calculation.

Adam: but the smaller the adjustment costs, the quicker the actual capital stock will adjust to the desired. If we want to go from 'change in desired K' to 'change in I', we are going from a stock to a flow, and we need some sort of assumption about the adjustment costs.

Also, I really don't get how he goes from change in taxes, via change in deadweight costs, to change in desired K. I would think you would need some sort of assumption about the elasticity of change in desired K wrt change in net (after-tax) return on K. A Cobb-Douglas production function, for example, should give you that, I think.

Yes, well it's also the degree of convexity of the adjustment cost. The adjustment cost may be small for small changes but increase very, very quickly for larger changes in the capital stock. If that is the case it may well be optimal to have the actual capital stock adjust very slowly, remember it has to be balanced against the cost of having the capital stock be different from the desired long-run level.

Here investment means net investment so it just equals dK and adjustment cost is a function of dK. But he clearly didn't give any clear statement of what adjustment costs he has in mind and the turnpike thing is just a guess on my part, I'd have to go look it up to be sure what it even says exactly.

Going from change in taxes to change in desired K, I understood him to be saying it's equivalent to a productivity shock. If the government will tax an extra 1% of your return that is the same as making capital 1% less productive. Wouldn't that reduce the desired capital stock?

PS: I'm not saying I think DeLong is right here, my first thought was that you are right. I just need to think about it more, maybe even go actually look stuff up.

DeLong did reply to Nick in comments, though I don't get what he means...

If nothing else, it's not cool that DeLong didn't provide sufficient detail for someone like Adam P to check his claim. If Nick and Adam have to scratch their heads over what he might mean, a mere mortal like me is hosed.

So Brad, let's take an human and slow down neuronal function (and yet allow the rest of her physiology to survive enough to continue to support neuronal function). At what point does the person go from being "conscious" to "unconscious" because the processing takes too long?

yeah, I saw DeLong's response to Nick. He made the distinction between a no-adjustment-cost model and one with adjustment costs which is what made me think of the turnpike thing.

The thing is, I usually think of the adjustment costs as applying to gross investment not net of depreciation investment, that is if investment is zero with zero costs then the capital stock falls at the deprecitation rate. If that is the case then DeLong is wrong because by the numbers he gives in the post a fall in the capital stock of $.10 can be accomplished by depreciation in one year, so you would reduce investment by $.10 in that year as Nick suggested.

What DeLong seemed to suggest is that you incur the adjustment cost from the total change in capital stock and not just from the investment part "active" part. So in DeLong's formulation zero gross investment with depreciation, meaning a net disinvestment, incurs the adjustment cost. That's not how I usually thought about it but maybe that is how the models are usually formulated, I don't know.

dlr: you are responding to Brad's post on philosophers? Good question.

I've just posted my own back of the envelope calculation. Let's switch this discussion to there.

Whoops. Sorry Nick, the risks of multiblogging...

The comments to this entry are closed.

Search this site

  • Google

Blog powered by Typepad