Paul Krugman on contractionary devaluation

This is from the 1970s, and with Lance Taylor:

The presumption that devaluation is expansionary is not supported by firm empirical evidence. Why, then, is it so widely accepted? Leftists have been known to suggest class bias — as we will argue later, devaluation does typically redistribute income from wages to profits — but this is too glib. We believe, instead, that the orthodox view of devaluation derives much of its strength from the persuasive power of the simple, elegant models in which it is presented. Since skeptics have mostly relied on Journalism or at best partial equilibrium analysis, it is not surprising that theoretical discussion is dominated by the belief that devaluation has an expansionary effect.

As just hinted, neglecting the contractionary impacts of devaluation amounts to ignoring income effects, especially those transferring real purchasing power toward economic actors with high marginal propensities to save. By redirecting income to high savers, devaluation can create an excess of saving over planned investment ex_ ante , and reductions in real output and imports ex_ post .

…Casual empiricism suggests that all three circumstances prevail in many countries, especially the less developed ones. In these
countries a deflationary impact from devaluation is more than a remote possibility; it is close to a presumption. The purpose of this paper is to show in a formal model how devaluation can cause an economic contraction. The results will come as no surprise to those concerned with policy in the underdeveloped world.

There is nothing wrong with changing your mind, as indeed I have myself on numerous issues.  The point is that most macro questions are not cut and dried, and opposing viewpoints are rarely stupid.  I also note a general tendency that, when critics attack other people, they are often attacking views they once held themselves.  I leave it to Adam Phillips and Darian Leader to tell us what that means.

The document you will find here.  For the pointer I thank Jay S.


Some grow a social conscience with time.
Plus that was before the big turning point in his life: Bush becoming president.

Bush becoming President changed the nature of devaluation?

Perhaps not, but it did change the nature of Krugman's mind.

Yes, he was an economist before he changed, no so much afterward.
As Tyler did say, it is perfectly acceptable to charge our mind over the years, but it's amazing how many of Krugman's post 2000 arguements are refuted quoting economist Krugman's pre 2000 work - the work he won the Nobel for.

err, 'change our minds'

There is always a place for the normative views of any philosopher, economist, or columnist in academia, in the blogosphere, and in the public. But he does a great disservice to the views he holds because of his style as a columnist. With him, It is difficult to keep an "ideas-only" exchange, because he pulls the author of not only contradictory view points, but also just different viewpoints, into the mud with him. Reading the comments on his blog, it is clear he is encouraging an army educated partisans to mimic this way, and I do not see how this helps the majority who are in the political middle have a framework for interweaving the variety of viewpoints into public policy which affect us all.

But given his launch into stardom as a vitrolic partisan, it seems that this is incentive enough for him to forever drop his former role as a thinker who can have a real exchange of ideas. Reminding him of his former self is fruitless.

Yes, but he is merely the latest economist to walk this path. We've seen many other scientists and none too few economists blaze this path before. Muckraking and partisan punditry is as old a profession as any.

Cherchez la femme? Or maybe it's cherchez la gloire?

You become an ever larger passive aggressive db with each passing post, TC. "I am so misunderstood" and "this guy is contradicting himself" have gotten very old.

This whole post was basically 'I know people change their views. Nonetheless here is something Krugman wrote 30-40 years ago that contradicts his blogging lately.'

Weak sauce TC. I come for teh interesting, not teh drama.

You would have a point if Krugman had not made this an issue by heaping personal abuse on people who did not change their minds when he did.

Mark, somebody needs to point out that Krugman oversells his arguments, if only so his legions of followers take one moment's pause.

Also, db? I am So Surprised you are a Krugman apologist! Remember, name calling adds to intelligent discussions ... But only when PK is doing it.

Who's apologizing for whom exactly? You are clearly apologizing for TC.

I don't like PK, don't read him, and am not apologizing for him. I am just very tired of TC's silly attack posts (this is one). It's not as bad as the faux-self-pity party that TC throws for himself in the "I'm misunderstood, no one is reading me correctly" posts, but still bad nonetheless

God forbid there should be any disagreement in life, or you will call it a silly attack post or faux-self-pity party.

This is going back an awfully long way, how old was Krugman back then? We have a lot more empirical evidence now. It's nostaligc just to compare how much the Republican party's platform has changed since 1976.

The Republican Party is not a Nobel Laureate (thank god).

No, the merely aspire to run the most powerful government and military in history.

Neither's he. Nobelish, perhaps, since it's the pretendy Swedish Central Bank Prize.

Pedantry aside, I'm sure you catch my drift...

This is a very mainstream open economy "toy" open economy macro/trade model in the MIT tradition. There are probably hundreds, if not thousands of ways a theorist can tweak these models to make devaluation deflationary, expansionary, neutral, whatever you like.

What about economies? Are there ways those could bet "tweaked"?

From what's posted, Krugman and Taylor are arguing that devaluation can have contractionary as well as expansionary effects; whether devaluation is contractionary or expansionary on net depends upon what conditions hold in the country under consideration; and present (1970s) empirical evidence supports those conditions often holding for underdeveloped countries, and thus devaluation having contractionary effects in those countries.

What exactly is Krugman saying now that is not compatible with any of that?

Quoting: "(i) When devaluation takes place with an existing trade deficit,
traded goods price increases iiranediately reduce real income at home
and increase it abroad, since foreign currency payments exceed receipts.
Within the home country the value of "foreign savings" goes
up ex ante , aggregate demand goes down ex_ post , and imports fall along
with it. The larger the initial deficit, the greater the contractionary
(ii) Even if foreign trade is initially in balance, devaluation
raises prices of traded goods relative to home goods, giving rise to
windfall profits in export and import-competing industries. If money
wages lag the price increase and if the marginal propensity to save
from profits is higher than from wages, ex ante national savings goes
up. The magnitude of the resulting contraction depends on the difference
between savings propensities of the two classes.
(iii) Finally, if there are ad valorem taxes on exports or imports,
devaluation redistributes income from the private sector to the government,
which has a saving propensity of unity in the short run. Once
again, the final outcome is reduction in aggregate demand.
Casual empiricism suggests that all three circumstances prevail
in many countries, especially the less developed ones."

Some combination of the above 3 conditions can be found to be true for most of the EU states, the PIIGS in particular.

Some combination of those conditions holding doesn't suggest anything to me about the net effect of devaluation in the PIIGS, other than that any expansionary effects will be at least partially mitigated.

But that is exactly the point, is it not?

Well, pretty much every policy in the world regarding anything will be partially mitigated by countervailing factors, so I really hope that wasn't the point.

I see two points:

1. macro is not cut and dried, therefore something about civility and moderation.

2. The other was that Krugman:

A. changed his mind
B. was attacking others because he changed his mind

(1) is uninteresting.
(2) is baseless

It would have been nice if the points had had something more to do with the content of the paper, and the content of the debates, rather than the style of the debates and mind-reading.

Very weak post, Tyler. There is no evidence whatsoever that PK has changed his mind. Where has he repudiated the theoretical arguments you quote?

These days he mostly talks about high income countries such as in the EU, where he accepts that the conditions for contractionary devaluation are less likely to hold. And the big counterexample for developing ones is China, with its export-led growth aided by undervaluation of its currency in the eyes of most. With its universally high savings rate, it also does not conform to the conditions that PK put forth for contractionary devaluation to hold.

I would have liked to see in a link for what is the contradiction.

Economics, it seems, is beginnining to look like Bible study, where you throw different passages at each other.

PK doesn't change his mind. His mind stands still while facts revolve about his head like the stars in a clear summer night sky.

The point isnt whether Krugman changed his mind or not on this. The point is that non of this is clear at all and instead of attacking other people you should be doubtful, bordering on agnostic.

Absolutely correct...thank you.

Nope. Macroeconomics, in the right hands, is SCIENCE! I think they even have white lab coats.

Isn't the primary point of devaluation to spread a negative shock broadly so that the economy can quickly move to a full employment equilibrium rather than achieving balance through years of high unemployment and disinflation/deflation?


What happens after a devaluation depends on a whole bunch of things, which is why open-economy macro is a difficult field, although in fact the original field that Krugman worked in and where he did some of his most original work, even if it is not what he was cited for in his Nobel, for which he received credit for work done earlier by others for some of it.

The quote that Tyler offers leaves out one of the most important conditions, elasticity of demand for a nation's exports, a condition known as far back as Marshall, and certainly one Krugman was and remains fully aware of. That has not been mentioned by anybody here so far, but is central to whether or not a devaluation will be contractionary or not.

Well, as always Tyler Cowen is not calling anybody names. Unfortunately! Because posts like this are his compensation for this alleged virtue: here it is a pseudo-gotcha from 30-40 years ago. But noooo, it isn't that, really - you'd NEVER guess what it really is, promise! "opposing viewpoints are rarely stupid." yes, generally speaking - sure, Krugman doesn't have anything specific in mind, today, but rather makes general comments... Uhm - - - In general this blog, while overall excellent, suffers from the disease of spending way too much time on Krugman as a target for not-so-empirical, mid-range conceptual blog posts which focus - - - wait, I know that sentence, somehow, where did I get that from??

I'm eagerly awaiting an aside how everybody misread TC, again, and didn't get the real argument beneath, that somehow cannot be expressed in plain language, never, ever...

Lars Christensen ("The Market Monetarist") responded to me when I asked him about this post:

There is a difference between having different views at different times and posting information that is mathematically incorrect or misleading:

The graphs that are pushed into the world in these posts incorrectly normalize logarithmic time series graphs to an arbitrary time and look at a long term outlook (rather than normalizing to features and looking at short term or looking at the trends). The last on the list overemphasizes levels to the point of obscuring the trends -- which are the subject of the argument.

I tend to suspect confirmation bias rather than mendacity or stupidity. The lack of a correction (and the arbitrary choice of normalization that by miracle of apparent random chance that makes your side of the argument look better) does suggest that you may be blind to your confirmation bias. In the study of these confirmation biases, it has been found that more intelligent people are worse at confirmation bias because they are clever at arguing their case in the face of conflicting evidence.

Is it better to be brilliant and suffer from confirmation bias or to be stupid and make mistakes?

** As an aside, this appears to be Krugman worrying about devaluation having some contractionary effects -- namely redistribution of wealth to the wealthy and in particular in the case of underdeveloped countries. There is no evidence either way whether Krugman has actually changed his mind! He still might think devaluation is redistributive to the wealthy and contractionary in underdeveloped countries. We are seem to be talking about Iceland and Greece, both wealthy nations.

"As just hinted, neglecting the contractionary impacts of devaluation amounts to ignoring income effects, especially those transferring real purchasing power toward economic actors with high marginal propensities to save."


Do you plan to respond to Jason's critique, in particular, regarding using log charts?

Just wondering,


He is the blog owner and can post what he wants. It's up to the readers to call him out when there are obvious errors. I did this a couple of weeks ago with one of Alex's posts where the chart really didn't make the point he was arguing.

Michelle Malkin.

Someone is in the mood to explode some heads.

How many times does one ask an obstreperous acquaintance to be polite and respectful before giving up? When we fired these kinds of clients, we called them "Not enough time in the day clients"

this is starting to look an awful lot like ankle biting. tyler, move this debate forward or be silent.

Well, there is no evidence that Krugman has changed his mind on any of this. Where Tyler might want to go, if he wishes to salvage his critique somewhat, is to focus on the original arguments themselves rather than attempting to claim that Krugman has somehow changed his mind.

So, the original argument claims that there exist models in which devaluation can be contractionary, and Krugman lists some conditions under which this might hold, also noting that these conditions may be more likely to hold in developing countries rather than developed. The argument there is clearly true. There are such models, and they are perfectly reasonable. The question then becomes how relevant are they empirically?

Here PK may have a weak spot. In the posted argument he claims that "casual empiricism" supports that argument, frequently a dangerous way to argue. Now it is certainly true that there were a number of countries in the 70s, mostly developing ones, that were engaging in frequent devaluations and were also exhibiting low growth. However, there are two problems with this. One is that if a country is growing slowly, then one would expect devaluation most of the time (unless the slow growth accompanied declining imports). Also, some of these cases involved entrenched high inflations as in parts of Latin America, in which one would expect a devaluation to exacerbate the ongoing inflation, and although Krugman did not model it, such inflations have long been argued to be damaging to growth, particularly when they are accelerating in an unpredictable way as many of them did.

So, the problem with Krugman on this matter is not some change of mind, for which there is no evidence, but that while his model was fine, his claim that there was "casual empiricism" support for it was at best very weak, although given that no nations that he was thinking of were mentioned, we cannot say that for sure.

It has been suggested to me via email that I am incorrectly assessing the degree to which Krugman's previous views are not held by him today. I think that this involves reading too much into his earlier views. They should be read very carefully.

So, I grant that we regularly see Krugman talking about how many eurozone countries would be better off if they could devalue. He does not very much talk about devaluation and developing countries, particularly the much poorer ones, so it is impossible to say that he has repudiated his views about the possibility of contractionary devaluations for those countries.

Look closely at the quotes from his earlier views. The crucial sentence, according to my email critic, is the second one in the third paragraph: "In these countries a deflationary impact from a devaluation is more than a remote possibility: it is a presumption."

Now, I think that claiming a deflationary impact was/is simply ridiculous, but claiming a contractionary one was not necessarily. "These countries" referred to those meeting the conditions for his model, which there is nothing conceptually wrong about. The problem is precisely what I state above, that he does not name any countries that do, rather he simply asserts that "many" do, particularly developing ones. And this is all based on "casual empiricism," which I have already noted is a weak basis for such claims, particularly when one is not naming any countries specifically.

So, I continue to maintain that the problem here is not that Krugman has obviously changed his mind (although maybe he has), but rather that his initial empirical claims were extremely weak.

Indeed, look at Krugman and Lance's conclusion:

"In any case, it is not the purpose of this paper to give policy advice' valid for all countries at all times. The important point is that devaluation may be deflationary, and one should be on the alert for that possibility."

What immediately came to my mind is that devaluation was used purely for "substitution" effects in developing countries. In other words, to get people who were sustenance farmers to instead become exporters.

In most of these countries, the government ran persistent deficits and there was already more than enough inflation to handle sticky wages. Instead, unemployment remained stubbornly high and real growth low due to structural factors.

In the case of driving substitution instead of pure expansion, the channel of monetary policy may matter in the short term. If the channel was to directly purchase foreign currency to devalue local currency, then the local currency will go into the hands of foreigners. The central bank action has directly raised import prices in local currency terms. It also should have lowered export prices in foreign currency terms. However, the exporters may merely raise prices for the demand instead of expanding production. In the long run, the exporters should eventually hire, but in the short-run standards of living would go down and real GDP in PPP terms would go down as well.

The issue is very different for rich countries, with well-functioning labor markets. QE does not buy foreign currency but instead exchanges money for Treasuries at market prices. The Treasury holder still has the same wealth and the holder's propensity to save and such should be the same. Even though the propensity to save is the same, the interest rates should go down with less Treasuries available. Lower interest rates then decrease savings and increase spending, the same monetary policy tool used in normal times. Aren't long-term Treasuries at historic lows though? Doesn't that mean the Fed is out of ammunition? Well, are Treasuries at zero? No, they're not? Then the Fed is not out of ammunition. The Fed can furthermore charge a penalty interest rate instead of paying 0.25% interest on excess reserves, effectively pushing interest rates negative.

Interest rates would probably actually increase after strong Fed action due to expectations of future inflation going up. But even if the market does not change expectations, such tools only go down the interest rate curve of people who already have money. It does not change their propensity to spend (i.e. shift the interest rate demand/loan supply curve) at all. That's exactly how Greenspan created the Great Moderation. So why shouldn't the tool be used?

Is anyone else noticing that Cowen's war on Krugman closely parallels the United States' undeclared, under-the-radar war on Iran?

Just sayin'...

You realize PK may order a Fatwa of death on you for this kind of comparison? Just saying.

"Opposing viewpoints are rarely stupid."

There is some truth to this. If you argue with a conspiracy theorist, it doesn't matter how much evidence exists showing the basic story of 9/11. Every piece will have some sort of argument to be explained away. With the mental acumen required to create some sort of plausible argument against the clear fact, I wouldn't call such an argument stupid in the IQ sense. Nevertheless, however possible the truthers make a vast government conspiracy sound, the probability goes down to near-zero once you add up all things that would have to fall in line for the truther's view.

Macroeconomics is less cut and dried, but reading through mainstream economics arguments linked to by Cowen and Mankiw, among others, I find a lot of good-sounding arguments instead of evidence and a lot of confirmation bias in evidence that is used. For example, on the Iceland issue, there is no skirting around the fact that Iceland has seen 6% unemployment while Ireland has 16% unemployment. Iceland did everything wrong, letting their banks default on their debts, letting their currency devaluate, and so on, but they're the ones with 6% unemployment. Were there some mitigating effects of the devaluation? Sure, but what would have happened if Iceland tried to peg the krona to the Euro? The central bank would have had to sell nearly all its assets and the high interest rates due to less currency would entice saving over spending. The government may have needed to engage in dramatic austerity as well. Where would unemployment in Iceland be then? All indications, from all evidence, is that unemployment would be well into the double digits.

So, again, why do mainstream economists have to be dragged kicking and screaming to policies which have greatly lowered unemployment? Why is there persistent favor for policies which dramatically increase human suffering on a massive scale?

Is there any human being who seriously thinks Iceland should not have defaulted on its banks' debts, or should not have devalued? It had absolutely no choice in the matter.

Looking at the factual matter, imagine Greece leaves the Euro and devalues to get to Drachma.

Are Greeks really so stupid that instead of being mad about having benefits from the state drop in nominal terms, that they will be happy when they drop in real terms instead, or perhaps stop lobbing their molotov cocktails at the police because the beer bottles have become too expensive?

Are none of Greece's state benefits linked to inflation?

My recipe for Greece is this: End all labor laws and eliminate all union contracts for five years. After that, they can figure out what labor regulations and wage rates are actually appropriate. Also sell off all State Owned Enterprises immediately. They can buy them back later if they are really worth it once the economy recovers. And if the economy doesn't recover, Greece can consider the liberal solution proven to be failed and go all socialist/communist.

Here is the last word on this topic. Look at the publication date. RL
An investment model with informational frictions and uncertainty is developed to capture the asymmetric dynamics of business cycles. When hit by a negative shock, the economy responds differently, in both size and recovery length, than when hit by a positive shock. In the model, the role for fiscal policy in smoothing the effects of business cycles fluctuations depends on initial conditions
at the time of the shock. Based on the degree of fiscal fragility of the government, expansionary fiscal policy may be expansionary or contractionary in terms of output.

Given asymmetric business cycles, I examine the effects of counter-cyclical fiscal policy that intends to smooth output fluctuations. The main policy conclusion of the paper is that implementing an expansionary fiscal policy will not always be expansionary in terms of
output—and could end up being contractionary. Rather, the effectiveness of expansionary fiscal policy depends on the degree of fiscal fragility. Specifically, for low levels of government debt to GDP, expansionary fiscal policy increases output. However, for a sufficiently high debt to GDP ratio, increasing government expenditures causes a reduction in aggregate output. It is simply a sustainability issue. A country with a low level of debt to GDP is able to borrow as much as required to finance a sharp increase in government expenditures,thus boosting aggregate demand. For countries with higher levels of debt to GDP, financing additional large increases in government expenditures is not possible since it will violate the ex ante intertemporal sustainability. In such cases, reductions in government expenditures are expansionary because they liberate resources in the credit market. The latter reduces market interest rates and increases investment. I also find that the optimal response of the government is not only to respond to lowerthan-expected aggregate demand, but also to increased uncertainty about the state of nature. Thus, the government can operate as a coordination device in the presence of information failures. Importantly, it can only do it conditional on its ex ante intertemporal sustainability. If not, its actions will only worsen the business cycle.
On asymmetric business cycles and the effectiveness of counter-cyclical fiscal policies
Nicolas E. Magud *
Department of Economics, University of Oregon, 1285 University of Oregon, Eugene, OR 97403, USA
Received 8 December 2005; accepted 23 February 2007

This post, coupled with the flip (and silly) non response to the Crooked Timber dialogue on the concept of "freedom" and the workplace, does it for me. See ya, Marginal Revolution, you are not worth my time. Those looking for good blogs should check out Timothy Burke (Easily Distracted) or Chris Blattman, or U.S. Intellectual History:


Tyler can be annoying, but Tim Burke?

"This post"? which post was that? Don't let the door hit you on your way out... As for today's blog entry in your recommended US Intelektual herstory site, whoever voted for "A People's History of the United States by Howard Zinn" as one of the worse history books should know it's rated in the "501 Must Read Books" of all time (2006, Bounty Books UK), as it was one of the stem books that tried to write about the 'commoner' in history (very trendy now, not so much back when it was published). Not that I'm defending Zinn's book, but it is a provocative read.

Noone responded to any of my reams of comments except to say that noone responded to CT.

I have no idea what you people are talking about.

I read both Krugman and Cowen's blogs regularly. Compared to Krugman I find Tyler quite the gentleman.

a) Interesting that the word "Journalism" is capitalized.

b) I'm not convinced that Paul has changed his mind about much of anything; see Barkley Rosser's comments above.

c) But let's say he did change his mind; what's wrong with that? Isn't that what any good Bayesian would do? (What's that old Keynes quote about "I change my mind when the facts change, what do you do?")

d) I know, I know, Tyler isn't saying there's anything wrong with changing his mind (Michelle Milken probably is, though), Tyler's only saying that the fact that Paul changed his mind suggests that maybe Paul should be more humble or uncertain or agnostic about his views. You know, more Tyler-ish. But at Tyler's level of fluency in Economics, a position of humility/uncertainty/agnosticism isn't necessarily any more humble or uncertain or agnostic than Paul's apparent certainty. Both Tyler and Paul require a significant degree of hubris to believe what they believe. *My* agnosticism comes from my awareness of my own ignorance, but Tyler's comes from the fact that his exceptionally good understanding of the subject leads him to believe that Paul shouldn't be so strident -- just as *Paul's* exceptionally good understanding of the subject leads him to believe whatever *he* believes about the subject.

Our Manifesto for Economic Sense has been signed by well over 8,000 people. Forward with the Intellectual Vanguard Leader PK!

Could not help thinking of the words attributed to Franklin Fisher of MIT: "Micro is all we know about; macro is all we care about." Which I as an economic historian manque have morphed into, "The past is all we know about, the future is all we care about."

Interesting. 1970's Krugman sounds an awful lot like an MMT'er, arguing that currency inflation will not be stimulative, contra Scott Sumner.

And so instead of building a powerful, unrelenting case for further fiscal easing, mainstream progressives are focused on the Fed, demanding that it do just as much to promote growth and employment as it does to promote price stability. How? By following Krugman’s advice and “credibly committing to a higher inflation target,” which, it is argued, will stimulate spending by lowering the real rate of interest. It’s a policy recommendation that only an economist (or someone with enough credit hours to be dangerous) could conjure up. I almost hope the Fed tries it so that we can banish this proposal to the wasteland of failed policy recommendations (along with QE1, QE2 and Operation Twist).

Comments for this post are closed