Commenter Matthew McOsker made this attempt:

I think the definition is simple where one country consciously manages the value of their domestic country against a foreign country.

What does it mean to “manage” the value of a currency. If you mean “enact government policies that impact the exchange rate”, then all countries are guilty.  If you mean enact government policies with the intention of affecting the exchange rate, then does it matter if that’s the sole purpose, or just one of many channels?  Didn’t Bernanke mention exchange rates as one channel by which QE could affect the US economy?  Some people argue that Japan’s recent adoption of negative IOR was (among other things) aimed at depreciating the yen.  But they had no specific exchange rate peg, and the yen continues to move up and down each day.  It probably wasn’t aimed at boosting the current account balance, but rather boosting prices and GDP. Is that currency manipulation?

People are going to need to be much more specific if they intend to convince me that there is a coherent definition out there.

Due to my recent move, I’m still catching up on the last few months in the blogosphere.  (Today I spent $750 (and many hours) registering my car in California.)  Thus I finally got around to reading an earlier post by David Glasner, which responds to my complaint that “currency manipulation” is an incoherent concept.  More specifically, I made this charge:

But I would go much further; there is no intellectually respectable definition of currency manipulation.

And David responded:

Well, my only response is that I consider Max Corden to be just about the most theoretically-respectable economist alive. So let me quote at length from Corden’s essay “Macroeconomic and Industrial Policies” reprinted in his volume Protection, Growth and Trade (pp. 288-301)

There is clearly a relationship between macroeconomic policy and industrial policy on the foreign trade side. . . . The nominal exchange rate is an instrument of macroeconomic policy, while tariffs, import quotas, export subsidies and taxes and voluntary export restraints can all be regarded as instruments of industrial policy. Yet an exchange-rate change can have “industrial” effects. It therefore seems useful to clarify the relationship between exchange-rate policy and the various micro or industrial-policy instruments.

The first step is to distinguish a nominal from a real exchange-rate change and to introduce the concept of “exchange-rate protection. . . . If the exchange rate depreciates to the same extent as all costs and prices are rising (relative to costs and prices in other countries) there may be no real change at all. The nominal exchange rate is a monetary phenomenon, and it is possible that it is no more than that. A monetary authority may engineer a nominal devaluation designed to raise the domestic currency prices of exports and import-competing goods, and hence to benefit these industries. But if nominal wages quickly rise to compensate for the higher tradable-goods prices, no real effects – no rises in the absolute and relative profitability of tradable-goods industries – will remain. Monetary policy can influence the nominal-exchange rate, and possibly can even maintain it at a fixed value, but it cannot necessarily affect the real exchange rate. The real exchange rate refers to the relative price of tradable and non-tradable goods. While its absolute value is difficult to measure because of the ambiguity of the distinction between tradable and non-tradable goods, changes in it are usually – and reasonably – measured or indicated by relating changes in the nominal exchange rate to changes in some index of domestic prices or costs, or possibly to the average nominal wage level. This is sometimes called an index of competitiveness.

A nominal devaluation will devalue the real exchange rate if there is some rigidity or sluggishness either in the prices of non-tradables or in nominal wages. The nominal devaluation will then raise the prices of tradables relative to wage costs and to labour-intensive non-tradables. Thus it protects tradables. This is “exchange-rate protection”. It protects the whole group of tradables relative to non-tradables. It will tend to shift resources into tradables out of non-tradables and domestic demand in the opposite direction. If at the same time macroeconomic policy ensures a demand-supply balance for non-tradables – hence decreasing aggregate demand (absorption) in real terms appropriately – a balance of payments surplus (or at least a lesser deficit than before) will result. This refers to the balance of payments on current account since the concurrent fiscal and monetary policies can have varying effects on private capital inflow.

If the motive for the real devaluation was to protect tradables, then the current account surplus will be only a by-product, leading to more accumulation of foreign exchange reserves than the country’s monetary authority really wanted. Alternatively, if the motive for the real devaluation was to build up the foreign-exchange reserves – or to stop their decline – then the protection of tradables will be the by-product.

The main point to make is that a real exchange-rate change has effects on the relative and absolute profitability of different industries, a real devaluation favouring tradables relative to non-tradables, and a real appreciation the opposite. A nominal exchange-rate change can thus serve an industrial-policy purpose, provided it can be turned into a real exchange-rate change and that the incidental effects on the balance of payments are accepted.

This does not mean that it is an optimal form of industrial policy. . . . [P]rotection policy could be directed more precisely to the industries to be protected, avoiding the by-product effect of an undesired balance-of-payments surplus; and in any case it can be argued that defensive protection policy is unlikely to be optimal, positive adjustment policy being preferable. Nevertheless, it is not difficult to find examples of countries that have practiced exchange-rate protection, if implicitly. They have intervened in the foreign-exchange market to prevent an appreciation of the exchange rate that might otherwise have taken place – or at least, they have “leaned against the wind.” – not because they really wanted to build up foreign-exchange reserves, but because they wanted to protect their tradable-goods industries – usually mainly their export industries.

Notice that the first part of this quotation repeatedly makes the point that what matters is not the nominal exchange rate, but rather the real exchange rate.  The nominal rate matters only to the extent that it impacts the real rate.  That’s progress, as David had previously cited data on China’s nominal exchange rate (specifically the peg to the dollar) as evidence of currency manipulation. I pointed out that the real value of the Chinese yuan had been appreciating during this period.   In his reply David notes (correctly) that the real appreciation does not prove that China was not manipulating the currency.  I’m happy to accept that response, and leave China as an open question.  Perhaps China did manipulate its currency, but we’d need to go beyond the nominal exchange rate peg.

But I’m still trying to discover a definition of currency manipulation in the Corden quote above.  Is it here?

A nominal devaluation will devalue the real exchange rate if there is some rigidity or sluggishness either in the prices of non-tradables or in nominal wages. The nominal devaluation will then raise the prices of tradables relative to wage costs and to labour-intensive non-tradables. Thus it protects tradables.

I can’t figure out what that means.  Taken literally it seems to imply that a nominal appreciation depreciation that is associated with a real appreciation depreciation is a form of protectionism.  But that’s obviously nonsense.  So what is he claiming?  We know the nominal exchange rate doesn’t matter; only the real rate matters.  But currency manipulation can’t be just a depreciation in the real exchange rate, as real exchange rates move around for all sorts of reasons. If a revolution broke out in Indonesia tomorrow, I don’t doubt that the real value the their currency would plummet.  But no one would accuse Indonesia of currency manipulation.

So we need to look deeper than the nominal exchange rate, and we need to look deeper than the real exchange rate.  How about a decline in the real exchange rate caused by government policy?  Maybe, but I don’t recall anyone accusing the Norwegians of currency manipulation when they set up a sovereign wealth fund for their oil riches.  That’s a government policy that encourages national saving and hence boosts the current account.  Nor was Australia accused of currency manipulation when they did tax reform in the late 1990s.

The term ‘motive’ seems to play a role in the passage above:

If the motive for the real devaluation was to protect tradables, then the current account surplus will be only a by-product, leading to more accumulation of foreign exchange reserves than the country’s monetary authority really wanted. Alternatively, if the motive for the real devaluation was to build up the foreign-exchange reserves – or to stop their decline – then the protection of tradables will be the by-product.

As an economist, references to “motives” make me very uncomfortable.  Let’s take the example of China.  Did China’s government try to reduce the real value of the yuan because they saw what happened during the 1997 SE Asia crisis, and wanted a big war chest in case they faced a balance of payments crisis?  Or did they do the weak yuan policy to shift resources from domestic industries to tradable goods industries?  I have absolutely no idea, nor do I see why it matters.  Surely if a concept of currency manipulation has any coherent meaning, it cannot depend on the motive of the policymakers in a particular country?  We aren’t mind readers.  This is especially true if we are to believe that currency manipulation hurts other countries, as its proponent seem to suggest.  How will it be identified?

In the spirit of Bastiat, consider the following analogy.  Suppose that for years we had been buying bananas from Colombia for 10 cents a pound.  American consumers got to eat lots of cheap tasty fruit, which don’t grow well in non-tropical countries.  Then in 2018, Trump sends a team of investigators down to Colombia, and finds out that we’ve been scammed.  It’s actually not a warm country, indeed quite cool due to its high elevation.  The Colombian government had spent millions building giant greenhouses to grow bananas.  We’ve been tricked into buying all these cheap bananas from Colombia, which artificially created a “competitive advantage” in the banana industry through subsidies.

Here’s my question:  Why does it matter why the Colombian bananas were cheap?  If we benefited from buying the bananas at 10 cents a pound, why would we care if the price reflected true competitive advantage or government subsidy?  Does the US benefit from buying 10-cent bananas, or not?

But that’s not all.  Even if you convinced me that we should worry about interventionist policies in our trading partners, I’d still want a definition of currency manipulation.  There are a billion ways that a foreign government could influence a real exchange rate.  Which ones are “manipulation”? It’s meaningless to talk about China depreciating its currency, without explaining HOW.  A currency is just a price, and reasoning from a currency change (real or nominal) is simply reasoning from a price change.  Which specific actions constitute currency manipulation?  I don’t want motives, I need verifiable actions.  And does this concept have to involve a current account surplus?  Australia’s been running CA deficits for as long as I can remember.  Suppose the Aussie government did enough “currency manipulation” to reduce their trend CA deficit from 4% of GDP to 2% of GDP.  But it was still a deficit.  Would that be “manipulation”.  Why or why not?

Should we care why a country has a big CA surplus? Suppose Switzerland has a big CA surplus due to high private saving rates, Singapore has a big CA surplus due to high public saving in common stocks, and China has a CA surplus due to high public saving in foreign exchange.  What difference does it make?  (And I haven’t even addressed Ricardian equivalence, which further clouds these distinctions.)

We know that the only way that governments can affect the real exchange rate is by enacting policies that impact national saving or national investment.  But almost all policies impact either national saving or national investment.  So which of those count as manipulation?  Is it merely policies that lead to the accumulation of foreign exchange?  If so, then won’t you simply encourage countries to use some other technique for boosting national saving? An alternative policy that avoids having them be labeled currency manipulators?

And why is this called “creating a competitive advantage”.  It doesn’t give an overall economy a competitive advantage, just one sector—exports. Other sectors are put in an equal and opposite disadvantage.  So why not call these currency manipulation policies “competitive disadvantage”.  Or conversely, why not label agricultural subsidies a form of “competitive advantage”.  After all, just as high saving policies boost the export sector, agricultural subsidies boost the agricultural sector.  Or consider America’s low saving fiscal policy, which boosts our service sector.  Do those policies also give America a competitive advantage?  After all, services are far bigger than exports.

When you read people use the term “competitive advantage” it’s hard to avoid the inference that they are claiming that these policies will somehow boost aggregate demand.  Paul Krugman rightly mocked those arguments back in the 1990s, pointing out that monetary policy determines AD.  So then is “currency manipulation” just a special theory that only applies at the zero bound?  It’s hard to tell, the descriptions of this concept are all so vague.

I’d still like a very short and highly precise definition of currency manipulation.  Just a couple sentences.  What data points do I look at to determine if a currency is being manipulated?  If it’s the nominal exchange rate then I disagree with the definition.  If it’s the real exchange rate then I disagree with the definition.  If it’s the current account balance then I disagree with the definition.  If it’s the accumulation of foreign exchange then I disagree with the definition.  A definition involving any of those criteria would be so flawed as to be meaningless.  So what is the definition?

PS.  David also asks the following question:

Scott often cites sticky prices as an important assumption of macroeconomics, so I don’t understand why he thinks that the nominal exchange rate has no effect on trade. If prices do not all instantaneously adjust to a change in the nominal exchange rate, changes in nominal exchange rates are also changes in real exchange rates until prices adjust fully to the new exchange rate.

Talking about the effect of a price on quantities is reasoning from a price change, and hence wrong.  If the currency depreciation is caused by monetary stimulus, then I agree that output is likely to rise in the short run, due to sticky wages.  But that’s equally true in a closed economy.  As far as the trade balance, it depends on the relative size of the income and substitution effects.  The massive dollar devaluation of 1933 had almost no effect on the trade balance, as the income and substitution effects were roughly equal size, but pushed in opposite directions.  So was that currency manipulation?  Did FDR’s motives matter?  I still don’t understand why any of this matters.

PPS.  David’s post also refers to “undervalued currencies”, but this term is just as vague as currency manipulation or competitive advantage.  What makes a currency undervalued?

Ben Cole directed me to this interesting story:

The National Association of Realtors released a report Tuesday that said foreign buyers and recent immigrants spent an estimated $153 billion on American properties in the year ending March 2017. That was a 49% increase over the previous year and the highest level since record-keeping began in 2009.

The purchases accounted for 10% of the total value of existing home sales in the U.S. The report did not include new homes.

The breakdown of sales between foreigners and recent immigrants was about 50:50.

[I wish they had data on new homes, as that’s the sort of home that foreigners tend to prefer.]

Of course the sale of homes to immigrants is not an export, but it does have a similar economic impact.  However the sale of homes to foreigners does represent a US export, and creates lots of goods jobs for American blue collar workers.  (Note that it doesn’t really matter whether they buy new or existing homes; the net effect on the housing market is the same.)  So the protectionists should be rejoicing, right?

Actually, just the opposite.  The US government does not even count these as exports.  Instead they are treated the same as net borrowing.  They are considered a part of America’s current account deficit, leading to all sorts of silly hand-wringing about how America is borrowing too much and living beyond our means.  In fact, we do borrow too much (due to the tax advantage of doing so), but that has nothing to do with the current account deficit.

I have a solution.  Treat international trade the way that we treat trade between American states.  Stop collecting records on imports and exports.  We don’t have data on the CA deficit of Texas or the CA surplus of Massachusetts, and that lack of data doesn’t seem to cause any problems. So stop doing so for the US as a whole.

You can still collect data on America’s net debt position (good luck with that!), if you wish to.

PS.  I have a post on “The German Problem” over at Econlog.

The back of the Economist magazine has current account deficits over the past 12 months for 42 major economies, comprising the vast majority of global GDP.  I noticed that they group neatly into 4 major blocks.  But first I’ll present the data for 5 regions:

1. Continental Europe (excluding Turkey):  $556.8 billion CA surplus

2.  East Asia:  $674.0 billion surplus

3.  English speaking countries (US/UK/Canada/ Australia)  $698.9 billion deficit

4.  Latin America:  $101.5 billion deficit

5.  Middles East/South Asia:   $116.0 billion deficit

Exceptions include Greece and Poland (which both have tiny deficits), Israel (which has a sizable surplus), and Indonesia and Philippines (which have tiny deficit). That’s it.

In theory, the numbers should add up to zero.  I presume the net global surplus reflects the fact that many of the smaller developing countries left off the list have CA deficits, and perhaps there is also a statistical discrepancy (measuring error.)

Let’s say I’m right that most of the missing countries are developing countries with deficits, and let’s mentally add them in with both Latin America and the Middle East/South Asia. Let’s call that big group “developing countries”, even though it technically excludes developing East Asia.  In that case the world has four blocks. Both English-speaking countries and developing countries have big CA deficits, and both continental Europe and East Asia have big surpluses.

So what’s going on here?  Basically, Europe and East Asia are investing lots of money in developing countries and English-speaking countries.  But why?

Developing countries have better growth prospects—thus India grows much faster than Switzerland and Singapore (two high surplus countries).  The English speaking world runs big deficits for several reasons:

1.  The Anglo-Saxon economic model is a good one, drawing in foreign capital.

2.  The Anglo-Saxon world is a safe place for investment, due to the legal system.

3.  The Anglo-Saxon world is a good place to move, as English is the global second language.  Thus if you are a Chinese person worried about the future of your country, Vancouver makes more sense than Vienna as a place to move if things go bad.  So you buy a house in Vancouver as a security blanket.

4.  The Anglo-Saxon world dominates high tech, which brings in lots of money that doesn’t show up in CA deficit data.

What am I missing?

I expect these “imbalances” to persist for many decades.  That’s because they are not “imbalances” at all, but rather equilibrium outcomes.  The world is basically Europe/East Asia on one side and English/developing on the other.  Why should that change in the future?

A number of people have asked me to comment on a new paper by Jonathan Rothwell, which criticizes a study by Autor, Dorn and Hanson (ADH) on the impact of Chinese imports on the US job market.

I conclude that the economic losses from trade are not as severe as the economics literature currently implies. Workers in the most import-exposed sectors face a risk of layoff and unemployment that is comparable to workers in other sectors, where competition comes almost exclusively from domestic businesses. While it is likely that less import competition would further lower the risk of displacement and boost wages for manufacturing workers, less competition would likely lead to a reduction in the ratio of product quality to price and a drop in consumer welfare. I accept the Autor et al. (2014) finding that import competition lowers wages for U.S. workers in the affected industries — but even still, I find that workers in the manufacturing sector continue to earn a sizable wage premium compared to those with similar experience and education levels in other sectors.

At the community level, these results should not be taken to mean that de-industrialization has been harmless to individuals or even communities. Rather, the results imply that deindustrialization as a result of Chinese import competition plays out no differently than deindustrialization as a result of other forces — such as domestic competition or technological change. Communities relying more heavily on industries facing import competition perform no  worse in this study on summary measures of economic development and consistently show higher growth rates in establishments. They seem to find ways to adapt, maintain wage growth and launch new enterprises.

That’s what I would have expected.  Not surprisingly, Autor, Dorn and Hanson contest Rothwell’s study.

Regardless of whether Rothwell is right or wrong, the press has done an extremely poor job in reporting the ADH study.  Trade economists already knew that specific industries, and even communities, can be hurt by import competition.  The press has suggested that the ADH study shows that China trade resulted in a net job loss to the US, a finding that really would be new.  But as Paul Krugman and I keep pointing out that’s just not so.  Their study is completely consistent with zero net job loss to the US.  That’s because the study looked at the period of 1990-2007, when monetary offset was fully engaged.  So there’s no plausible AD channel.  Of course you can make other arguments, but you can’t show aggregate effects with a cross-sectional study.

All the press coverage of ADH is much ado about nothing.  Maybe China did hurt the overall US labor market, but their study doesn’t show it.  I’m not surprised that the press ignores me, but I am a bit surprised they ignore Krugman, particularly since he has occasionally argued that China was stealing US jobs during the Great Recession.  No one can claim his critique of ADH was based on ideological bias.

PS.  Nor can Autor, Dorn and Hanson be accused of ideological bias.  For instance, they favor TPP.

PPS.  Before you try defending ADH based on non-AD channel arguments, you might consider that at various times in their paper they imply they do have an AD channel in mind.  For instance, when contrasting Germany’s trade surplus with the US trade deficit.