Tuesday, December 27, 2016

Paul Krugman on Protectionism and Trade Balances: Wonkish and Wrong

Fighting off a Trump-induced melancholy, Krugman has posted a brief analysis of the expected effect of a potential across-the-board tariff.  After a few sensible observations about why a VAT is not protectionist, he continues with:
The starting point for a simple analysis of trade balances is the accounting identity, 
Current account + Capital account = 0 
where the current account is the trade balance broadly defined to include services and income from investments. The standard story then runs as follows: the capital account is determined by international differences in savings and investment opportunities, with capital inflows to countries that offer good returns. The real exchange rate then adjusts to ensure that the trade balance offsets these desired capital flows.
The rest of his argument can be distilled into these further claims:

1. The first-level effect is that, if the trade balance is fixed, a tariff will have to be offset by appreciation, restoring the pre-existing balance at a lower level of both imports and exports.

2. But reduced openness will make US financial assets less attractive, reducing net capital inflows.

3. This will put downward pressure on exchange rates, reducing the trade deficit and eventually moving the account into surplus.

4. But this would happen anyway without protectionism, since the attractiveness of a country’s capital assets depends on its expected future exports.  A trade deficit (insufficient export) requires an ever-weaker currency to attract investment (cheapening domestic assets), which sooner or later turns it into a surplus.

Says Paul, summing up: “.... trade deficits are always a temporary phenomenon, to be followed eventually by surpluses, and vice versa.”

5. The net effect of achieving the inevitable trade turnaround by protectionism rather than letting nature take its course is that it will require a greater depreciation to produce a given shift in flows under reduced openness.

Before jumping on this extraordinarily mistaken analysis, I should say in all fairness that PK himself is not to blame for it.  He is presenting in his usual limpid way the mainstream view of international macro, one you would find in almost any textbook.  He is transmitting error, not creating it.  Still, the error is there.

The starting point, as you would expect if you’ve read my previous posts (or my textbook, for that matter) is that the equation Krugman begins with is not an equation at all, but an identity.  It is not that some causal process (like exchange rate adjustment) makes the current account equal (with opposite signs) to the capital account; they are two ways of measuring the same thing.  Let’s put it in the terms Krugman does:

A trade deficit, assuming it is reflected as a current account deficit, is an inflow of savings.  If I buy a bottle of wine produced in the US, I spend a sum of money for it, and that same sum is received by other people in the US as income.  (I’m putting aside the possibility that some payments may leave the country, for instance if the winery is foreign-owned.)  Individually, I’m doing the spending and someone else is getting the revenue, but when you add it up at the national level, the expenditure from my purchase is exactly the same as the income from it.  Again, this is an identity, not an equality.  There is no process that causes one to equal the other; they are two different measures of the same transaction.

Now, suppose I go for a foreign wine instead.  Let’s say I buy it on a trip abroad, so I bypass domestic retailers.  Now an expenditure is recorded for “people in the US” but no corresponding income, just as income is recorded in the selling country but no corresponding expenditure.  There’s still an identity at the world level, but not nationally.  The immediate effect of my purchase in the US is that there has been an increase in spending without an increase in income: that is a reduction in savings, identically.  Abroad there is the opposite, an increase in savings, identically.  This is what the balance of payments identity, current account ≡ - capital account, is telling us.

What’s the point of harping on the identity?  First, an identity is instantaneous, true at each moment in time.  If A≡B it doesn’t mean that first A is what it is and then, some time later, B changes to match it.  Second, there is no process that causes B to equal A.  Processes could fail, after all, but identities can’t.  All the reasoning Krugman provides for how the current account should adjust to a balance on the capital account is not simply superfluous but wrong.  No process occurring over time can relate the two.

So to be specific:

Krugman’s premise is wrong.  The current account balance is not passively determined by the capital account balance, any more than vice versa.  They are two measures of the same thing.  The balance of payments position of a country is the outcome of all the forces acting on the choices people make that cause this balance to be what it is.  That includes micro-level decisions about what products, domestic or foreign, to purchase, portfolio decisions about which assets to hold, speculative pressures in foreign exchange markets—everything.  If I develop a taste for foreign wine, that has a tiny but nonzero effect on the US balance of payments, including the capital account position, just as if I become convinced that another country’s assets are a better investment than domestic ones.  Once you understand what an identity means you can’t claim one kind of factor is determinate at the BOP level, while the other is the outcome of some causal process resulting from the first.

Therefore claim (1) is false.  The trade balance is not fixed.  If protectionist policies (including likely responses by other countries) induce expenditure-shifting, that should influence the payments position.  Of course, insofar as protectionist measures may also alter desired international capital flows, the outcome may be difficult to predict in advance.

Claim (2), although Krugman doesn’t state it, is the result of a limit assumption.  Foreign capital inflows are attracted by the promise of repatriated earnings and capital gains.  A country with a current account deficit/capital account surplus cannot continue to sustain this position indefinitely; at some point depreciation must occur to reverse it, but expected future depreciation makes foreign earnings less attractive.  If you also assume rational expectations or its near-equivalent, the anticipation of BOP reversal moves its date forward to the near-present.  (The role of the limit assumption makes BOP reversal structurally the same as Ricardian Equivalence.)  Whether the global economy obeys limit-cycle dynamics in this fashion is really an empirical question, however.

Is it actually true that countries cycle back and forth between external surpluses and deficits?  The view of international political economy, which I share, is no.  There is tremendous serial correlation.  (Check it out.)  For extended periods of time, countries sort themselves into chronic surplus or deficit entities, characterized by various policies and institutions that are conducive to one or the other.  (In my macro book I loosely characterize three types of surplus countries—resource exporters, followers of the east Asian development model, and social democratic collective competitors—and four types of deficit countries—less dynamic developing countries, former members of the Soviet Bloc, English-speaking developed countries, and peripheral Europe.)  Incidentally, the factors that tend to consign countries to one side of the ledger or the other are both micro (“competitiveness”) and macro (mobilization of savings).

What enables countries to run external surpluses and deficits for extended periods of time?  We don’t have a clear answer, in my opinion.  First of all, we don’t have long enough time series to draw conclusions, given that the world economy is punctuated by cataclysms (wars) that cause immense discontinuities in capital stocks, financial obligations and trade flows.  There are also periodic crises at the national level that cause an economic reset.  Perhaps the eruption of crisis is the form that sustainability limits take in a world of limited foresight.

In any case, the textbook (wrong textbook) analysis Krugman applies to a possible Trump tariff tells us almost nothing about the effects such a policy would have on the economy.


rosserjb@jmu.edu said...

You are right, Peter, thta international econ is both micro and macro. Also it is common to see nations go for long periods of time running either chronic current account surpluses or deficits. It also happens to be the case that the current account does not always offset the capital account and vice versa, even though they are theoretically supposed to. But go look at data. In most countries they do not, although they do tend somewhat to be on opposite sides of being in surplus and deficit. There is a lot of noise in the data, a lot.

Anyway, bottom line is taht I am not sure what the outcome of a Trump tariff would be. What I do know is that foreign exchange rates are the hardest of all macro variables to actually forecast, so, who knows? Forex rates do the darnedest things.

Peter Dorman said...

Barkley, the BOP discrepancies are almost certainly measurement error, in part due to tax and regulatory avoidance. We can probably trust the COFER data on foreign exchange holdings (for reporting countries), and they don't suggest much buffering via changes in the reserve account.

And also, yes, any theory that implies a known relationship between macro variables and forex rates is *very* counter-empirical. And Krugman knows this! I don't get it.

Unknown said...

Krugman says: “The standard story then runs as follows: the capital account is determined by international differences in savings and investment opportunities, with capital inflows to countries that offer good returns. The real exchange rate then adjusts to ensure that the trade balance offsets these desired capital flows”,

I’m not at all sure how this is supposed to work when countries like China (as opposed to individuals) purchase dollars with their currencies and purchase US government and agency bonds with the dollars in order to keep their exchange rates low. There’s a risk of capital loss in terms of their currency if and when exchange rates adjust, but this is not necessarily a deterrent when the accumulated foreign assets are held by the country's central bank or government. It is the central bank or government that will take the loss rather than those who earn their incomes in the exporting industries or otherwise benefit from the lower exchange rate.

Hobson’s analysis of this kind of situation seems more reasonable to me:

"It is here enough to repeat that Free Trade can nowise guarantee the maintenance of industry, or of an industrial population upon any particular country, and there is no consideration, theoretic or practical, to prevent British capital from transferring itself to China, provided it can find there a cheaper or more efficient supply of labour, or even to prevent Chinese capital with Chinese labour from ousting British produce in neutral markets of the world. What applies to Great Britain applies equally to the other industrial nations which have driven their economic suckers into China. It is at least conceivable that China might so turn the tables upon the Western industrial nations, and, either by adopting their capital and organisers or, as is more probable, by substituting her own, might flood their markets with her cheaper manufactures, and refusing their imports in exchange might TAKE HER PAYMENTS IN LIENS UPON THEIR CAPITAL, REVERSING THE EARLIER PROCESS OF INVESTMENT UNTIL SHE GRADUALLY OBTAINED FINANCIAL CONTROL OVER HER QUONDAM PATRONS AND CIVILISERS. This is no idle speculation. If China in very truth possesses those industrial and business capacities with which she is commonly accredited, and the Western Powers are able to have their will in developing her upon Western lines, it seems extremely likely that this reaction will result." John Atkinson Hobson, Imperialism, A Study, 1902.

For a country to accumulate foreign debt as it runs a persistent trade deficit is not, in itself, a bad thing. We followed this course in the 19th century and into the 20th, but we used that debt to import capital goods and foreign technology and to invest in public education and infrastructure that led to increases in productivity. We built railroad and telegraph systems and created steel, oil, gas, electrical, automobile, and aviation industries. Our trade policies protected manufacturing as our economy grew more rapidly than our foreign debt, and as Europe squandered its resources in senseless conflicts, by the end of WW I the US had become a net creditor nation and the economic powerhouse of the world.

This is not the course we have followed over the past 40 years. We have exported rather than imported capital goods and technology, and, in return, we borrowed to import consumer goods. We invested less rather than more in our public education and infrastructure than other countries. We made huge advances in electronics and computers, but our trade policies have not protected manufacturing, and we have outsourced the manufacturing and technological components of these industries to foreign lands. As a result, our economy is not growing more rapidly than our foreign debt, and it is the United States that is squandering its resources in senseless conflicts: http://rwEconomics.com/htm/WDCh_2.htm and http://rwEconomics.com/htm/WDCh3e.htm

ProGrowthLiberal said...

Again the issue may be writing style. The (1) which draws your objection:

"The first-level effect is that, if the trade balance is fixed, a tariff will have to be offset by appreciation, restoring the pre-existing balance at a lower level of both imports and exports."

Of course no one believes the trade balance is fixed. Rather it is determined by all sorts of things such as the difference between national savings and investment. With national savings being even lower than investment now, unless we have an increase in national savings or a fall in investment (neither of which I'm advocating), then the standard model does say trade protection will lead to currency appreciation. I think this is the point Krugman (and Mankiw) are trying to make. Focusing on identities is a bit of a distraction but there is a coherent model behind all of this. Not saying the standard model is the right one but Team Trump has no model.

ProGrowthLiberal said...

To put some numbers on my comment, I went back to my Dec. 15 comment:I

"Taking a look at this source and updating what Summers said for 2015, it seems net investment was only 5.5% of NNP. With the trade deficit running at about 3.75% of GDP, net national savings is still only 1.75% of NNP."

The source is www.bea.gov. I'm hoping for an increase in investment demand. This should increase real GDP as we are not at full employment (that's my view and I'm sticking to it). If that means a larger current account deficit - so be it. What we do not need is an increase in interest rates which would not only lower investment but also lead to even more dollar appreciation and hence less net exports.

Thornton Hall said...

The issue is not writing style. That would be hilarious. The truth is, in this case, tragic and deadly.

A previous example of Krugman reasoning from an identity to an empirical conclusion is what awoke me from my dogmatic slumber.


The crucial point is this: you would be laughed at, hard, in a fourth year philosophy seminar for reasoning the way Krugman does. You would be called "stupid".

What kind of empirical discipline gives a Nobel Prize to a laughably stupid philosophy undergraduate?

"Of course no one believes..." Beware this clause. Economics makes thousands of claims about the world. When pressed to defend one, they place the clause in front of several others. When pressed to defend the several others, they place it before the one they just defended.

rosserjb@jmu.edu said...


You are probably right that most of the BOP gaps on capital and current account are stat error, although in many cases it is a simply illegal transactions. I mean, the three largest exports of Colombia are all illegal and thus contribute to a screaming gap between the measured levels of the two accounts.

Another oddity is that the accumulated capital account balance may not be reflected in current account flows. This has been the case for some time with the US, where we are on the one hand a huge net debtor on the capital account, but continue to have a surplus on foreign income flows in the current account due to our people owning assets abroad having ones that pay a lot whereas foreigners have been willing to buy very low yielding very low returns. This has been going on for quite some time now.

In his old Principles text, one of the better things that Samuelson die was at one point to have an econ history of the US from the BOP standpoint. So, prior to the Civil War we ran deficits on the current account while borrowing money from abroad (not all of which we paid back, such as to those British banks who financed the building of the Erie Canal). Then between the Civil War and WW I roughly (probably just up to before the war), we ran current account surpluses in our industrial expansion, but ran capital account deficits as we essentially paid off all those old debts. Then from WW I until somewhere in the 1960s or thereabouts, we still ran current account surpluses, but our capital account deficit was now us investing abroad.

Since then we have been running current account deficits, especially dramatically so since the mid-1980s, while borrowing from abroad. But this whole time we have managed to maintain a positive net income flow from abroad. Weird and not much noticed by most Americans. If that were to flip around, things could get unpleasant.

Peter Dorman said...

Barkley, the gap between the net foreign earnings flows and the imputed external net liability position of the US is a stock valuation problem and doesn't affect the flow identity. True, some of the dark money crowd like to apply an imputed flow of some secret investment sauce as a US export that "explains" our favorable net earnings. Of course, they have it exactly backwards: if it were not for the measured net earnings no one would be talking about our implicit export of capital allocation wisdom -- of which there is no evidence in our internal affairs, needless to say. (I think global portfolios are willing to include safe but low yield US financial assets to offset risks they incur elsewhere; we have a comparative advantage in generating those assets. And there are a lot of low yield securities held by east Asian and other CB's as post-1998 buffers.)

In any case, I'm with Keynes on this. It's a monetary economy. We want to measure flows of money and monetary commitments whatever strange items they may be transacted for.

Owen Paine said...

Pk in the service of his grand model
Has telescoped the time line

A sin of theorists
When trading regions collide
Goodly hunks of a nations raw humanity
can be devoured
before the ineluctably magnificent workings of institution embedded markets
Complete the full circle of outcomes
from deficit to surplus ..... from surplus to deficit .....