Sunday, June 22, 2014

Net Savings and Trade: Krugman is Simply Wrong

I was going to let Paul Krugman’s aside about the trade deficit being “determined” by net national savings pass, but Dean Baker has unintentionally prodded me into action.

It’s actually worse than what Dean says.  The sum of a country’s domestic budgets, private and public, is not “equal” to its current account (mostly trade) position; it is the current account position.  The two are the same thing.  It’s as if my team played your team and you won while we lost.  Your victory didn't “determine” our defeat or vice versa: they are one and the same.  It’s such a simple idea, but top flight economists like Krugman (who knows trade and open economy macro about as well as anyone on the planet) mess it up just as readily as my introductory students.  I've wondered why this is, but it is.  (You can read some speculation about why economists don’t distinguish between identity and equality here; see p. 169.)

Imagine if you will an economy in which people are earning income, borrowing, saving, generating income for one another by buying domestically produced goods and selling exports, and withdrawing income by buying imports.  To contribute to a trade deficit is identically to increase expenditures relative to incomes.  Neither determines the other.

It is, of course, a reasonable question to ask what causes domestic budgets and the current account to have the configuration they have.  It’s a bit like asking, did your team play great or did mine play lousy?  We don’t have a language for this sort of distinction in economics.  In the article I wrote on this topic I used the phraseology “active” and “passive” and argued that trade decisions are more likely to be active.  Actually, I’m not sure about this, either the active-passive frame or the bias toward “trade did it”.

Let me give a real world example, Germany.  Germany has a persistent current account (trade) surplus.  In fact, its economy is brilliantly organized to generate surpluses.  But it is also a country in which saving is practically a sacrament.  Germany is currently freaked out that the ECB will charge interest on reserves, which puts even more downward pressure on non-policy interest rates.  How horrible to punish virtuous savers like this!  So there you have it: a big external surplus and lots of domestic savings.  Does either “determine” the other?  No, they are identically the same thing.  If Germans ran out and spent their money on imports, they wouldn't save and they wouldn't have a trade surplus.  (If they spent their money on domestic goods only, with no ripple effects on imports, they’d have exactly that much more money so the impact on savings would cancel out.)  If they decided this summer to forego their beloved Frisian and Baltic islands and all go for vacations in the Mediterranean, their external surplus would fall, they would generate less domestic income, and their savings would be down equivalently.  The accounting identities simply reflect the fact that, when you add up the two sides of the ledger, you have two measures of the same thing.

To move the discussion back to the US, our situation can be described in terms of our awful trade performance (which Krugman acknowledges) and the willingness of borrowers to borrow and lenders to lend to finance consumption as income fails to keep up with expenditure (as it will if there is current account leakage).  All of this is part of the same, single, identical story.

You all played better and we all played worse.

UPDATE:  Why does it matter?  Is this whole business just pedantry, taking a small point, the difference between an equality and an identity, and making a big deal out of it?  No.  The argument that net national savings “determines” the trade balance has enormous practical consequences.  It means that trade policy can simply ignore issues of trade surpluses and deficits.  Trade agreements, offshoring, global regulatory arbitrage, industrial policy—all of this, by theoretical edict, is allowed to affect only the composition of trade, not its balance.  Meanwhile, for a country like the United States, with a chronic and obviously debilitating current account deficit, the remedy is said to be save, save, save.  So fiscal austerity, cuts in Social Security, defined contribution pension plans and anything else that seems to contribute to net savings rises to the top of the policy agenda.

To his credit, Krugman himself has not been a part of the savings mob, at least not since 2008.  But why give credence to the erroneous reasoning that fuels it?

UPDATE II:  And another thing: the claim that the trade balance is “determined” by net saving, which in turn is the result of other stuff that have nothing to do with trade, is the basis for the assumption of fixed balances that underlies the entire microeconomic edifice of trade theory.  All those lovely theorems simply assume that a given change in trade policy, or technology or whatever leads to perfectly offsetting changes in the value of imports and exports.  Drop the assumption—allow the trade balance to change when policy or some other factor changes—and you’re in a different world, one where the comforting strictures of comparative advantage theory no longer apply.

Mainstream trade theory has two separate steps.  First net savings determines the trade balance, then the various micro and policy forces determine trade composition.  If you recognize the identity between net savings and the current account, however, any two-step process is a fundamental conceptual error.  One thing, one step.


blissex said...

«We don’t have a language for this sort of distinction in economics. In the article I wrote on this topic I used the phraseology “active” and “passive” and argued that trade decisions are more likely to be active. Actually, I’m not sure about this, either the active-passive frame or the bias toward “trade did it”.»

and from your linked paper:

«Eventually, it dawned on me that the difference between identity and functional relations vanishes if one considers only equilibrium states, in which functional relations, by definition, hold. And this is, in fact, the methodology of nearly all modern professional Economics.»

You critique of mistaking identities as functional is very welcome, as is realuzing that is done under the influence of the theory that proves the truthyness that income distribution depends solely on productivity.

But your discussion of identity versus equality is almost as misleading for reason of a rather different nature, even if they are related;

* What is far more important in studying the political economy is the distinction between ex-ante and ex-post quantities, and the ability to estimate them. You may be hinting this when you write «limited attention economists give to mechanisms rather than end states«. Fortunately ex-ante and ex-post are well established terms.

* More formally in "ordinary" maths and theories based on it there is no difference between equality and identity because they are based on "extensional" logic. For that difference to arise one must use "modal" logic and maths based on it. In other words to introduce "time" in the discourse. When studying political economy this is done by introducing a theory of capital, because the difference between ex-ante and ex-post manifests as changes in stocks. But introducing time in Economics via a theory of capital makes it even more impossible to prove the truthyness that income distribution depends solely on productivity, and that's what motivates «limited attention economists give to mechanisms rather than end states».

* Last but not least the terms used in the national accounts use common words like investment, household, government, saving, but they are all defined in a special way that is often quite far from the meaning of the same term used not just in ordinary discourse but also in other parts of Economics, never mind of political economy.

For example insisting on the distinction between ex-ante and ex-post is a large part of understanding the original thought of Keynes, and I suspect that his sufferered evolution out of Economics into his General Theory was largely motivated by reflecting on the differences between ex-ante/ex-post.

Note: in the above I have tried to use "distinction" to indicate the distinct concepts, and "difference" to indicate the different states of the economy.

Ben Johannson said...

Krugman knows the domestic balance is the foreign balance, he just has so many conflicting and zombie ideas colliding around in his head he can't keep it straight any more. It's why I don't listen to him much.