Revisiting the Euro Crisis

The euro crisis of the 2010s is well in the past now, but it remains one of the central macroeconomic events of our time.  But the nature of the crisis remains widely misunderstood, not only by the mainstream but also — and more importantly from my point of view — by economists in the heterodox Keynesian tradition. In this post, I want to lay out what I think is the right way of thinking about the crisis. I am not offering much in the way of supporting evidence. For the moment, I just want to state my views as clearly as possible. You can accept them or not, as you choose. 

During the first 15 years of the euro, a group of peripheral European countries experienced an economic boom followed by a crash, with GDP, employment and asset prices rising and then falling even more rapidly. As far as I can tell, there are four broad sets of explanations on offer for the crises in Greece, Ireland, Italy, Portugal and Spain starting in 2008. 

(While the timing is the same as the US housing bubble and crash, that doesn’t mean they are directly linked — however different they are in other respects, most of the common explanations for the European crisis I’m aware of locate its causes primarily within Europe.s)

The four common stories are:

1. External imbalances. The fixed exchange rate created by the euro, plus some mix of slow productivity growth in periphery and weak demand growth in core led to large trade imbalances within Europe. The financial expansion in the periphery was the flip side of a causally prior current account deficit.

2. Monetary policy. Both financial instability and external imbalances were result of Europe being far from an optimal currency area. Trying to carry out monetary policy for the whole euro area inevitably produced a mix of stagnation in the core and unsustainable credit expansion in the periphery, since a monetary stance that was too expansionary for Greece, Spain etc. was too tight for Germany.

3. Fiscal irresponsibility. The root of the crisis in peripheral countries was the excessive debt incurred by their own governments. The euro was a contributing factor since it led to an excessive convergence of interest rates across Europe, as markets incorrectly believed that peripheral debt was now as safe as debt of core countries.

4. Banking crises. The booms and busts in peripheral Europe were driven by rapid expansions and then contractions of credit from the domestic banking systems, with dynamics similar to that in credit booms in other times and places. The specific features of the euro system did not play any significant role in the development of the crisis, though they did importantly shape its resolution. 

In my view, the fourth story is correct, and the other three are wrong. In particular, trade imbalances within Europe played no role in the crisis. In this post, I am going to focus on why I think the external balances story is wrong, since that’s the one that people who are on my side intellectually seem most inclined toward.

As I see it, there were two distinct causal chains at work, both starting with a credit boom in the peripheral countries.

easy credit —> increased aggregate spending —> increased output and income —> increased imports —> growing trade deficit —> net financial inflows

easy credit —> rising asset prices —> bubble and/or fraud —> asset price crash —> insolvent banks —> financial crisis

That the two outcomes — external imbalances and banking crisis — went together is not a coincidence. But there is no causal link from the first to the second. Both rather are results of the same underlying cause. 

Yes, in the specific conditions of the late-2000s euro area, a credit boom led to an external deficit. But in principle it is perfectly possible to have a a credit-financed asset bubble and ensuing crisis in a country with a current account surplus, or one with current account balance, or in a closed economy. What was specific to the euro system was not the crisis itself, but the response to it. The reason the euro made the crisis worse because it prevented national governments from taking appropriate action to rescue their banking systems and stabilize demand. 

This understanding is, I think, natural if we take a “money view” of the crisis, thinking in terms of balance sheets and the relationship between income and expenditure. Here is the story I would like to tell.

Following the introduction of the euro in 1998, there were large credit expansions in a number of European countries. In Spain, for example, bank credit to the non-financial economy increased from 80 percent of GDP in 1997 to 220 percent of GDP in 2010. Banks were more willing to make loans, at lower rates, on more favorable terms, with less stringent collateral requirements and other lending standards. Borrowers were more willing to incur debt. The proximate causes of this credit boom may well have been connected to the euro in various ways. European integration offered a plausible story for why assets in Spain might be valued more highly. The ECB might have followed a less restrictive policy than independent central banks would have (or not — this is just speculation). But the euro was in no way essential to the credit boom. Similar booms have happened in many other times and places in the absence of currency unions — including, of course, in the US at roughly the same time.

In most of these countries, the bulk of the new credit went toward speculative real estate development. (In Greece there was also a big increase in public-sector borrowing, but not elsewhere.) The specifics of this lending don’t matter too much. 

Now for the key point. What happens when a a Spanish bank makes a loan? In the first step the bank creates two new assets – a deposit for the borrower, and the loan for itself. Notice that this does not require any prior “saving” by a third party. Expansion of bank credit in Spain does not require any inflow of “capital” from Germany or anywhere else.

Failure to grasp is an important source of confusion. Many people with a Keynesian background talk about endogenous money, but fail to apply it consistently. Most of us still have a commodity money or loanable-funds intuition lodged in the back of our brains, especially in international contexts. Terms like “capital flows” and “capital flight” are, in this respect, unhelpful relics of a gold standard world, and should probably be retired.

Back to the story. After the deposits are created, they are spent, i.e. transferred to someone else in return, in return for title to an asset or possession of a commodity or use of a factor of production. If the other party to this transaction is also Spanish, as would usually be the case, the deposits remain in the Spanish banking system. At the aggregate level, we see an increase in bank credit, plus an increase in asset prices and/or output, depending on what the loan finances, amplified by any ensuing wealth effect or multiplier.

To the extent that the loans finance production – of beach houses in Galicia say — they generate incomes. Some fraction of new income is spent on imported consumption goods. Probably more important, production requires imported intermediate and capital goods. By both these channels, an increase in Spanish output results in higher imports. If the credit boom leads Spain to grow faster relative to its trade partners — which it will, unless they are experiencing similar booms — then its trade balance will move toward deficit.

(That changes in trade flows are primarily a function of income growth, and not of relative prices, is an important item in the Keynesian catechism.)

Now let’s turn to the financial counterpart of this deficit. A purchase of a German good by a Spanish firm requires a bank deposit to be transferred from the Spanish firm to the German firm. Since the German firm presumably doesn’t hold deposits in a Spanish bank, we’ll see a reduction in deposits in the Spanish banking system and an equal increase in deposits in the German banking system. The Spanish banks must now replace those deposits with some other funding, which they will seek in the interbank market. So in the aggregate the trade deficit will generate an equal financial inflow — or, better said, a new external liability for the Spanish banking system. 

The critical thing to notice here is that these new financial positions are generated mechanically by the imports themselves. It is simply replacing the deposit funding the Spanish banks lost via payment for the imports. The financial inflow must take place for the purchase to happen — otherwise, literally, the importer’s check won’t clear. 

But what if there is an autonomous inflow – what if German wealth owners really want to hold more assets in Spain? Certainly that can happen. These kinds of cross-border flows may well have contributed to the credit boom in the periphery. But they have nothing to do with the trade balance. By definition, autonomous financial flows involve offsetting financial transactions, with no implications for the current account. 

Suppose you are a German pension fund that would like to lend money to a Spanish firm, to take advantage of the higher interest rates in Spain. Then you purchase, let’s say, a bond issued by Spanish construction company. That shows up as a new liability for Spain in the international investment position. But the Spanish firm now holds a deposit in a German bank, and that is an equal new asset for Spain. (If the Spanish firm transfers the deposit to a Spanish bank in return for a deposit there, as I suppose it probably would, then we get an asset for Spain in the interbank market instead.) The overall financial balance has not changed, so there is no reason for the current account to change either. Or as this recent BIS paper puts it, “the high correlations between gross capital inflows and outflows are overwhelmingly the result of double-entry bookkeeping”. (The importance of gross rather than net financial positions for crises is a pint the bIS has emphasized for many years.)

It may well happen that the effect of these offsetting financial transactions is to raise incomes in Spain (the contractor got better terms than it would at home) and/or banking-system liquidity (thanks to the fact that the Spanish banking system gets the deposits without the illiquid loan). This may well contribute to a rise in incomes in Spain and thus to a rise in the trade deficit. But this seems to me to be a second-order factor. And in any case we need to be clear about the direction of causality here — even if the financial inflows did indirectly cause the higher deficit, they did not in any sense finance it. The trade balances of Germany and Spain in no way affect the ability of German institutions to buy Spanish debt, any more than a New Yorker’s ability to buy a house in California depends on the trade balance between those states.

At this point it’s important to bring in the TARGET2 system. 

Under normal conditions, when someone wants to take a cross-border position within the euro systems the other side will be passively accommodated somewhere in the banking system. But if a net position develops for whatever reason, central banks can accommodate it via TARGET2 balances. Concretely, let’s say soon in Spain wants to make a payment to someone in Germany, as above. This normally involves the reduction of a Spanish bank’s liability to the Spanish entity and the increase in a German bank’s liability to the German entity. To balance this, the Spanish bank needs to issue some other liability (or give up an asset) while the German bank needs to acquire some asset. Normally, this happens by the Spanish bank issuing some new interbank liability (commercial paper or whatever) which ends up, perhaps via various intermediaries, as an asset for a German bank. But if foreign banks are unwilling to hold the liabilities of Spanish banks (as happened during the crisis) the Spanish bank can instead borrow from its own central bank, which in turn can create two offsetting positions through TARGET2 — a liability to the euro system, and a reserve asset (a deposit at the ECB). Conceptually, rather than the transfer of the despot being offset by a liability fro the Spanish to the German bank the interbank market, it’s now offset by a debt owed by the Spanish bank to its own national central bank, a debt between the central banks in the TARGET2 system, and a claim by the German bank against its own national central bank.

In this sense, within the euro system TARGET2 balances stand at the top of the hierarchy of money. Just as non financial actors settle their accounts by transfers of deposits at commercial banks, and banks settle their balances by transfers of deposits at the central bank, central banks settle any outstanding balances via TARGET2. It plays the same role as gold in the old gold standard system. Indeed, I sometimes think it would be better to describe the euro system as the “TARGET2 system.” 

There is however a critical difference between these balances and gold. Gold is an asset for central bank; TARGET2 balances are a liability. When a payment is made from country X to country Y in the euro area, with no offsetting private payment, the effect on central bank balance sheets is NOT a decrease in the assets of the central bank of X (and increase in the assets of the central bank of Y) but an increase in the liabilities of the central bank of X. This distinction is critical because assets are finite and can be exhausted, but new liabilities can be issued indefinitely. The automatic financing of payments imbalances through the TARGET2 system seems like an obscure technical detail but it transforms the functioning of the system. Every national central bank in the euro area is in effect in the situation of the Fed. It can never be financially constrained because all its obligations can be satisfied with its own liabilities. 

People are sometimes uncomfortable with this aspect of the euro system and suggest that there must be some limit on TARGET2 balances. But to me, this fundamentally misunderstands the nature of a single currency. What makes “the euro” a single currency is not that it has the same name, or that the bills look the same in the various countries, or even that it trades at a fixed ratio of one for one. What makes it a single currency is that a bank deposit in any euro-area country will settle a debt in any other euro-area country, at par. TARGET2 balances have to be unlimited to guarantee the this will be the case — in other words, for there to be a single currency at all.

(In this sense, we should not have been so fixed on the question of being “in” versus “out” of the euro. The relevant question is the terms on which payments can be made from one bank account another, for settlement of which obligations.)

The view of the euro crisis in which trade imbalances finance or somehow enable credit expansion is dependent on a loanable-funds perspective in which incomes are fixed, money is exogenous and saving is a binding constraint. It’s implicitly based on a model of the gold standard in which increased lending impossible without inflow of reserves — something that was not really true in practice even in the high gold standard era and isn’t true even in principle today. What’s strange is that many people who accept this view would reject those premises – if they realized they were applying them.

Meanwhile, on the domestic side, abundant credit was bidding up asset prices and encouraging investment that was, ex post, unwise (and in some case fraudulent, though I have no idea how important this was quantitatively). When asset prices collapsed and the failure of investment projects to generate the expected returns became clear, many banks faced insolvency. There was a collapse in activity in the real-estate development and construction activity that had driven the boom and, as banks tightened credit standards across the board, in other credit-dependent activity; falling asset values further reduced private spending; all these effects were amplified by the usual multiplier. The result was a steep fall in output and employment.

I don’t believe there’s any sense in which a sudden stop of cross-border lending precipitated the crisis. Rather, the “nationalization” of finance came after. Banks tried to limit their cross-border positions came only once the crisis was underway, as it became clear that there would be no systematic euro-wide response to insolvent banks, so that any rescues or bailouts would be by national governments for their own banks.

Credit-fueled asset booms and crashes have happened in many times and places. There was nothing specific to the euro system about the property booms of the 2000s. What was specific to the euro system was what happened next. Thanks to the euro, the affected governments could not respond as developed country governments have always responded to financial crises since World War II — by recapitalizing insolvent banks and shifting public budgets toward deficit until private demand recovers. 

The constraints on euro area governments were not an inevitable feature of system, in this view. Rather, they were deliberately imposed through discretionary choices by the authorities in order to use the crisis to advance a substantive political agenda.

 

New-Old Paper on the Balance of Payments

Four or five years ago, I wrote a paper arguing that the US current account deficit, far from being a cause of the crisis of 2008, was a stabilizing force in the world economy. I presented it at a conference and then set it aside. I recently reread it and I think the arguments hold up well. If anything the case that the US, as the center of the world financial system, ought to run large current account deficits indefinitely looks even stronger now, given the contrasting example of Germany’s behavior in the European system.

I’ve put the paper up as a working paper at John Jay economics department site. Here’s the abstract:

Persistent current account imbalances need not contribute to macroe- conomic instability, despite widespread claims to the contrary by both mainstream and Post Keynesian economists. On the contrary, in a world of large capital inflows, a high and stable level of world output is most likely when the countries with the least capacity to generate capital inflows normally run current account surpluses, while the countries with the greatest capacity to generate capital inflows (the US in particular) normally run current account deficits. An emphasis on varying balance of payments constraints is consistent with the larger Post Keynesian vision, which emphasizes money flows and claims are not simply passive reflections of “real” economic developments, but exercise an important influence in their own right. It is also consistent with Keynes’ own views. This perspective helps explain why the crisis of 2008 did not take the form of a fall in the dollar, and why reserve accumulation in East Asia successfully protected those countries from a repeat of the crisis of 1997. Given the weakness of the “automatic” mechanisms that are supposed to balance trade, income and financial flows, a reduction of the US current account deficit is likely to exacerbate, rather than ameliorate, global macroeconomic instability.

You can read the whole thing here.

Stimulus Around the World

Interesting new working paper out from the NBER today, on Net Fiscal Stimulus During the Great Recession. It purports to compare the level of fiscal stimulus across 28 rich and developing countries, with results that are decidely gratifying for a Keynesian.

Purports, I say, because unfortunately their chosen measure of fiscal stance makes it hard to know how seriously to take their results. They look only at final expenditures by government, ignoring both transfers and taxes. While there are certainly contexts in which this is the right approach — where the alternative would be double-counting with private expenditures — it’s not at all clear that it’s right for the questions they are trying to answer. From the stimulus side, in theory one would expect the demand effect of final government purchases to be qualitatively greater than the effect of transfers or tax cuts only if the recipients of the latter don’t face credit constraints, so that temporary changes operate only through wealth effects. And while I do think that the importance of credit constraints in the Great Recession may be overstated for businesses, they’re clearly very important for households, especially the ones most likely to receive transfers like UI. On the debt burden side, obviously deficits add the same whatever their source. On the other hand, it may well be that changes in final expenditure by government is a good proxy for the fiscal stance in general, and perhaps a better one for discretionary stimulus spending. It would be nice to see the paper redone with other measures of stimulus, but let’s tentatively accept their findings. What do they show?

First, as Krugman says, if stimulus didn’t work in the US, it’s because it wasn’t tried. The US ranks 9th from the bottom of the 28 countries in the growth of government spending, and even that is only thanks to spending in 2007-08; taking all levels of government together public consumption and investment didn’t rise at all in 2009. Of course we knew that already (And we also knew, as Aizenman and Pasricha seem not to, that the earlier increase was almost all military spending.) But it’s useful to see it in comparative perspective.

Second, the most interesting finding, that countries with the biggest increases in public spending did not see any larger increase in real interest rates on public debt, either contemporaneously or in the following year, but they did see faster growth. This means the real debt burden – measured as (r – g) * d, where r is the real interest rate on public debt, g is the real GDP growth rate, and d is the debt-to-GDP ratio – fell in those countries where public spending rose the most. If it holds up, this is obviously a very interesting result.

Finally, there’s a point they don’t make. They observe, correctly, that the US is far from any objective financial constraint on public spending. And they observe; also correctly, that the most aggressively countercyclical fiscal policy is found in middle-income countries like Korea, in sharp contrast to previous downturns, especially the late 90s. But they don’t offer any explanation for this change except a vague suggestion that countries chastened by the Asian crisis got their fiscal houses in order, leaving them plenty of space for stimulus. But that’s obviously not right. As they themselves note, there’s no correlation between the public debt burden prior to the crisis and the trajectory of government spending over the past few years. As I’ve pointed out before, what’s different in countries like Korea in the period before this crisis compared with the Asian Crisis isn’t the fiscal balance, but the external balance. They were running external deficits then, external surpluses this time. That’s what created the extra space for stimulus. (Same thing in Europe: public-sector surpluses in Spain and Ireland didn’t matter because the countries had big current account deficits. It was the corresponding private liabilities thy ended up on public balance sheets in the crisis and created the pressure for spending reductions.) Which brings me to the punchline: If the US had had a smaller trade deficit with,say, Korea in the past few years, that would have had a negligible direct effect on US demand and there’s no reason to believe that it would have created space for more expansionary fiscal policy, since we’re using nowhere near the space we have. But it very well might have forced Korea to adopt a more contractionary policy, just as other not-exorbitantly-privileged countries without external surpluses have had to. In that sense, though they certainly don’t draw this conclusion, I think this paper supports the view that global imbalances have moderated rather than exacerbated the crisis.

Does the Level of the Dollar Matter?

Mike Konczal has kindly reposted my back-of-the-envelope estimate of how much a dollar devaluation would boost US demand. (Spoiler: Not much.)

I am far from an expert on international trade and exchange rates. (Or on anything else.) Maybe some real economist will see the post and explain why it’s all wrong. But until then, I’m going to continue asking why Krugman and others who claim that exchange rates are an important cause of unemployment in this country, never provide any quantitative analysis to back that opinion up.

More abstractly, one might ask: Is the time it takes for demand to respond to changes in relative prices, minus the time it takes for exchange rate changes to move relative prices, greater than the time it takes for exchange rates changes to move relative costs (or to be reversed)? Just because freshwater economists say No for a bad reason (because relative costs adjust instantly) doesn’t mean the answer isn’t actually No for a good reason.

How Much Would a Lower Dollar Boost Demand?

Lots of economists of the liberal Keynesian persuasion (Paul Krugman, Dean Baker, Robert Blecker [1] — very smart guys all) think dollar devaluation is an important step in getting back toward full employment in the US. But have any of them backed this up with a quantitative analysis of how much a lower dollar would raise demand for American goods?

It’s not an easy question, of course, but a first cut is not that complicated. There are four variables, two each for imports and exports: How much a given change in the dollar moves prices in the destination country (the passthrough rate), and how much demand for traded goods responds to a change in price (the price elasticity.) [2] We can’t observe these relationships directly, of course, so we have to estimate them based on historical data on trade flows and exchange rates. Once we choose values of them, it’s straightforward to calculate the effect of a given exchange rate change. And the short answer to this post’s title is, Not much.

For passthrough, estimates are quite consistent that dollar changes are passed through more or less one for one to US export prices, but considerably less to US import prices. (In other words, US exporters set prices based solely on domestic costs, but exporters to the US “price to market”.) The OECD macro model uses a value of 0.33 for import passthrough at a two-year horizon; a simple OLS regression of changes in import prices on the trade-weighted exchange rate yields basically the same value. Estimates of import price elasticity are almost always less than unity. Here are a few: Kwack et al., 0.93; Crane, Crowley and Quayyum, 0.47 to 0.63; Mann and Plück, 0.28; Marquez, 0.63 to 0.92. [3] (Studies that use the real exchange rate rather than import prices almost all find import elasticities smaller than 0.25, which also supports a passthrough rate of about one-third.) So a reasonable assumption for import price elasticity would be about 0.75; there is no support for a larger value than 1.0. Estimated export elasticities vary more widely, but most fall between 0.5 and 1.0.

So let’s use values near the midpoint of the published estimates. Let’s say import passthrough of 0.33, import price elasticity of 0.75, and export passthrough and price elasticity both of 1.0. And let’s assume initial trade flows at their average levels of the 2000s — imports of 15 percent of GDP and exports at 10.5 percent of GDP. Given those assumptions, what happens if the dollar falls by 20 percent? The answer is, US net exports increase by 1.9 percent of GDP.

1.9 percent of GDP might sound like a lot (it’s about $300 billion). But keep in mind, these are long-run elasticities — in general, it takes as much as two years for price movements to have their full effect on trade. And the fall in the dollar also can’t happen overnight, at least not without severe disruptions to financial markets. So we are talking about an annual boost to demand of somewhere between 0.5 and 1.0 percent of GDP, for two to three years. And then, of course, the stimulus ends, unless the depreciation continues indefinitely. This is less than half the size of the stimulus passed last January (altho to be fair, increased demand for tradables will certainly have a higher multiplier than the tax cuts that made up a large share of the Obama stimulus.) The employment effect woul probably be of the same magnitude — a reduction of the unemployment rate by between 0.5 and 1.0 points.

I would argue this is still an overestimate, since it ignores income effects, which are much stronger determinants of trade than exchange rates are — to the extent the US grows faster and its trading partners grow more slowly as a stronger US current account, that will tend to cancel out the initial improvement. I would also argue that the gain to US employment from this sort of rebalancing would be more than offset by the loss to our trade partners, who are much more likely to face balance of payments constraints on domestic demand.

But those are second-order issues. The real question is, why aren’t the economist calling for a lower dollar providing quantitative estimates of its effects, and explicitly stating their assumptions? Because on its face, the data suggests that an overvalued dollar plays only a modest role in US unemployment.

[1] I was going to include Peter Dorman on this list but I see that while he shares the IMO misplaced concern with global imbalances, he says, “Will a coordinated dollar devaluation do the trick? Maybe, if you can get coordination (no easy feat), but it is also possible that US capacity in tradables has deteriorated too far for price adjustment alone to succeed.” Which is a more realistic view of the matter than the one Krugman seems to hold. On the other hand, Dorman was also writing just a couple years ago about The Coming Dollar Crash. That dog that didn’t bark is something I’ll hopefully be writing about in a future post.

[2] Many studies collapse passthrough and price elasticity into a single measure of real exchange rate elasticity. While this is a standard approach — about half the published papers take it — I would argue it’s not the right one for either analytic or policy purposes. Analytically, the real exchange rate elasticity doesn’t distinguish between the behavior of buyers and sellers: A low value could mean either that consumers are not responsive to price, or that sellers are holding price stable in the face of exchange rate changes. And on the other side, it’s the nominal, not real, exchange rate that’s accessible to policy. Policy-induced movements in the nominal exchange rate only translate into movements in the real rate if we assume that price levels (and real wages, if we’re deflating by labor costs) don’t respond to movements in the exchange rate, which is not generally a safe assumption.

[3] Price elasticities are all negative of course. I’m omitting the negative sign for simplicity.

Those Who Forget History, Are Probably Historians

There are hardly any economists or economic historians who have contributed more to our understanding of the role of international finance in the Great Depression than Barry Eichengreen and Peter Temin. [1] So it’s disappointing to see them so strenuously refusing to learn from that history.

They start by correctly observing that the fatal flaw of the gold standard was the “asymmetry between countries with balance-of-payments deficits and surpluses. There was a penalty for running out of reserves .. but no penalty for accumulating gold.” Thus the structural tendency toward deflation in the gold standard era, and the instability of the system once workers recognized that lower wages for “sound money” wasn’t such a great deal. If Temin and Eichengreen want to draw a parallel with the Euro system today, well, I’m not sure I agree, but it’s an avenue worth pursuing. But as they want to apply it, to the US and China, it’s unambiguously wrong, as economics and as history.

“The point,” say Temin and Eichengreen, “is not to let deficit countries off the hook.” Barry, Peter — read your books! Letting the deficit countries off the hook is exactly the point. If there’s one lesson in Lessons from the Great Depression, it’s that no practical response to the crisis was possible until the idea that a trade deficit represented a kind of moral failing was abandoned. The whole point, first, of leaving the gold standard, and later, of the Bretton Woods institutions, was to free deficit countries from the obligation to “live within their means” by curtailing domestic investment and consumption.

Keynes couldn’t have been clearer on this. The goal of postwar monetary reform, he wrote, was “A system which would maintain balance of payments equilibrium without trade discrimination but also without forcing unemployment .. on deficit countries,” [2] in other words, a system in which governments’ efforts to pursue full employment was not constrained by the balance of payments. We needn’t take Keynes as holy writ, but if we’re going to analyze current arrangements in light of his writings in the 1940s, as Temin and Eichengreen claim to, we have to be clear about what he was aiming for.

One would expect, then, that they would go on to show how “global imbalances” are constraining national efforts to pursue full employment. But they don’t even try. Instead, they offer ambiguous phrases whose vagueness is a sign, perhaps, of a bad conscience: Keynes “wanted measures to deal with chronic surplus countries.” What kind of surpluses, exactly? and deal with how?

The beginning of wisdom here is the to recognize the distinction between the balance of payments and the current account. Keynes was concerned with the former, not the latter. Keynes didn’t care if some countries ran trade surpluses or deficits, temporarily or persistently; what he cared about was that these imbalances did not interfere with other countries’ freedom “to pursue full employment and progressive social policies.” In other words, current account imbalances were not a problem as long as the financial flows to finance them were guaranteed.

“Creditor adjustment” is rightly stressed by Eichengreen and Temin as a central feature of Keynes’ vision of postwar monetary arrangements, but they seem to have forgotten what it meant. It didn’t mean no one could run a trade surplus, it just meant that the surplus countries would be obliged to lend to the deficit ones as much as it took to finance the trade imbalances. As Keynes’ follower Roy Harrod put it,”The most important requirement [is] to get the United States committed to creditor adjustment. …. Creditor adjustment could be secured most simply by an agreement that the creditor would always accept cheques from the deficit countries in full discharge of their debts. … So long as their credit position cannot cause pressure elsewhere, there is no harm in allowing a further accumulation.” All of Keynes’ proposals at Bretton Woods were oriented toward committing the countries with surpluses to lend, at concessionary rates if necessary, to the deficit ones.

China today accepts American checks in full discharge of our debts; they don’t demand payment in gold. The Chinese surplus isn’t putting upward pressure on US interest rates, or constraining public spending. All Keynes ever wanted was for all surplus countries to be like China.

“Sixty-plus years later, we seem to have forgotten Keynes’ point,” Eichengreen and Temin conclude. True that.

[1] The strangely forgotten Robert Triffin is one.

[2] The historical material in this post post, including all quotes, is drawn from chapters 6 and 9 of the third volume of Robert Skidelsky’s biography of Keynes.