Michael Pettis Savings&Loans


Foreign Saving Gluts and American Financial Imbalances

The idea that trade imbalances are more likely to be the result of credit imbalances than of savings imbalances ignores the role of savings imbalances in creating credit imbalances. When a surplus country demands to be paid for its trade surplus with claims on American assets, the U.S. economy must adjust to create these assets—and one of the most common ways it does so is by expanding credit.

On November 27, 2020, I sent out a rather long tweet in response to a very good article in the Economist called “Why It Is Misleading to Blame Financial Imbalances on a Saving Glut.” I later received a lot of positive responses asking me to expand my discussion further. Because Twitter was obviously not the right forum in which to expand what was already a long tweet, I decided to do so in this short blog post.

The Economist article illustrates just how much confusion there is over the accounting identities that describe the balance of payments. The piece starts out by reminding readers of a reference Ben Bernanke made in 2005, when he was chair of the Federal Reserve, to a “remarkable reversal in the flows of credit” to emerging economies, particularly in East Asia. These countries had begun to save more than they invested at home, morphing into a “net supplier of funds” to other parts of the world. He called this a “saving glut.”

The article then cites a number of economists, including Michael Kumhof of the Bank of England and Andrej Sokol of the European Central Bank, who were dissatisfied with the concept and called for, as the Economist put it, “a careful distinction between flows of saving and flows of finance.”

“The two are not the same,” they continued. “They need not even move together. The implication is that Mr. Bernanke may have got things the wrong way around.”

The article goes on to stress the difference between the use of the word “savings” to mean “money accumulating in a bank account” and “saving” as “the opposite of consumption.” It then notes:

By producing something that is not consumed, the economy is saving. Thus someone who spends all their earnings on home improvements is saving, however stretched they may seem, because a house is a durable asset, not a consumer trifle. Similarly a farmer who stores his harvest in a barn, rather than eating it, is saving—even if he never deposits money in a bank.


The problem, the author asks, is how savings, properly defined, can flow across borders. The article suggests that perhaps excess savings is not necessarily the right causal starting point:

Any output that is not consumed meets one of two fates: it is either invested or exported. It follows that anything that is neither consumed nor invested at home must be exported. (A farmer might, for example, export wheat to a barn overseas.) What flows across borders are the unconsumed goods and services themselves. “Other countries are not sending saving to America to give it ‘funds’ to finance their imports,” argue Mr. Kumhof and Mr. Sokol. “Their net exports are the saving, by definition.”

The rest of the article goes on to discuss the problem of conflating financial flows with capital flows, and it asserts that confusion over the two can easily lead to confusion over the meaning of a saving glut. The piece is well worth reading. It concludes:

For many people (including some economists), it is natural to think that saving must precede investment and that deposits must precede bank lending. It is therefore tempting to see saving as a source of funding and the prime mover in many macroeconomic developments. Mr. Kumhof and his co-authors see things differently, giving banks a more active, autonomous role. They give less credit to saving and more to credit.

This issue has come up repeatedly in discussions I have had about savings imbalances and in reviews of my last two books, Trade Wars Are Class Wars (with Matthew Klein) and The Great Rebalancing. Yet I think it poses a false distinction between saving and credit when it assumes that the two act separately. As we show in the book on trade wars, when a country like Japan exports its excess savings to a country like the United States, these excess savings have a financial impact on the United States even though the saving glut takes the form of an overproduction of goods. This flow of excess savings must manifest itself as a diversion of assets or a creation of assets, including most commonly an increase in credit. In such cases, it makes no sense to separate the two.

The reason such a distinction between saving and credit does not make sense here is because the U.S. economy must react to the additional goods it has imported from Japan. One way to do that would be to use the additional goods directly or indirectly to support an increase in investment, in which case the claims on this investment would either be exchanged for the goods exported by Japan or these claims would be exchanged for other claims that would be delivered to Japan. If the United States were still a developing country with substantial unmet investment needs (such as canals to be dug, railroads and airports to be built, and farmland to be cleared), this is probably what would happen. But as an advanced economy with plentiful access to cheap capital, the United States is not an economy where investment is constrained by limited access to capital.

That being the case, changes in total investment in the United States would mainly reflect business expectations of changes in demand for goods and services. The fact that the United States is importing Japanese goods, however, doesn’t increase demand for American-produced goods—if anything it reduces it—so there is likely to be no increase in investment in the United States.

There may even be a reduction in U.S. investment, in which case U.S. unemployment would rise and U.S. savings would fall, so U.S. securities that had previously been purchased by U.S. savers, such as owners of bank deposits, would now effectively be transferred to Japanese savers. Alternatively, to forestall a rise in U.S. unemployment, the Fed might reduce interest rates or otherwise encourage household borrowing to expand consumption by a wide enough margin to absorb the excess Japanese imports while still maintaining full employment at home. If that were the case, there would be an expansion in U.S. credit that arose because of the saving glut in Japan.


There are many other ways the United States could react, but the point is that credit creation at home can be and usually is one of the automatic consequences of a saving glut abroad. These are not two independent variables but simply the obverse of each other. The reason this confusion so often emerges is probably because, when discussing the sources of imbalances, it is frequently necessary to separate the bilateral capital account and bilateral current account, a bit of arithmetic that seems to suggest to some that the export of savings is different from the export of excess domestic production.

But they are not different. In fact, when a country saves more than it invests, there is no difference between its running a current account surplus and its running a capital account deficit. One cannot lead to the other because they are simply obverse sides of the same coin. In either case, because savings is the part of production that isn’t consumed, a country that saves more than it invests has insufficient domestic demand to absorb all it produces. By definition, in other words, it produces more goods and services than it can consume or invest at home, so it must export its excess savings in the form of real resources, including goods, commodities, or services.

But, of course, the country exporting these surplus savings must get paid, and it gets paid for the delivery of excess goods and services to foreigners by receiving real claims on foreign assets. The former is called the current account surplus, and the latter is called the capital account deficit. Analysts only talk about the capital account driving the current account, or vice versa, as a shorthand way of explaining the source of individual bilateral imbalances.

And this is where things get complicated. What seems to create much of the confusion is the fact that the claims on foreign assets through the capital account that a surplus country receives do not have to be from the country the surplus country is running the current account surplus with. All that must happen is that the surplus-running country must receive claims on foreign assets for its net delivery of excess goods and services to foreigners. What’s more, while all sorts of financial claims may be created to clear the transactions, they all net out to a delivery of goods and services by the surplus country against a receipt of claims on foreign assets.


I think this is where most of the confusion lies. To return to the hypothetical case above, if Japan has excess savings (meaning that Japan’s domestic savings exceed domestic investment), it can run a current account surplus with France, for example. But while Japan may immediately resolve the transaction through the creation of a series of financial claims involving France, Japan can decide to convert those claims directly or indirectly into claims on U.S. assets.

If that happens, while France runs a bilateral deficit with Japan, the French economy—by effectively having to swap claims on its own assets for claims on U.S. assets—also has to adjust by running a current account surplus with some other country that matches its deficit with Japan. In short, France has to run a current account surplus to get the claims on foreign assets that it must deliver to Japan.

For the sake of convenience, we will assume that this other country is the United States. But while that need not be the case, the current accounts of each country involved do have to keep adjusting, so the United States ultimately runs the current account deficit that corresponds to the original Japanese surplus. This is because—by giving up claims on American assets to the Japanese—the United States must run a current account deficit in which it receives goods and services from abroad—whether from France, Japan, or anywhere else.

The way these corresponding deficits and surpluses balance out can be complicated and can involve multiple countries, but ultimately all economies involved must adjust in ways that ensure that the trade ledgers are balanced globally, and these economies must also adjust to ensure that the net export or import of goods and services must balance with the net acquisition or loss of claims on foreign assets. If there is a country in which savings exceeds investment, there must be some other country in which investment exceeds savings, and that country will be the one that delivers claims on domestic assets to foreigners.

The most controversial part of the two books seems to be the implication that the United States has no control over its current account or its domestic savings rate. But this insight follows automatically if there are no constraints on the ability of foreigners to acquire U.S. assets and if foreigners do indeed choose to acquire U.S. assets as payment for their net exports of goods and services. (I discuss this topic further here and here.) If foreigners want to hold claims on American assets in exchange for the excess savings they have shipped to some other country, and if there is no constraint imposed by Washington on their ability to do so, the United States must either

  • divert existing claims to foreigners because unemployment rises,
  • create new claims because investment rises, or if investment cannot rise,
  • create new claims by increasing household or government debt to prevent unemployment from rising.

This is how a foreign saving glut can expand credit in the United States, and this is why it might seem like the U.S. current account deficit is driven by domestic credit expansion rather than by excess foreign savings. In reality, though, the excess foreign savings result in domestic credit creation. Atif Mian, Ludwig Straub, and Amir Sufi explain a very similar process, albeit one that involves a rise in domestic savings rather than foreign savings, in an important recent paper.

Note in this case that it is Japan that, for lack of a better term, is responsible for the U.S. current account deficit, even though the bilateral deficit shows up with France. That is why Matthew Klein and I employed the fiction in our book that it is the capital account that drives the current account imbalances. Technically this isn’t true—as stated earlier, the capital account is simply the obverse of the current account—but it is the Japanese decision to be paid with claims on U.S. assets that forces the United States to run the current account deficit that corresponds to Japan’s export of its excess savings, even though Japan exported its foreign savings to France.


This relationship between excess foreign savings and the U.S. current account deficit, by the way, is why Trump’s tariffs never had a chance of working. Assume that the United States imposed tariffs on French goods so as to resolve its deficit with France. As long as Japan continues to export its excess savings in the form of goods and services to France (or indeed to any other country) and then demands to be paid directly or indirectly with claims on U.S. assets, all the countries involved would have to adjust so as to let Japan run a current account surplus, force the United States to run a current account deficit, and ensure that everyone else’s trade is balanced (albeit with two or more bilateral imbalances that net out to zero). Given these circumstances, tariffs on French goods simply distort trade and raise overall costs for American consumers and French producers without in any important way affecting the overall U.S. imbalances.

If the United States does not want to be forced to absorb Japan’s domestic demand deficiency, it cannot put tariffs on French goods (or even on Japanese goods) in isolation because doing so would just cause the rest of the world to adjust in ways that shift the American bilateral imbalance with France to some other country. It turns out that the United States can only do one of two things.

  • It can prevent Japan (or other foreigners) from making net acquisitions of claims on U.S. assets. Going this route would interrupt the financial adjustments in the U.S. economy that would force the United States into running a trade deficit that corresponds with the Japanese surplus. If foreigners were not able to demand claims on American assets in exchange for their exported savings, the United States would not have to adjust domestically in the form of a trade deficit and—as is likely to be the case—an increase in outstanding credit.
  • Second, the United States could raise tariffs on all foreign imports high enough to force a downward adjustment in the savings rates of the rest of the world that corresponds with the deficiency in Japanese demand driven by excess Japanese savings. Tariffs on foreign imports affect savings rates in the same way that currency depreciation does. They raise the price of goods in the country imposing the tariffs and lower the price of goods abroad. The latter effectively transfers income from foreign manufacturers to foreign households (who consume these items) and, by raising the relative share of household income, raises the consumption share and reduces the savings share. Because tariffs on imported goods work by shifting savings rates at home and abroad, they can be much less efficient than addressing capital flows directly.

It is true, as the Economist points out, that the flow of financial savings is not the same as the flow of savings that occurs through the current account, but the concept of a saving glut nonetheless remains a very useful concept in understanding trade imbalances. As Matthew Klein and I discuss in the book Trade Wars Are Class Wars, ever since the late twentieth century, the main sources of trade imbalances in the advanced economies have been savings imbalances caused by distortions in the distribution of income in surplus countries. But while these savings imbalances are exported through the current account, the impact they have on global imbalances can best be understood by the capital account, namely the way in which current account surpluses are paid for by the acquisition of claims on foreign assets.

Aside from this blog, I write a monthly newsletter that focuses especially on global imbalances and the Chinese economy. Those who would like a subscription to the newsletter should write to me at chinfinpettis@yahoo.com, stating affiliation. My Twitter handle is @michaelxpettis.

Carnegie does not take institutional positions on public policy issues; the views represented herein are those of the author(s) and do not necessarily reflect the views of Carnegie, its staff, or its trustees.More on: 


Waleed AddasDecember 03, 20204:22 am

World Total Exports Must Be Equal to World Total Imports: it’s the bilateral imbalances that cross each other’s out but in these crossings lie the answer to reasons behind all fluctuations in exchange rates…

REPLY TO THIS POSTcstevensDecember 11, 20202:00 am

No, it is the multilateral.

Waleed AddasDecember 11, 20204:45 pm

Multilateral always balances as mentioned all world exports equal all world imports.

DamienDecember 04, 20201:18 am

So this means that the current tariffs China has imposed on Australian wine, barley, lobster etc causes a transfer of value from Australian wine producers (for instance) to me, an Australian household. I imagine this would be most likely in the form of a reduction in the price of local wine, which is the topic of about 3 articles in the local press today. As we are trying to boost demand to recover from the pandemic driven recession, this diversion of value to households is probably a “good thing”, and may explain why the government is fairly placid about the impost. Although not so good if you own shares in Treasury Wine.

REPLY TO THIS POSTYokDecember 09, 20203:54 am

Domestic Wealth and Power never sees an advantage in expanding domestic consumption. It is willing to endure whatever discount necessary, to dump its’ products abroad. Hence the sustained assault on organized labor, minimum wages, disability, social safety nets.

REPLY TO THIS POSTcstevensDecember 27, 202010:30 pm

They might not have, but it seems with the way the financial community is scrambling, the traders with a few million, who discuss in the MAcro community, that they have started smoking the gold and bitcoin joint. Fact is, as they will realize, those who never thought equities could become so untethered, bonds so low in return, that they will come to realize that key to rectification is corporate revenues under terms of increasing wages and growing domestic demand, or it ends in tears for capital.

SHDecember 14, 20205:39 am

“the main sources of trade imbalances in the advanced economies have been savings imbalances caused by distortions in the distribution of income in surplus countries.” Does it mean, the output surplus country’s ability to create surplus is because there is imbalance in its domestic factor prices, especially, price of human resources?

REPLY TO THIS POSTYokDecember 17, 20209:19 pm

Yes SH. That’s exactly what it means. The people of the surplus country are not allowed to consume all the goods and services they produce.

Moses LiaoMarch 23, 20216:59 am

Michael I truly understand what you have written. Unfortunately do you realise why Asia have such particular huge savings glut. the Asian Financial Crisis in 1997 – 98 force many poor asian countries to save and export out the savings as turning to IMF for help always mean austerity measures which Asian societies can ill afford. If back then the IMF allows the affected government to spend till recovery then pay back the loan I believe this savings glut won’t happen

REPLY TO THIS POSTcstevensMarch 25, 20214:19 pm

Michael might not respond because this is incorrect. The IMF doesn’t allow countries to do things, or force them to do things. The IMF does not have this power. It provides assistance to countries on a regular basis in monitoring their Financial System and Economies and helps them to increase their capabilities in doing this themselves. When a country has a problem, they can ask them the IMF to help. The IMF will work with the country to see what they can do. It makes a proposal and often this requires the country to make changes that will enable loans to be repaid, after all it was the inability to manage the previous debt that was the problem. So, where you are getting your perspective is likely a source that is incorrectly editorializing the issue. Indeed, in repsonse to the crisis, the IMF provided a very large support, especially large against the size of these economies and their financial systems at that time. Since then, the countries have taken to printing a lot of new money, or debt, in their local domestic currency units, where before USD debt comprised a much larger share of total debt then, than it does now. Since 1997, they have printed a lot of domestic currency debt. ” As mentioned above, the IMF intervened, providing loans to stabilize the Asian economies—also known as “tiger economies”—that were affected. Roughly $110 billion in short-term loans were advanced to Thailand, Indonesia, and South Korea to help them stabilize their economies.” Many people around the world are confused about what the IMF does, what it can and can not do. It doesn’t force countries on a regular basis to do anything, indeed it relies on cooperation with governments and assists them to increase in capacity, only provides loans with terms when a country is in a crisis. Often tells them about the possibility for crisis before hand, but it is countries that choose activities within their financial system; insofar as anyone does.

YokMarch 26, 20216:54 pm

Moses, I think your argument has merit. After all, once burned, twice shy! For Cstevens, I’ve read many sources that would completely Disagree with the characterization of the IMF as Not Coercive. Here are two books for Steve: Confession of an Economic Hitman by John Perkins and Inside the Company by Philip Agee. Two people from inside the American Empire. As recently as 2012 the IMF was involved with the economic collapse of Greece. Steve, I don’t know where the money printing comes in, but for all of us to become more affluent, someone has to print more money? After all, unless printing happened, the only way someone could do better would be at the expense of another.

cstevensApril 07, 20213:52 am

Yok The Perkins book is completely incoherent and what is purported doesn’t even reconcile to reality as there is no nefarious other able to do as imagined. I have read your comments down thru the ages and am surprised that you don’t know that, as your perspectives seemed to have been far more refined previously. It’s almost as if you are a copletely different person. And yes it has been quite common to criticize the IMF, doesn’t make the criticism, which is often political or social in nature, rather than economic or financial, and those criticizing tend not to be worthy of reading, as they don’t understand they issues, nor have the correct data. Then, like the PErkins book, any who has a shred of knowledge couldn’t take it seriously, because what he imagines able to be able to be done and by whom, isn’t able to be done, by whom. Completely only relevant to those who are ignorant. I always thought that someone as you understood that before, almost as if there is another Yok occupying an alternate universe from postings of years ago if I remember.