News Ticker

How much longer can the global trading system last? – Michael Pettis

My last blog entry inspired an old Brazilian friend of mine, with whom I hadn’t had any contact for years, to comment on this section of the interview:

It seems to me that the US is becoming increasingly isolationist, largely because it is increasingly uncertain that the benefits to the US of a US-dominated world order still exceed the costs. When the US comprised a much larger share of the “globalized” part of the world, it retained a greater share of the benefits of a stable trading environment and it cost less to maintain that environment. As the US becomes a declining share of the globalized world, the costs of imposing stability (and I have no illusions that this is done for charity) rise, and its share of the benefits decline. It is only a matter of arithmetic that at some point the costs will exceed the benefits.

My friend is a very thoughtful economist who writes often about global governance and trade, and I really enjoyed and learned from the subsequent discussion, which quickly became a three-way conversation with one of his friends. In the conversation I tried to explain why I think the break-up of the current monetary and trading regime that governs much of the world, and an American turn inward towards isolation, are very likely over the next few years, and indeed almost inevitable.

As always it helps to understand the historical context. A potted history of the current regime under which we live actually begins in the chaos of the 1920s and 1930s. The Great War, whose centennial was marked this year, had thoroughly undermined a fragile, rotted through, but still functional global economic and monetary regime with relatively clear rules. It left in its wake a system with few mechanisms with which to address the trade and capital imbalances that were all but inevitable consequence of the war.[1]

In 1944, the Allies, determined to prevent a repetition of the chaos they believed had led to war, met at Bretton Woods in New Hampshire to design and implement for the first time in history a global monetary and trade system. With a few modifications, most notably the Nixon shock of 1971, which eliminated the pretence that the global monetary system was still underpinned by gold, this is the system under which we have operated ever since.

From its inception the system was unsustainable. Backed by the unassailable power of the US, and perhaps a twinge of jealousy, Harry Dexter White, the American representative who turned out all along to have been a Soviet operative, was able to reject the alternative proposed by John Maynard Keynes in favor of his own. Keynes understood that individual countries might have great difficulty in reconciling domestic balance and external balance, and because policymakers tend to prioritize the domestic needs of the economy, the system had to be protected from the tendency for large countries to create the kinds of external imbalances that had proven so destabilizing in the 1920s and 1930s.[2]

Keynes therefore wanted mechanisms that constrained the ability of countries to create large external imbalances, whether these were in the form of large current account surpluses or of large current account deficits. White, on the other hand, believed only deficit countries needed to be disciplined, perhaps because the fact that the US had been running large surpluses for over two decades had convinced him that surpluses were an indication of moral superiority.

What kinds of imbalances are healthy?

It may make sense to stop for a moment to distinguish the advantages and disadvantages of net capital exports and imports, which of course are simply the obverse of current account surpluses and deficits. For simplicity’s sake we will assume that the world consists of two countries (relaxing the constraint does not change the analysis). If Country X begins to export large amounts of capital to Country Y on a net basis, Country X must begin to run large current account surpluses equal to the amount of its net capital exports, and Country Y will run the corresponding current account deficit. The impact of this net capital export on Country X can vary:

  1. If Country X is an advanced economy with easy access both to domestic and foreign savings, like Germany today, or if it is an undeveloped country whose policies have forced up the savings rate above domestic investment needs, like China today, the impact of capital exports is usually positive for growth and employment because they fund foreign purchases of domestically produced tradable goods. This is the case of a “healthy” current account surplus.
  1. If Country X is a developing country with insufficient domestic savings to fund domestic investment, net capital exports are probably caused either by flight capital or by the net repayment of external debt. Brazil the 1980s suffered from both, and its large current account surplus was an “unhealthy” one.

Similarly, the impact of this net capital import on Country Y can also vary:

  1. If Country Y is an advanced economy with easy access both to domestic and foreign savings, like the US, the impact of capital imports depends on whether it is “pulled in” because of a large amount of productive investment that needs financing (in which case investment in infrastructure and manufacturing would rise and real interest rates would also rise to attract foreign savings) or “pushed in” as Country X tries to balance excess production at home with insufficient domestic demand by exporting excess savings to Country Y (in which case debt-funded speculative investment and debt-funded consumption would rise, and real interest rates would drop). The former case is quite rare. Advanced economies whose capital imports are simply the obverse of capital exports elsewhere must respond with either an increase in debt or an increase in unemployment. These are “unhealthy” current account deficits.[3]
  1. If Country Y is a developing country without enough domestic savings to fund domestic investment, like the United States for much of the 19th Century, capital imports permit an increase in domestic productive investment. In that case growth will pick up and unemployment decline. Its current account deficit is “healthy”.

As should be obvious, current account surpluses or deficits are implicitly neither good nor bad. Depending on what caused them surpluses were as likely to be destabilizing to the system as deficits, and so Keynes argued that the institutions that regulated global trade and capital flows should include constraints on policies that resolved domestic imbalances (mainly high unemployment), by exporting capital. Without such constraints, it was always likely that the costs for the institution responsible for enforcing stability (and from the start White ensured that this role would be played by the United States) would at some point exceed the benefits of participating in the global trading regime. If not corrected these costs could destabilize the global trading system.

The Nixon shock

As Keynes would have predicted, destabilizing imbalances began to emerge fairly soon. In the 1950s in response to a global “dollar shortage” that had impeded the return of international trade in the late 1940s and 1950s, Germany and other countries implemented policies, including sharply undervalued currencies, aimed at acquiring dollars by running large trade surpluses. At first the US supported these policies, with the Marshall Pan contributing to the foreign accumulation of dollars from 1948 to 1952.

By the 1960s, however, there was no longer a dollar shortage. By this time, however, the policies designed to accumulate dollar had succeeded so well that the world experienced a dollar glut, but as is often the case, these policies, including undervalued currencies and export subsidies, were too firmly entrenched among powerful local manufacturers to be easily reversed. This was when a number of countries, led by France, began to game the system in a way even unanticipated by Keynes, taking advantage of the dollar overvaluation to acquire as much American gold as possible, at the price agree upon within the Bretton Woods framework.

The US response was not especially helpful. Forced to choose between unemployment (in the form of higher interest rates) and continued deficits, the US responded to German imbalances and French gold purchases in the worst way possible, by accelerating fiscal expenditures on social welfare and stepping up the war in Vietnam. They tried to reduce balance of payments pressures by imposing in 1963 a “temporary” tax (which was not withdrawn until 1974) that probably only worsened the imbalances by restricting US outflows. Given its double commitments, it was irresponsible for Washington to fund the full increase in fiscal expenditures by borrowing, but the unpopularity of the Vietnam war made Congress and presidents Johnson and Nixon reluctant to fund higher fiscal expenditures with higher taxes.

At that point the global trading and capital flows system nearly collapsed. It was clear that the cost to the United States of maintaining the regime exceeded the benefits. Rather than opt out of the system, however, the US was able to “resolve” the crisis in August 1971 by reneging on its Bretton Woods commitment to convert dollars into gold, following this up with a series of agreements in 1972 and 1973 in which Japan and Europe took steps to reduce their external imbalances by adjusting their currencies. Because there was no automatic adjustment mechanism, it was wholly up to the US to decide whether to abandon its role altogether and allow the system to collapse, or to improvise a negotiated resolution in which other counties agreed to take on part of the adjustment cost.

The “Nixon shock” in 1971, by severing the link with gold, eliminated one of the few formal constraints left within Harry Dexter White’s system, although because the gold link only existed to the extent that no one tested it, it was a fairly meaningless constraint. Since then the system has continued to function erratically and continues to be hostage to American domestic priorities. Periods of stability have been followed by periods during which one or more countries, including the US, has resolved domestic imbalances by generating large external imbalances that were themselves resolved only when they caused or threatened to cause financial crises, and then only in the painful and uncertain form of negotiated ad hoc agreements (e.g. the 1985 Plaza Accord).

These imbalances must continue to occur, and as I will show, they will become larger and increasingly difficult to resolve. It is not an accident that the deepest set of trade imbalances, followed by worst crisis since the establishment of the Bretton Woods system, was the 2007-08 crisis. Even if the US had been determined to manage its economy prudently so as to minimize the cost of its having to absorb global imbalances – and it has often chosen instead reckless monetary and fiscal policies that convert external imbalances into domestic asset bubbles – the burden placed on the United States made the global trading system unsustainable, especially as other large economies also chose reckless monetary and fiscal policies to resolve their own domestic imbalances, thereby creating or exacerbating large external imbalances.

What is the cost of stability?

How should we think about the benefits the US derives from a stable and regulated system of global trade and capital flows, the costs it must pay to maintain that system, and the stability of the relationship between the two? As I see it this question has a pretty uncomplicated calculus. Increases in global trade raises total global productivity in at least three ways.

  1. Specialization increases productivity in the ways Adam Smith described, and in many industries there are economies of scale, so presumably the larger the market, the easier it is for production to be concentrated in scale and separated into its components. Given the size of the US and European markets, there is some evidence that the marginal benefits of additional specialization are quite low but of course this depends on specific sectors of the economy.
  1. David Ricardo showed that as regions increasingly specialize in comparative advantage, total output will rise. There is a great deal of controversy over the idea that regions should specialize in comparative advantage, but the controversy tends to be about either the unequal distribution of the increase in output or about whether or not comparative advantage is static. There is little disagreement even among the strongest proponents of protection that specializing in comparative advantage increases total output.
  1. Higher levels of trade and capital flows spread the diffusion of technology and institutions more rapidly and efficiently and increase the speed of economic convergence.

It is notoriously difficult to capture the real economic value of the increase in global output created by globalization, but the tendencies that enhance the value of increased trade – specialization, comparative advantage, and the diffusion of more productive technology and institutions – are probably not linear. Increases in global trade integration, in other words, probably increase the value of the global production of goods and services at a declining rate. This suggests that the really big increases in total global output probably occur when a large economy that was previously not part of the global trading system is suddenly and quickly integrated into that system – the most obvious case is that of China in the 1990s and 2000s, and it may also be the last such case, although of course India may or may not be another such case.

Although there is no good way to determine what share of the total increase in global output the US captures (nor even a way to determine what that increase is), it is probably proportional to the US share of the relevant part of global GDP. This number has clearly been declining, both because the world has growing faster than the US and because the relevant part of the world has expanded. By the end of the 1940s and into the 1950s the US comprised, if I remember correctly, about a third of global GDP.

Because the world then really consisted of three separate groups, for my purposes the relevant US share was actually much higher. The global trading system consisted not of the entire world but rather of North America, Australia, Western Europe, and parts of Latin America, Asia and northern Africa. Communist countries were largely excluded from this system, as were countries, or parts of countries, that were too poor and too backwards to have much of an impact on global trade.

As a result the US GDP share of the relevant “global economy” was probably more than one third, and perhaps even close to one half. Whether the US captured a disproportionately large or disproportionately small share of the total benefits, and arguments can be made in either direction, it is probably safe to say that the US share of total benefits has declined since the 1940s and 1950s in line with the decline of the US share of the global trading system. As the world has grown faster than the US, and especially as countries that were once excluded fro the global trading system have joined – Russia, China, and large parts of Africa and Asia – the US share of the relevant world has declined very sharply.

But there are substantial costs to maintaining this system, and these have risen sharply. As the creator of the rules, and as by far the largest player within the system, the US is not able to game the system in the way other countries can. And other countries do often game the system – among the most obvious examples being cases that I discussed above, for example Germany and France (in two very different ways) in the 1960s, Japan in the 1980s, and China in the 2000s – not for evil intent but simply because policymakers everywhere always prioritize the resolution of domestic imbalances over external imbalances, and domestic and external balances are often difficult to manage simultaneously.

Put differently, the world economy is necessarily volatile, and to the extent that the US tries to limit destabilizing volatility, it can only do so by finding a way to absorb that volatility itself. The most obvious way the US absorbs external volatility is by absorbing trade and capital flow distortions, and the associated cost is likely to be higher to the extent that other countries try to game the system to generate more growth at home.

As I showed earlier in my potted history of the global financial system, whenever a country uses external demand to increase domestic employment and domestic production, it effectively does so by exporting capital, and in most cases the capital exports take the form of central bank purchases of foreign government bonds. Although there is a widely-held view that reserve currency status creates tremendous economic value (the “exorbitant privilege” of the US dollar), in fact most countries act as if reserve currency status conveys prestige, but at a huge cost.

Foreign capital go home!

Most countries with reserve currencies, for example, actively discourage large purchases by foreign central banks of their government bonds except in very specific cases – mainly cases in which low credibility and a declining currency creates financial or inflationary pressure. Large-scale foreign purchases of local currency assets tend to push up the value of their currency and to cause domestic demand to flow abroad. As a result, and as I showed in my example of capital flows between to countries, except for countries whose domestic investment needs are constrained by insufficient savings, or for whom foreign investment helps diffuse more efficient technology and institutions (in both cases these are mainly undeveloped countries), foreign investment is likely to lead either to higher unemployment or higher debt.

This to me is the main cost associated with enforcing a global trading and capital flow regime (I have excluded military expenses as being separate, if there are others I would be glad to hear them), and it seems to me that whether or not there are it leaves us with two relevant points here. First, there are significant costs associated with implementing and enforcing a global trade and capital flow regime. If there were not, the chaos that we saw in the 1920s and 1930s would probably have not occurred and the role of enforcer would have been voluntarily taken up, then and now, by international organizations. Second, these costs are likely to be a function of the number of players in the system and the extent to which they design policies at least in part to use external demand to generate domestic growth.

The main measure of that cost is the US current account deficit. Growth in the global economy should naturally require that the rest of the world accumulate dollar reserves, and so it is natural that the US run current account deficits as the world accumulates dollars. This permanent exchange of a small amount of dollars for real goods is the total extent of the exorbitant privilege.

But this privilege comes at an enormous risk, one which no other country wants to take. As the rest of the global trading system grows relative to the US, the need for a rising US current account deficit poses the problem identified by Robert Triffin.

The Triffin Dilemma, as this problem is known, points out that if foreign growth is high enough relative to US growth that the need for US dollar reserves grows faster than the US economy, the resulting US current account deficit will require that the US sell assets fast enough, or that US obligations to foreigners grow fast enough, eventually to put the US economy at risk. What is more, when large countries, like Japan in the 1980s or China in the 2000s, try to generate very rapid domestic growth by repressing domestic interest rates and undervaluing the currency, because of the resulting surge in their reserve accumulation, their soaring current account deficit must be balanced by a soaring US current account surplus, which exacerbates the Triffin Dilemma significantly.

This leaves us with two important points:

  1. As the US economy becomes a smaller share not so much of the global economy but of the parts of the global economy that participate in international trade, its share of the total benefits must decline, and because so many new countries and regions have joined the relevant “world” in the past three decades, the US share of total benefits has declined very rapidly. As trade impediments are further gradually reduced, the growth in benefits overall is likely to decelerate, so the US retains a declining share of a more slowly growing number.
  1. The costs it bears, however, are likely to grow inversely with its share of the relevant global economy. What is more, as more players with increasingly varied agenda join the system, the costs are unlikely to decelerate and may in fact accelerate. The costs include, but are probably not limited to, the risks identified as the Triffin Dilemma.

It seems to me then purely a matter of logic that as the world grows, as there is convergence in income disparities between rich and poor countries, and as more countries choose to join the global trading system, at some point the two lines – the higher line representing a declining share of total benefits, whose slope is likely to be slightly positive tending towards flat, and the lower line, representing rising costs, whose slope is steeply positive and becoming more so – must cross, after which point the costs exceed the benefits.

I would argue that we have probably already passed that point, and that the US would be better off today by significantly modifying the way it participates in the global trading system. The longer it waits to do so, the riskier it will be, and either the more debt or the more unemployment it will have to accept. Among other things, the US must address the role of the US dollar as the world’s reserve currency and the way this role forces the US into absorbing volatility and shortfalls in demand that originate abroad.

What is unsustainable eventually stops

Many economists may disagree with me that the costs of the current role the US plays in the global trade regime exceeds the benefits, but the point of this essay is to show that even if I am wrong, as long as the world grows faster than the US, more of the world is incorporated into the global trading system, and more countries design growth models that suppress domestic consumption in order to subsidize domestic growth, there must of necessity be a point at which it makes sense for the US to opt out of its role as shock absorber, and – by raising tariffs, intervening actively in the currency, restricting foreign purchases of US assets and especially US government bonds, or otherwise reducing capital inflows – become simply one more member of a system with no automatic adjustment process.

The current system, in other words, is inherently unstable and will sooner or later force the US economy into a position of choosing either to take on excessive risk or to abdicate its role as shock absorber. In my email exchange with my Brazilian friends, we discussed and speculated on a number of other geopolitical implications, but this is about as far as I want to go on the subject. By the way I am not making the argument, which perhaps was a little more popular a few years ago but remains popular today, of the decline of the US or of the rise of Asia. I have never really believed in either, except in the sense that in the aggregate the Asian share of the world economy is likely to rise (although not nearly as fast as some of the more intoxicated proponents of the Asian Century suggest), mainly at the “expense” of Europe and Russia, whose demographic profiles make it almost impossible for them to maintain their current share of global GDP (Japan and China too have very ugly demographic profiles that will limit the growth of their relative sizes, but of course they are part of Asia).

It seems to me however that this rise in the Asian share of global GDP will be accompanied by an even faster rise in destabilizing geopolitical tendencies within the region, so that overall the relative rise in Asian GDP will not be matched by its political rise. As most of the readers of my blog know, I expect that over the next decade we will see a number of Asian countries undergo very difficult adjustments, and I would imagine that unless handled much more carefully than it has in the past, this adjustment process is likely to increase these tensions.

But my argument does not need the 21st Century to be an American century, an Asian one, or a multi-polar one. All it requires is that the “globalized” world experience faster economic growth than the US. If this happens, to the extent that more countries with a wider range of political goals and institutions join the system created by the Bretton Woods conference, and to the extent that geopolitical tensions rise in Eurasia, it seems to me that the flat or mildly upward sloping line that represents the benefits to the US of the global trading system and the steep upward sloping line that represents the costs, if they haven’t already crossed, must cross soon. And if the recent changes in high-tech manufacturing and in the distribution of energy resources favor the US, as a lot of excited talk seems to suggest although I am an expert in neither manufacturing, high tech, nor energy), it seems to me that this would cause the flattish curve that represents the benefits to the US of stabilizing the global trading system actually to turn downwards.

If I am right, and a resurgence of some kind or other of US isolationism is simply a matter of time, the US and the world should be considering, and perhaps even already designing, the alternative sooner rather than later. Otherwise, and because this regime was created very specifically to avoid the economic chaos of the 1920s and 1930s, the reversal of this regime could very easily return us to that chaos.

Academics, journalists, and government and NGO officials who want to subscribe to my newsletter, which sometimes includes portions of this blog and sometimes (as in this case) does not, should write to me at, stating your affiliation, please.  Investors who want to buy a subscription should write to me, also at that address.

[1] I am often asked for book recommendations, and while I have read more books than I can possibly count on the economic, political and cultural history of the first three decades of the 20th Century, when it comes to political economy there are two books that, in my opinion, stand above all others. For an understanding of how the Great War undermined the global system, not just Europe and the US, the focus of most histories, but throughout the world, the best book I have read by far is The Deluge: The Great War and the Remaking of the Global Order by Adam Tooze (I also strongly recommend his book on the German economy in the 1930s, The Wages of Destruction: the Making and Breaking of the Nazi Economy). When it comes to understanding the monetary and financial forces that buffeted the US and Europe, I have long considered Barry Eichengreen’s book, Golden Fetters: the Gold Standard and the Great Depression 1919-1939, as the best book on a topic about which many great books have been written.

[2] One of the clearest descriptions of the difficulty policymakers have in managing both internal and external balance is Peter Temin and David Vines, The Leaderless Economy (Princeton University Press, 2013

[3] I explain how this works in more detail in chapters 7 and 8 of my book The Great Rebalancing, (Princeton 2013)

© Michael Pettis 2014, reproduced with permission, original article 28th September 2014,

About Andrew Palmer (275 Articles)
Book by Andrew Palmer explores today's fundamental & systemic problems of the world. Proposes a framework for understanding the forces that are driving change.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.