Chairman Ben S. Bernanke
At the Bundesbank Lecture, Berlin, Germany
September 11, 2007
Global Imbalances: Recent Developments and Prospects
In a speech given in March 2005 (Bernanke, 2005), I discussed anumber of important and interrelated developments in the globaleconomy, including the substantial expansion of the current accountdeficit in the United States, the equally impressive rise in thecurrent account surpluses of many emerging-market economies, and aworldwide decline in long-term real interest rates. I argued thatthese developments could be explained, in part, by the emergence of a global saving glut,driven by the transformation of many emerging-marketeconomies--notably, rapidly growing East Asian economies andoil-producing countries--from net borrowers to large net lenders oninternational capital markets. Today I will review those developmentsand provide an update. I will also consider policy implications andprospects for the future.
A principal theme of my earlier remarks was that a satisfyingexplanation of the developments in the U.S. current account cannotfocus on developments within the United States alone. Rather,understanding these developments and evaluating potential policyresponses require a global perspective. I will continue to take thatperspective in my remarks today and will emphasize in particular howchanges in desired saving and investment in any given region, throughtheir effects on global capital flows, may affect saving, investment,and the external balances of other countries around the world.
The Origins of the Global Saving Glut, 1996-2004
I willbegin by reviewing the origins and development of the global savingglut over the period 1996-2004, as discussed in my earlier speech, andwill then turn to more-recent developments.
As is well known, the U.S. current account deficit expanded sharplyin the latter part of the 1990s and the first half of the presentdecade. In 1996, the U.S. deficit was $125 billion, or 1.6 percent ofU.S. gross domestic product (GDP); by 2004, it had grown to $640billion, or 5.5 percent of GDP.1 National income accounting identities imply that the current accountdeficit equals the excess of domestic investment in capital goods,including housing, over domestic saving, including the saving ofhouseholds, firms, and governments. The proximate cause of theincrease in the U.S. external deficit was a decline in U.S. saving;between 1996 and 2004, the investment rate in the United Statesremained almost unchanged at about 19 percent of GDP, whereas thesaving rate declined from 16-1/2 percent to slightly less than 14percent of GDP.2 Domestic investment not funded by domestic saving must be financed bycapital flows from abroad, and, indeed, the large increase in the U.S.current account deficit was matched by a similar expansion of netcapital inflows.
Globally, national current account deficits and surpluses mustbalance out, as deficit countries can raise funds in internationalcapital markets only to the extent that other (surplus) countriesprovide those funds. Accordingly, it is not surprising that thewidening of the U.S. current account deficit has been associated withincreased current account surpluses in the rest of the world.
What is surprising, however, in light of historical patterns, isthat much of the increase in current account surpluses during thisperiod took place in developing countries rather than in the industrialcountries.3 The table shows current account balances for various countries andregions in selected years. The aggregate current account balance ofindustrial countries other than the United States did increase between1996 and 2004, by a bit less than $200 billion, much of that rise beingaccounted for by an increase in Japan's current account balance; theaggregate balance of the euro area rose only slightly.4 In comparison, the aggregate current account position of developingcountries swung from a deficit of about $80 billion in 1996 to asurplus of roughly $300 billion in 2004, a net move toward surplus of$380 billion.
In the aggregate, the shift from deficit to surplus in the currentaccount of the emerging-market world over this period largely reflectedincreased saving as a share of output rather than a decline in the rateof capital investment. However, changes in saving and investmentpatterns varied by countries and regions. For example, in thecountries of developing Asia excluding China, most of the $150 billionswing toward external surplus between 1996 and 2004 was attributable todeclines in domestic investment. In China, rates of both saving andinvestment rose, but saving rates rose more, leading to an increase inthat country's current account surplus of about $60 billion.
Outside of developing Asia, oil exporters in the Middle East and theformer Soviet Union were also important contributors to the largeincrease in emerging-market current account balances. The combinedcurrent accounts of the two regions increased from a surplus of $20billion in 1996 to a surplus of $162 billion in 2004, an increase ofabout $140 billion. This rise largely reflected higher saving rates,as domestic consumption fell behind the surge in oil revenues. Amongother emerging-market economies, higher saving also accounted for anincrease in the aggregate current account balance of Latin America. Ofcourse, as emerging-market countries switched from being net borrowersto being net lenders, they began to pay down their international debtsand to acquire assets of industrial countries.
I have noted the expansion of the U.S. current account deficit andthe associated increases in current account surpluses abroad over the1996-2004 period. A third key development in that period was asustained decline in long-term real interest rates in many parts of theworld. For example, the real yield on ten-year inflation-indexed U.S.Treasury securities averaged about 4 percent in 1999 but less than 2percent in 2004. The difference between the nominal long-term Treasuryyield and the trailing twelve-month rate of consumer price inflation,another measure of the U.S. real interest rate, showed a similarpattern, falling from about 3.5 percent in 1996 to about 1.5 percent in2004. Similar movements were observed in other industrial countries: In the United Kingdom, the real yields on inflation-indexed governmentbonds fell from an average of 3.6 percent in 1996 to just below 2percent in 2004; in Canada, the analogous figures were 4.6 percent in1996 and 2.3 percent in 2004. Real interest rates measured as thedifference between government bond yields and consumer inflation alsofell in Germany, Sweden, and Switzerland. However, in Japan, realinterest rates remained low throughout the period.
In sum, considering the 1996-2004 period, we have three facts toexplain: (1) the substantial increase in the U.S. current accountdeficit, (2) the swing from moderate deficits to large surpluses inemerging-market countries, and (3) the significant decline in long-termreal interest rates. Many observers have focused on the expansion ofthe U.S. current account deficit in isolation and have argued that itis due largely to domestic factors, particularly declines in bothpublic and private saving rates. But accounting identities assure usthat any movement in the current account must involve changes inrealized saving rates relative to investment rates. The question atissue, therefore, is whether the decline in the realized saving rate inthe United States reflected a decline in desired saving or wasinstead a response to other, possibly external, economic developments. Or, in textbook terms, did the fall in the realized saving rate in theUnited States reflect a shift in the demand for savings at any giveninterest rate (a shift in the saving schedule) or a decline in savingsinduced by a change in the interest rate (a movement along the savingschedule)?
In fact, there is no obvious reason why the desired saving rate inthe United States should have fallen precipitously over the 1996-2004period.5 Indeed, the federal budget deficit, an oft-cited source of the declinein U.S. saving, was actually in surplus during the 1998-2001 periodeven as the current account deficit was widening. Moreover, a downwardshift in the U.S. desired saving rate, all else being equal, shouldhave led to greater pressure on economic resources and thus toincreases, not decreases, in real interest rates. As I will discusslater, from a normative viewpoint, we have good reasons to believe thatthe U.S. saving rate should be higher than it is. Nonetheless,domestic factors alone do not seem to account for the largedeterioration in the U.S. external balance.
In my earlier speech, I put forth an alternative explanation that isconsistent with each of the three basic facts I listed earlier. Thatexplanation takes as a key driving force a large increase in netdesired saving (that is, desired saving less desired domesticinvestment) in emerging-market and oil-producing economies, a changethat transformed these countries from modest net demanders tosubstantial net suppliers of funds to international capital markets. This large increase in the net supply of financial capital from sourcesoutside the industrial countries is what, in my earlier remarks, Icalled the global saving glut.
To interpret the rise in net saving in emerging-market countries ascausal, we need to identify factors in those countries that may havecaused their desired saving to rise, or their desired investment tofall, or both. In fact, several factors appear to have contributed tothe increase in the supply of net saving from emerging-marketcountries. First, the financial crises that hit many Asian economiesin the 1990s led to significant declines in investment in thosecountries (in part because of reduced confidence in domestic financialinstitutions) and to changes in policies--including a resistance tocurrency appreciation, the determined accumulation of foreign exchangereserves, and fiscal consolidation--that had the effect of promotingcurrent account surpluses. Second, sharp increases in crude oil pricesboosted oil exporters' incomes by more than those countries were ableor willing to increase spending, thereby leading to higher saving andcurrent account surpluses. Finally, Chinese saving rates rose rapidly(by more even than investment rates); that rise in saving was, perhaps,a result of the strong growth in incomes in the midst of anunderdeveloped financial sector and a weak social safety net thatincreases the motivation for precautionary saving.
The combined effect of these developments, I argued, raised desiredsaving relative to desired investment in the emerging markets, which inturn led to current account surpluses in those countries. But for theworld as a whole, total saving must equal investment, and the sum ofnational current account balances must be zero. Accordingly, in theindustrial economies, realized saving rates had to fall relative toinvestment, and current account deficits had to emerge as counterpartsto the developing countries' surpluses. This adjustment could beachieved only by declines in real interest rates (as well as increasesin asset prices), as we observed. The effects were particularly largein the United States, perhaps because high productivity growth and deepcapital markets in that country were particularly attractive to foreigncapital. The global saving glut hypothesis is thus consistent with thethree key facts I noted earlier.
To be sure, the global saving glut was not the only factor behindthe decline in long-term real interest rates since the 1990s. As Idescribed in subsequent remarks (Bernanke, 2006), term premiums alsodeclined during this period for reasons that are debated but may haveincluded a perceived reduction in uncertainty regarding inflation andthe real economy as well as increased demand for longer-term securitiesby various institutional investors, including pension funds and foreigncentral banks. Changes in the global pattern of saving and investmentsurely played an important role in the decline in long-term rates,however.
Recent Developments
I turn now to a review of developmentssince I last spoke on these issues two and a half years ago. In brief,external imbalances have become wider since 2004. Both thegeographical pattern of these imbalances and their sources in terms ofsaving and investment rates have changed a bit. Nevertheless, thebroad configuration that developed after 1996 still seems to be inplace today.
As the table shows, the U.S. current account deficit has widenedfurther in the past two years, from $640 billion in 2004 (5.5 percentof GDP) to $812 billion in 2006 (6.2 percent of GDP), although it fella bit in the first quarter of this year, to $770 billion at an annualrate. In an accounting sense, the increase in the U.S. deficit overthis period reflects primarily an increase in the investment rate fromabout 19 percent of GDP in 2004 to 20 percent of GDP in 2006. The U.S.national saving rate did not change significantly over that period.
Meanwhile, the aggregate current account surplus of emerging-market economies expanded about $350 billion, from$297 billion in 2004 to $643 billion in 2006; almost all the increasewas attributable to a higher aggregate rate of saving. A significantportion of this further growth is due to China, whose current accountsurplus swelled an additional $180 billion, rising from 3.6 percent ofnational output in 2004 to 9.4 percent in 2006. The increase in theChinese surplus can be attributed primarily to an increase in thesaving rate between 2004 and 2006. The increase in China's saving ratecould, in part, be a consequence of the rapid pace of growth in thecountry. That is, with income growing very rapidly, but with consumercredit not readily available and precautionary motives for savingremaining strong, consumption is failing to catch up.6 Also contributing to high saving rates was the authorities' decision tolimit currency appreciation, thereby restraining import demand andboosting exports.
Oil exporters have also contributed significantly to the recentincrease in the aggregate current account balance of developingcountries. The combined current account balance of the countries ofthe Middle East and the former Soviet Union (which include a number oflarge oil exporters) rose about $150 billion between 2004 and 2006. Again, the increase is almost entirely reflected in higher savingrates, as the oil exporters continue to save a large portion of theincreased revenue resulting from higher oil prices.
In contrast to the situation in emerging markets, the aggregatecurrent account surplus for industrial countries other than the UnitedStates declined recently, from almost $350 billion in 2004 to about$200 billion in 2006; most of the decline reflected a sharp drop in theeuro-area balance. Thus, unlike in the 1996-2004 period, industrialcountries other than the United States have absorbed part of theincrease in the net supply of capital coming from the emerging-marketeconomies. In aggregate, the recent decline in the current accountbalances of non-U.S. industrial economies reflects an increase ininvestment rates; saving rates have generally remained little changed.7 Inshort, in the emerging markets, realized saving and current accountsurpluses have increased since 2004. In the industrial countries, overthe same period, current accounts have moved further into deficit,primarily because of higher realized rates of investment.
What about real interest rates? Since I discussed these issues inMarch 2005, real interest rates have reversed some of their previousdeclines. For example, in the United States, real yields oninflation-indexed government debt averaged 2.3 percent in 2006 ascompared with 1.85 percent in 2004. In the past few weeks, that yieldhas averaged about 2.4 percent. Inflation-adjusted yields in otherindustrial countries have also started to move back up after falling in2005.8
How does this all fit together? My reading of recent developmentsis that although some of the details have changed, the fundamentalelements of the global saving glut remain in place. Most important,the emerging-market countries and oil producers remain large netsuppliers of financial capital to global markets. The mix of suppliersof funds and the factors motivating that supply have changed a bit: China and the oil exporters account for a larger share of thedeveloping countries' aggregate surplus, and developing Asia excludingChina accounts for somewhat less. Also, the further expansion of theregion's net supply of saving in the past two years appears to reflectprimarily an increase in desired saving by the emerging-marketcountries, whereas the previous increase in net saving also involvedsome decline in desired investment in East Asia after the financialcrises of the 1990s. Exchange rate policies in Asia have alsoinfluenced desired saving in that region.
Further increases in net capital flows from the developingeconomies, all else being equal, should have further depressed realinterest rates around the world. But as I have noted, in the past fewyears, real interest rates have moved up a bit. This increase does notimply that the global saving glut has dissipated. However, it doessuggest that, at the margin, desired investment net of desired savingmust have risen in the industrial countries enough to offset anyincrease in desired saving by emerging-market countries. Thischaracterization is certainly consistent with the pickup in investmentrates in the industrial countries, which I noted earlier, and it isalso consistent, more generally, with the recovery of domestic demandgrowth in Europe, Japan, and other parts of the industrial world. Insummary, economic growth over the past few years, especially inindustrial countries, has apparently been sufficient to increase thenet demand for saving and thus to raise global real interest ratessomewhat.
Once again, however, I do not want to rely exclusively on this lineof explanation for the behavior of long-term real interest rates, asother factors have no doubt been relevant. In particular, termpremiums appear recently to have risen from what may have beenunsustainably low levels, in part because of the greater recentvolatility in financial markets and investors' demands for increasedcompensation for risk-taking.
Are Current Account Imbalances a Problem?
This analysis ofthe sources of global imbalances does not address the criticalnormative question: Are the current account imbalances that we seetoday a problem? Not everyone would agree that they are, for severalreasons.
First, these external imbalances are to a significant extent amarket phenomenon and, in the case of the U.S. deficit, reflect theattractiveness of both the U.S. economy overall and the depth,liquidity, and legal safeguards associated with its capital markets.9 Of course, some foreign governments have intervened in foreign exchangemarkets and invested the proceeds in U.S. and other capital markets,which most likely has led to greater imbalances than would otherwiseexist. But the supply of capital from foreign governments is not aslarge as that from foreign private investors. From 1998 through 2001,even as the U.S. current account deficit widened substantially,official capital flows into the United States were quite small. Duringthe years 2002 through 2006, net official capital inflows picked upsubstantially but still corresponded to less than half (47 percent) ofthe U.S. current account deficit over the period. On a gross basis,during the same period, private foreign inflows were three timesofficial capital flows.10 Moreover, even public investors are motivated to some extent by the attractions of the U.S. economy and U.S. capital markets.
Second, current account imbalances can help reduce tendencies towardrecession, on the one hand, or overheating and inflation, on the other.11 During the late 1990s, for example, the developing Asian economies thathad experienced financial crises and consequent collapses in domesticinvestment benefited from being able to run trade surpluses, whichhelped strengthen aggregate demand and employment. During that sameperiod, the trade deficits run by the United States allowed domesticdemand to grow strongly without creating significant inflationarypressures. Until a few years ago, the euro area was growing slowly andthus also benefited from running trade surpluses; more recently, asdomestic demand in Europe has recovered, the trade surplus hasdeclined.
Third, although the U.S. current account deficit is certainly notsustainable at its current level, U.S. liabilities to foreigners arenot, at this point, putting an exceptionally large burden on theAmerican economy. The net international investment position (NIIP) ofthe United States, although at a substantial negative 19 percent ofGDP, is still smaller than the negative NIIP of several otherindustrial economies. As a fraction of net household wealth, whichtotaled almost $56 trillion in 2006, the negative NIIP is evensmaller--less than 5 percent. Moreover, the U.S. investment incomebalance, which essentially represents the debt service on the NIIP,remains positive, at least for now. Thus, even after years of currentaccount deficits and corresponding increases in net liabilities, theUnited States continues to earn more on its foreign investments than itpays on its foreign liabilities. And, as best we can tell, the shareof U.S. assets in foreign portfolios does not seem excessive relativeto the importance of the United States in the global economy.
All that said, the current pattern of external imbalances--theexport of capital from the developing countries to the industrialeconomies, particularly the United States--may prove counterproductiveover the longer term. I noted some reasons for concern in my earlierspeech, and they remain relevant today.
First, the United States and other industrial economies face theprospect of aging populations and of workforces that are growing moreslowly. These trends enhance the need to save (to support futureretirees) and may reduce incentives to invest (because workforceseventually will shrink). If the United States saved more, one likelyoutcome would be a reduction in the U.S. current account deficit and inthe rate at which the country is adding to its liabilities to the restof the world.
Second, the large U.S. current account deficit cannot persistindefinitely because the ability of the United States to make debtservice payments and the willingness of foreigners to hold U.S. assetsin their portfolios are both limited. Adjustment must eventually takeplace, and the process of adjustment will have both real and financialconsequences. For example, in the United States, the growth ofexport-oriented sectors such as manufacturing has been restrained bythe shifts in relative prices and foreign demand associated with theU.S. trade deficit. Ultimately, the necessary reduction in the tradeand current account deficits will entail shifting resources out ofsectors producing nontraded goods and services to those producingtradables. The greater the needed adjustment, the more potentiallydisruptive and costly these shifts may be. Similarly, externaladjustment for China and other surplus countries will involve shiftingresources out of the export sector and into industries geared towardmeeting domestic consumption needs; that necessary shift, too, willlikely be less disruptive if it occurs earlier and thus less rapidlyand on a smaller scale.
On the financial side, if U.S. current account deficits were topersist at near their current levels, foreign investors wouldultimately become satiated with dollar assets, and financing thedeficit at a reasonable cost would become difficult. Earlier reductionof global imbalances would reduce the potential strains associated withfinancing a large quantity of international liabilities and likelyallow a smoother adjustment in financial markets.
Finally, in the longer term, the developing world should be therecipient, not the provider, of financial capital. Because developingcountries tend to have high ratios of labor to capital and to be awayfrom the technological frontier, the potential returns to investment inthose countries are high. Thus, capital flows toward those countriesshould benefit both them and the countries providing the capital.
Prospects for Reducing External Imbalances
What are theprospects for a gradual and orderly rebalancing of spending andexternal accounts around the world? The brief answer is that signs ofprogress have appeared but that most countries have only just begun toundertake the policy changes that will ultimately be needed.
Recently, the pickup in economic growth outside the United States,together with changes in the real exchange rate and other relativeprices, has assisted the process of current account adjustment. Notably, during 2006, foreign growth helped U.S. real exports of goodsand services grow 9.3 percent, and exports of capital goods rose 10.8percent. Some of the gain in foreign growth is cyclical, but some isdue to economic reforms (in both industrial and non-industrialcountries) and thus may be more persistent. Overall, we have seen somemodest indications of improvement in the U.S. external balancerecently. For example, the non-oil trade deficit has declinedmodestly, from 3.7 percent of U.S. GDP in 2004 to 3.5 percent of GDP in2006. In addition, in 2006, net exports made a positive contributionto U.S. real GDP growth, the first year that had happened since 1995. Net exports also contributed to U.S. growth in the first half of 2007.
As is well known, however, further progress on the U.S. currentaccount seems unlikely without significant increases in public andprivate saving in the United States. The U.S. federal budget deficithas declined recently and is officially projected to improve furtherover the next few years. Unfortunately, as I have noted, the UnitedStates has already reached the leading edge of major demographicchanges that will result in an older population and a more slowlygrowing workforce. A major effort to increase public and privatesaving is needed to prepare for the economic consequences of thisdemographic transition and to address external imbalances.
As the global perspective makes clear, the reduction of the U.S.current account deficit also requires efforts on the part of thesurplus countries to reduce the excess of their desired saving overdesired investment. Over the longer term, the current accountsurpluses of the emerging-market countries seem likely to narrow asdomestic spending catches up with income. Economic policies in thesecountries can assist this process. For example, the oil exporters havecollectively saved much of the windfall arising from higher crudeprices in recent years; they should spend more in the future to developand diversify their domestic economies. China has officiallyrecognized the need to increase its domestic spending and scale backits reliance on exports. Measures that could help achieve these goalsinclude further reforms of the financial sector; increased governmentspending on infrastructure, environmental improvement, and the socialsafety net; and currency appreciation. In East Asia excluding China,continued efforts to strengthen and deepen the banking sector andfinancial markets would help domestic investment recover from thelingering effects of the financial crises of the 1990s. In each ofthese cases, the indicated policies would reduce global imbalances. Moreover, as with U.S. saving efforts, these actions would conveyimportant economic benefits to the countries undertaking them even ifcurrent account balances were not an issue.
What implications would a gradual rebalancing have for long-termreal interest rates? The logic of the global saving glut suggeststhat, as the glut dissipates over the next few decades and therebyreduces the net supply of financial capital from emerging-marketcountries, real interest rates should rise--a tendency that seemslikely to be only partly offset by increased saving in the industrialcountries. However, factors other than the saving-investment balanceaffect long-term interest rates, including the relative supplies of,and demands for, long-term securities and changes in the requiredcompensation for the risk embedded in term premiums. Moreover, distantone-year forward interest rates remain low, an indication that marketscurrently do not expect much change in the global balance of desiredsaving and investment or that they expect the effects of such a changeto be offset by other developments. Accordingly, we are again remindedof the need to maintain appropriate humility in forecasting returns andasset prices.
(Billions of U.S. dollars)
Country or region | 1996 | 2000 | 2004 | 2005 | 2006 |
|---|---|---|---|---|---|
Industrial | 31.1 | -304.7 | -296.5 | -502.5 | -607.3 |
United States | -124.8 | -417.4 | -640.2 | -754.8 | -811.5 |
Japan | 65.7 | 119.6 | 172.1 | 165.7 | 170.4 |
Euro area 1 | 77.3 | -37.0 | 115.0 | 22.2 | -11.1 |
France | 23.4 | 22.3 | 10.5 | -19.5 | -28.3 |
Germany | -14.0 | -32.6 | 118.0 | 128.4 | 146.4 |
Italy | 36.8 | -6.2 | -15.5 | -28.4 | -41.6 |
Spain | -1.4 | -23.1 | -54.9 | -83.0 | -108.0 |
Other | 12.9 | 30.0 | 56.6 | 64.4 | 45.0 |
Australia | -15.4 | -14.9 | -38.5 | -41.2 | -40.9 |
Canada | 3.4 | 19.7 | 21.3 | 26.3 | 21.5 |
Switzerland | 22.0 | 30.7 | 50.4 | 61.4 | 69.8 |
United Kingdom | -10.5 | -37.6 | -35.4 | -53.7 | -88.3 |
Memo: | 155.9 | 112.7 | 343.7 | 252.3 | 204.2 |
Developing | -82.8 | 124.7 | 296.5 | 507.9 | 643.2 |
Asia | -40.2 | 77.0 | 172.4 | 245.1 | 352.1 |
China | 7.2 | 20.5 | 68.7 | 160.8 | 249.9 |
Hong Kong | -4.0 | 7.0 | 15.7 | 20.3 | 20.6 |
Korea | -23.1 | 12.3 | 28.2 | 15.0 | 6.1 |
Taiwan | 10.9 | 8.9 | 18.5 | 16.0 | 24.7 |
Thailand | -14.4 | 9.3 | 2.8 | -7.9 | 3.2 |
Latin America | -39.1 | -48.1 | 20.4 | 34.6 | 48.7 |
Argentina | -6.8 | -9.0 | 3.2 | 3.5 | 5.2 |
Brazil | -23.5 | -24.2 | 11.7 | 14.2 | 13.6 |
Mexico | -2.5 | -18.7 | -6.7 | -4.9 | -1.5 |
Middle East | 15.1 | 72.1 | 99.2 | 189.0 | 212.4 |
Africa | -5.2 | 7.2 | 0.6 | 14.6 | 19.9 |
Eastern Europe | -18.5 | -31.8 | -58.6 | -63.2 | -88.9 |
Former Soviet Union | 5.2 | 48.3 | 62.6 | 87.7 | 99.0 |
Memo: | -47.4 | 56.5 | 103.7 | 84.3 | 102.2 |
Statistical discrepancy | -51.6 | -180.0 | 0.0 | 5.4 | 35.9 |
1. Calculated as the sum of the balances of the thirteen euro-area countries. Return to table
Source: For the United States, Department of Commerce, Bureau ofEconomic Analysis. For some countries other than the United States,national sources; for most countries, however, International MonetaryFund (IMF), World Economic Outlook Database 
, April 2007 (www.imf.org/external/pubs/ft/weo/2007/01/data/index.aspx); some values for 2006 are IMF estimates.
References
Bernanke, Ben S. (2005). "The Global Saving Glut and the U.S. Current Account Deficit,"speech delivered for the Sandridge Lecture at the Virginia Associationof Economists, Richmond, March 10,www.federalreserve.gov/boarddocs/speeches/2005/200503102/default.htm. Similar remarks with updated data were presented for the Homer Jones Lecture, St. Louis, April 14, 2005, www.federalreserve.gov/boarddocs/speeches/2005/20050414/default.htm.
------------ (2006). "Reflections on the Yield Curve and Monetary Policy,"speech delivered at the Economic Club of New York, New York, March 20,www.federalreserve.gov/newsevents/speech/bernanke20060320a.htm.
Caballero, Ricardo J., Emmanuel Farhi, and Pierre-Olivier Gourinchas (2006). "An Equilibrium Model of 'Global Imbalances' and Low Interest Rates ,"NBER Working Paper Series 11996. Cambridge, Mass.: National Bureau ofEconomic Research, January, www.nber.org/papers/w11996.pdf.
Mendoza, Enrique G., Vincenzo Quadrini, and Jose-Victor Rios-Rull (2007). "Financial Integration, Financial Deepness, and Global Imbalances ," NBER Working Paper Series 12909. Cambridge, Mass.: National Bureauof Economic Research, February, www.nber.org/papers/w12909.pdf.
Footnotes
1. The shift was almost wholly attributable to asimilar expansion of the trade deficit. The balance on investmentincome actually improved over the period. Return to text
2. More precisely, investment grew from 19.0percent to 19.3 percent of GDP, and saving declined from 16.5 percentto 13.8 percent of GDP, for a net change in investment less saving of3.0 percent of GDP. As implied by data noted earlier in thisparagraph, the net change in the U.S. current account deficit over thesame period was 3.9 percent of GDP. In principle, the change in theexcess of investment over saving and the change in the current accountdeficit should be the same. The difference between the two figures isaccounted for by statistical discrepancies, both within the nationalincome and product accounts (NIPA) and between the balance of paymentsdefinitions and NIPA definitions of certain international transactions. Return to text
3. I am using the terms "emerging-market" and "developing" interchangeably. Return to text
4. As shown in the table, the surplus ofindustrial countries other than the United States increased from about$150 billion to nearly $350 billion over the period, and the Japaneseexternal balance rose from $66 billion to $172 billion. The increasein the Japanese current account balance as a share of GDP, from 1.4percent to 3.7 percent, occurred despite a substantial fall in the GDPshare of the saving rate, from 30.4 percent to 26.8 percent, as the GDPshare of the investment rate fell even more dramatically, from 28.9percent to 23.0 percent. For the euro area as a whole, the currentaccount balance remained at about 1 percent of GDP between 1996 and2004, as aggregate investment and saving ratios remained largelyunchanged. Within the euro area, Germany's current account balanceincreased almost 5 percentage points of GDP--from -0.6 percent in 1996to 4.3 percent in 2004--as saving moved up and investment decreased. However, this development was offset by declines in the balances ofsome other euro-area countries, including France, Italy, and Spain; thedecreases were mostly associated with higher investment rates. Data onsaving, investment, and current account balances for countries otherthan the United States are drawn primarily from the InternationalMonetary Fund, World Economic Outlook Database , April 2007 (www.imf.org/external/pubs/ft/weo/2007/01/data/index.aspx); in some cases, data are drawn from national sources. Return to text
5. During the first part of the period, the risein U.S. productivity and higher stock prices likely contributed to theU.S. current account deficit by increasing desired investment andreducing desired saving. However, some of the increase in stock pricesmay have been the endogenous result of factors discussed later, and inany case the effects of the stock market on investment dissipated by2004. Finally, as noted in the text, if the driving force behind thechanges in external balances was a decline in desired saving in theUnited States, world real interest rates would have risen rather thanfallen. Return to text
6. The combined current account balance ofdeveloping Asia excluding China narrowed a bit as a share of GDPbetween 2004 and 2006, as the investment rate edged up while the savingrate was little changed. Nevertheless, investment rates in this regionstill remain substantially below their 1996 levels. Return to text
7. The combined current account balance for theeuro area moved from a surplus of $115 billion in 2004 to a deficit ofabout $10 billion in 2006, largely because of an increase in theaggregate investment rate. Large declines in the balances of France,Italy, and Spain more than offset a higher surplus in the balance ofGermany. For the euro area as a whole, the movement into deficit haslargely reflected an increase in the euro-area investment rate fromabout 20 percent of GDP in 2004 to about 21 percent of GDP in 2006. Japan's current account surplus was almost unchanged at around $170billion in both 2004 and 2006, as an increase in the rate of investmentwas matched by a higher saving rate. Return to text
8. Inflation-adjusted bonds in the United Kingdomhad a yield of 2.19 percent, on average, in July 2007 as compared witha yield of 1.65 percent, on average, in July 2005. In Canada, yieldson inflation-adjusted bonds moved from 1.76 percent in July 2005 to2.18 percent in July 2007. Real interest rates, calculated asgovernment bond yields minus twelve-month inflation rates, have alsomoved up since 2005 in Germany, Sweden, and Switzerland. Return to text
9. An interesting vein of recent researchsuggests that one of the reasons that developing countries seek to runcurrent account surpluses is to finance the acquisition of high-qualityassets they cannot produce in their own economies. Refer to Caballero,Farhi, and Gourinchas (2006) and Mendoza, Quadrini, and Rios-Rull(2007). Return to text
10. During 2002-06, gross foreign officialinflows totaled $1,491 billion; net official inflows were only slightlyless, as U.S. official outflows were negligible. Private foreigninflows net of private U.S. outflows totaled $1,659 billion during thesame period; gross foreign private inflows were $4,697 billion. Return to text
11. Another way to make this point isthat current account balances and surpluses give countries theflexibility to spend more or less than their current output, asdictated by economic conditions and needs. Return to text
