David Gruen and Jason Harris1
It is often argued that the United States cannot continue for long to run current account deficits of their current size of 5percent of Gross Domestic Product (GDP). This article questions this conventional wisdom by examining the implications were the United States to continue to run current account deficits of 5percent of GDP for the next ten years. It suggests:
- Despite the size of the United States, the asset-composition of the wealth portfolio of the average non-United States citizen would need to change by less than might be anticipated.
- The risk premium on United States assets might remain quite small.
- The most serious risk to the medium term outlook for the United States is a continuation of large fiscal deficits, not large current account deficits, per se.
Might the United States continue to run large current account deficits?
At the end of the 1980s, after a decade in which Australia’s current account deficit averaged 4percent of GDP, the high current account deficit and the associated strongly rising level of foreign liabilities was widely regarded as the most serious economic problem facing Australia.2 Over the subsequent nearly decade-and-a-half to the present, the Australian current account deficit has averaged a slightly smaller 4percent of GDP.3 Despite this, the performance of the Australian macroeconomy has been impressive over this time. Are there any lessons for the United States from this experience?
The United States net external asset position
As a result of persistent current account deficits, the foreign asset position of the United States moved from net external asset holdings of 13percent of GDP in 1980 to net external liabilities of 21.3percent of GDP in 2002. If the current account deficit were to remain at 5percent of GDP over the next ten years, United States net external liabilities would rise to around 56percent of GDP in 2014.4
This would represent the highest ratio of net external liabilities to GDP in United States history. Nevertheless, there are a number of advanced economies with ratios of net external liabilities to GDP higher than this: the Scandinavian countries in the mid-1990s; and Canada, New Zealand and Australia at present. In terms of their wider macroeconomic performance, these countries do not seem to have been adversely affected by these relatively large stocks of net external liabilities.5
In contrast to these small open economies, the trade-to-GDP ratio in the United States is relatively low, which means that a given percentage-point-of-GDP adjustment to the trade balance would require a larger real depreciation than it would were the United States economy more open to trade. Perhaps more importantly, the United States is the world’s largest economy, and the United States current account deficit currently represents a daily flow, to the rest of the world, of claims on United States assets of around US$1.5billion.
Over the past few years, much of the United States current account deficit has been funded by Asian central bank purchases of United States assets—largely United States Treasury securities. In 2003, for example, Asian official purchases funded about three-quarters of the United States current account deficit.
This article examines the implications were the United States to continue to run sizeable current account deficits, but with foreign central banks not continuing to raise their holdings of United States assets. It begins by examining whether the private sectors in the economies of the rest of the world could comfortably absorb a flow of asset claims of this magnitude for an extended period.
Implications for the global wealth portfolio
To derive quantitative estimates of what would be implied, we begin by generating estimates of the stock of non-United States private-sector wealth. For simplicity, the analysis is restricted to the non-United States Organisation for Economic Co-operation and Development (OECD) private sector. Results are presented for two measures of non-United States OECD private-sector wealth. The first measure is non-United States OECD private-sector net wealth, which is financial and non-financial assets less liabilities. Since not all these assets and liabilities are traded, results are also presented using a second measure: non-United States OECD private-sector net financial wealth. This measure includes tradeable financial assets, such as currency, equities, other securities and accounts receivable/payable, but excludes largely non-internationally tradeable assets such as dwellings and non-durable goods. All financial liabilities with the exception of mortgage liabilities are subtracted from the financial asset measure to give net financial assets.
Lacking wealth data for the whole non-United States OECD, we use OECD data on wealth and GDP for the G6 (the G7 excluding the United States), and scale the numbers up.
That is, to estimate each measure of non-United States OECD private-sector nominal wealth, the corresponding estimate of G6 private-sector nominal wealth is scaled up by the ratio: (non-United States OECD nominalGDP)/(G6nominal GDP). To convert wealth measures and GDPs into a common currency, December 2003 exchange rates are used.6
It will come as no surprise that the private sector owns large stocks of wealth. In 2002, the derived measure of non-United States OECD private-sector net wealth was 8.1times as large as annual United States GDP, while the derived measure of non-United States OECD private-sector net financial wealth was 4.1 times annual United States GDP.
In 2002, non-United States public sector holdings of United States assets (official foreign exchange reserves) amounted to 54percent of United States net external liabilities. (United States official holdings of foreign exchange reserves were trivial by comparison, and we ignore them.) So, to fund cumulated past United States current account deficits, the non-United States private sector needed to hold 46percent of United States net external liabilities in 2002.
Of course, the non-United States private sector’s actual holdings of United States assets were (and are) much larger than this—because there are big gross asset holdings in both directions, with the United States private sector holding sizeable stocks of foreign assets and vice versa. But we are interested in how much of the non-United States private-sector wealth portfolio must be taken up by United States assets to fund past and future United States current account deficits, and so these are the numbers on which we focus.
In 2002, the 46percent of United States net external liabilities that needed to be held by the non-United States private sector amounted to 1.2percent of non-United States OECD private-sector net wealth, or 2.3percent of non-United States OECD private-sector net financial wealth (from now on, we assume for simplicity that non-OECD private sector holdings of United States assets are zero).
To project these numbers out over the next ten years, the entire flow of assets (5percent of United States GDP) is assumed to be purchased by the non-United States OECD private sector. That is, there is no change in the non-United States public sector’s holdings of United States assets, and no flow of assets to the non-OECD. As a result, the required adjustment in the wealth portfolio of the non-United States OECD private sector is likely over-estimated.
Assumptions about how each of the relevant aggregates grows over time are also needed. We assume average annual growth
rates equal to those over the period since 1990, which are 5.0percent for United States nominal GDP, 2.5percent for non-United States OECD private-sector net wealth, and 5.6percent for non-United States OECD private-sector net financial wealth.
Finally, no further depreciation of the United States dollar is assumed and therefore December 2003 exchange rates are used to convert future stocks of non-United States OECD wealth into United States dollars.
In contrast, if the United States dollar were to depreciate further over this period as part of the adjustment to the large current account deficits, our assumption would again overestimate the required adjustment in the wealth portfolio of the non-United States OECD private sector, since non-United States OECD citizens would then be able to buy United States assets at lower prices in the future with a depreciated United States dollar.7
Chart 1 shows projections of the two measures of non-United States OECD wealth over the next decade, based on these assumptions. By 2014, United States assets would need to make up 8.3percent of non-United States OECD private-sector net wealth, or 11.2percent of non-United States OECD private-sector net financial wealth. Of course, the results change with changes in the assumptions. But the chart suggests, for both wealth measures, that the proportion of United States assets that must be held in the non-United States OECD private sector wealth portfolio is quite modest and rises only gradually over time.
The willingness of foreign investors to increase their holdings of United States assets in the future presumably depends on the current composition of their asset holdings. A study by Bertaut and Griever (2004), examining the foreign equity and debt holdings of the world’s major developed economies other than the US, found that almost all of them were underweight in United States assets at the time of the latest available data in 2001.
This applied to holdings of United States assets relative to total assets, which we would expect because of ‘home bias’—the widespread tendency for individuals and firms to hold a disproportionate share of their wealth in domestic, rather than foreign, assets. But it also applies to holdings of United States assets relative to holdings of other foreign assets. Bertaut’s and Griever’s results therefore suggest that there remains substantial room for increased holdings of United States assets in the non-United States OECD wealth portfolio.
On the basis of this evidence, it seems possible that the United States could continue to run sizeable current account deficits for quite some time.
Chart 1: Proportion of United States assets that would need to be held
in the wealth portfolio of the non-United States OECD private sector to fund a United States current account deficit of 5percent of GDP
Source: OECD Economic Outlook 74, Bureau of Economic Analysis, Treasury calculations.
A rising risk premium on United States interest rates?
If the United States were to continue to run sizeable current account deficits, would there be significant implications for the risk premium on United States interest rates? Recent cross-country evidence on the link between net foreign assets and real interest rates seems unconvincing.8 Australia’s experience may, however, be illuminating.9
In his report to the Government on Australian national saving, FitzGerald (1993) provided estimates of the real interest differential between 10-year bonds in Australia and in the US, Germany, and Japan. He argued, on the basis of data up to the end of 1992, that a sizeable risk premium was being added to Australian real interest rates because of the high level of Australia’s net external liabilities, which at the time amounted to around 50percent of GDP.
Chart 2 shows the relevant real interest differentials, with the dark vertical line marking the end of the sample used by FitzGerald. While the data up to the end of 1992 do suggest the emergence of a sizeable Australian risk premium, the experience of the subsequent decade or so has not been kind to this hypothesis. Australian net foreign liabilities are now close to 60percent of GDP, but the more recent experience suggests a much smaller risk premium on Australian real long bond yields than appeared to be the case on the basis of data up to 1992.10
Chart 2: Real long term interest differentials
Australia versus US, Germany and Japan(a)
(a) 10-year government bond yields deflated by core consumer price inflation over the previous year in each country.
Source: RBA, OECD, Thomson Financial.
A natural explanation for the relatively high Australian real bond yields in the early 1990s, and their subsequent fall, is that markets took a long time to be convinced that the early 1990s step-down in inflation in Australia would be sustained.
But this argument can be taken further. Along with the well established floating exchange rate regime, it has surely been the development of medium-term macroeconomic frameworks for both monetary and fiscal policy in Australia in the 1990s—and the demonstrated commitment to use these frameworks to discipline monetary and fiscal outcomes—that has been important in retaining the confidence of foreign investors, and enabling Australia to borrow in international capital markets on favourable terms.11 Despite a backdrop, since the early 1990s, of sustained large current account deficits and a gradually rising ratio of net external liabilities to GDP, over the years 1999-2003, rating agencies Moody’s and Standard and Poors unwound the downgrades they had imposed in the 1980s, and returned Australia to Aaa and AAA ratings. The treatment of Australia as a financial safe haven during the Asian crisis, with Australian bond yields falling despite a rise in the current account deficit of more than twopercent of GDP, also suggests that the combination of the floating exchange rate and the medium-term macroeconomic policy frameworks played an important role in retaining international confidence at that time.
In the United States, the Federal Reserve has substantial anti-inflation credibility, having maintained low inflation in the United States since the early 1980s. Although it does not have a formal inflation target, the Federal Reserve aims for ‘practical price stability’ which amounts to much the same thing.
In contrast, the rapid deterioration of the United States fiscal position over recent years has been a cause for rising concern, as explained by Rubin et al (2004), Kohn (2004) and the IMF (2004). Public dissaving now accounts for almost all of the United States current account deficit, with baseline fiscal deficits projected out to 2013, and likely legislative c
hanges expected to further add to them (Congressional Budget Office, 2004).
The relevance of the Australian and New Zealand experiences may be that the United States could perhaps continue to run sizeable current account deficits for many years with no obvious harmful side-effects—provided the United States fiscal deficit is significantly reduced (or eliminated). If, over time, the United States fiscal deficit was significantly reduced, that might also see a significant narrowing of the United States current account deficit—but the experiences of Australia and New Zealand caution against automatically assuming that outcome.
Continued large fiscal deficits seem to pose a much more serious risk to the United States outlook than continued large current account deficits on their own. If government debt continues to rise inexorably, with no convincing signs that it will be brought under control, it makes sense for investors to demand a rising risk premium—or at some point, to rush for the exits. On the one hand, investors may come to doubt the commitment to avoid inflation, since inflation would erode the real value of government debt. On the other hand, high public debt, in the absence of substantial inflation, eventually needs to be repaid, implying future surpluses and probably higher future taxes. Given that these taxes could be imposed on investment income earned in the United States—including interest and dividend payments on foreign owned United States assets—investors might at some point begin to demand an additional premium on United States assets.
Just how long the United States could continue to run fiscal deficits of their current size without a serious loss of confidence is a question we hope will not be answered.
Bertaut C.C. and Griever W.L. (2004). Recent Developments in Cross-Border Investment in Securities, Federal Reserve Bulletin, Winter 2004.
Treasury (2001). The Net Income Deficit Over the past Two Decades, Economic Roundup – Centenary Edition 2001, pp. 105-135.
Congressional Budget Office (2004). The Budget and Economic Outlook, Fiscal Years 2005-2014, January 2004.
FitzGerald VW (1993). National Saving: A Report to the Treasurer, Australian Government Publishing Service, Canberra.
IMF (2004). US. Fiscal Policies and Priorities for Long-Run Sustainability, Mhleison M. and C. Towe (eds.), Occasional Paper 227.
Kohn D (2004). The United States and the World Economy, Remarks to the Federal Reserve Bank of Atlanta’s Public Policy Dinner, Atlanta, Georgia, 7 January 2004,
Lane PR and GM Milesi-Ferretti (2001). Long-Term Capital Movements, NBER Macroeconomics Annual, pp. 73–116.
Rubin RE, Orszag P.R. and Sinai A (2004). Sustained US Budget Deficits: Longer-Run US Economic Performance and the Risk of Financial and Fiscal Disarray, Paper presented at the American Economic Association annual meeting, San Diego, 4 January 2004, http://www.brook.edu/views/papers/orszag/20040105.htm.
1 The authors are from Macroeconomic Group, Australian Treasury. They are grateful for comments and suggestions from Steven Kennedy and Martin Parkinson. The views in this article are those of the authors and not necessarily those of the Australian Treasury.
2 A couple of newspaper headlines give a flavour of the times: ‘Australia’s [Foreign] Debt Hovering On The Cliff-Face Of Crisis’, Australian Financial Review 21 September 1989; ‘Australia Sliding Into [Foreign] Debt Trap Says Bank’, The Age, 20 November 1989. And not to forget Macquarie Bank’s November 1989 contribution to the debate, A Boiling Frog: Australia’s Economic Challenge, a booklet that argued that Australia’s current account problem could be likened to that of a frog immersed in initially cold, but gradually warming, water. Failing to realise the impending danger, the frog is eventually boiled.
3 The averages over the two sub-periods of the 1990s, and thus far in the 2000s, are both 4per cent of GDP.
4 This estimate ignores any valuation effects on the stocks of non-United States$-denominated external assets and liabilities from possible future changes in the value of the United States dollar.
5 Of course, a similar statement would not apply to developing countries, with their less-developed financial markets and significant mis-matches in the currency denomination of their foreign debts.
6 Market exchange rates seem more appropriate for our purposes than Purchasing Power Parity exchange rates, because United States assets change hands at market exchange rates.
7 If there was a widely held view that the United States dollar would continue to depreciate, investors might demand a higher US-dollar return on United States assets. This argument seems a valid one, but the Australian experience suggests that a long history of large current account deficits does not lead to a noticeable premium on domestic real interest rates (see Chart 2 and the associated discussion).
8 Influential observations in recent cross-country regressions run with 1990s data (Lane and Milesi-Ferretti, 2001) are those for Australia and New Zealand, with high levels of net external liabilities and high real interest rates, and Japan, with a high level of net foreign assets and low real interest rates. But, as we argue in the text, Australia had high real interest rates for much of the 1990s for another reason (and the same argument applies to New Zealand). And we suspect that Japan had low real interest rates because of the parlous state of its economy in the 1990s, not because of its high stock of foreign assets. It follows that the negative correlation between net foreign assets and real interest rates found in cross-country regressions in the 1990s may be largely spurious.
9 See also Treasury (2001).
10 Japan is something of an outlier in Chart 2, with lower real bond yields than Australia, United States or Germany since the mid 1990s. This is presumably for domestic Japanese reasons, with the Bank of Japan holding the policy rate at zero for the past several years.
11 Of course, much the same argument, and timing, applies to New Zealand.