When Paul Krugman explained in a Senate hearing that the US current account deficit reflected low US saving rates and high investment, an astonished senator responded that this was “a very interesting theory”. Economists agree that Krugman made here one of his less controversial statements by simply stating an accounting identity about the balance of payments. Matters become a bit more complicated when the issue is what such a deficit means for the future of the economy.
Accounting goes some way in providing insight. Persistent US current account deficits imply that foreigners are accumulating net claims on US assets. Assuming their willingness to hold dollar-denominated claims has some limit, the dynamic version of the current account identity tells us that repayment can be made in two ways:
- future US trade surpluses, or
- a depreciation of the dollar;
a combination of the two is of course possible.1
But where is the limit where these options must kick in? How willing are foreign investors to swap their domestic assets for US assets? Only at this stage of the discussion do economists leave the certainties of accounting and enter the realm of “interesting theory”.
Questions about the willingness of investors to hold foreign currency assets are hardly new. In the early 1980s, the Finnish economist Pentti Kouri, among others, developed the so-called “portfolio balance theory” before financial globalisation became a reality more than a decade later. The basic assumption is that foreign and domestic assets are imperfect substitutes and that investors care about the currency denomination of their assets. A current account deficit requires more foreign asset holdings. In equilibrium this can be achieved only if the deficit country depreciates and thereby makes the asset acquisition more attractive.
Portfolio theory thus provides the link between the current account deficits and the exchange rate dynamics.2 Unfortunately, the determinants of asset substitutability were largely left unspecified. The lack of microeconomic foundations meant that many economists did not regard the theory with much sympathy. What made things worse was the lack of empirical support in aggregate data.
Evidence hard to find in the macro data
Finding direct evidence on portfolio rebalancing in macroeconomic data may be exceedingly difficult.3 At a country level, many additional factors drive exchange rates and capital flows, and the small number of exchange rates limits the statistical power of any test.
We look for the effect in micro data
We use a new microeconomic approach based on individual equity fund holdings to test the empirical validity of the portfolio rebalancing paradigm. Our data set comprises more than 6,500 international equity funds domiciled in four different currency areas, namely Canada, the euro area, Britain and the US. For each fund position, we estimate the risk of the fund portfolio. We use the entire cross-section of funds and their heterogeneous portfolios to identify the effects of changes in asset valuations on portfolio reallocations.4
Simply describing the data provides some interesting insights:
- The so-called home bias at the fund level is radically different across countries even within the sample of equity funds that hold at least five foreign stocks in their portfolio.5
International equity funds domiciled in the US, for example, hold a relatively small share of their value in foreign equity, unlike their UK counterparts. The median foreign holding share is only 10% for a US fund and 85% for a UK fund.
- Differences arise concerning the exchange rate risk of any given level of international diversification.
For example, Canadian international funds benefit from a negative correlation between the Canadian/US exchange rate and the US stock market index during the period 1998-2002. This implies that exchange rate risk typically makes a negative contribution to the overall portfolio risk of an international equity fund that measures performance in Canadian dollars. But generally, higher levels of foreign investments tend to increase the share of exchange rate risk in the total portfolio risk.
Do funds rebalance their portfolios?
The most interesting question concerns how funds react to changes in their risk exposure – the question at the heart of the portfolio balance theory. Exposure changes at the fund level occur whenever foreign stocks perform differently from domestic stock investments. If foreign assets perform a lot better than domestic assets, their share in the portfolio of a fund will increase automatically, and so will the exchange rate risk associated with them. An equity fund based in France, for example, faces increased exchange-rate exposure if its euro-area investment considerably underperforms its overseas investment. Does this performance difference generate rebalancing towards euro assets? Or does the portfolio manager regard his higher investment share in dollar assets, say, as a good substitute for euro assets?
Funds do rebalance to offset valuation effects on risk
The evidence clearly supports portfolio rebalancing strategies. Our research shows that approximately 25% to 30% of the risk increase due to valuation effects is reversed by active portfolio changes within the same half-year. The rebalancing occurs following changes in exchange-rate exposure as well as equity-risk changes. International equity funds do not consider assets in different currencies to be close substitutes.6 They sell foreign assets when valuation effects lead to an increase in their portfolio share.
Bad news for sustainability of US current account deficit
This is bad news for those who think that accumulating current account deficits can be financed at constant returns to the foreign investor. In practice, foreign investors require a currency depreciation before financing new deficits.
The portfolio balance approach has other implications. Depreciation or devaluation of deficit countries’ currencies may be most likely in period of high asset price volatility or investor risk aversion. Under such circumstances foreigners might be least willing to increase their foreign asset positions. This could perhaps explain why the link between current account deficit and exchange rate depreciations has become significantly positive over the last six months.7
To the US senator questioning Krugman, the simple idea of limited international asset substitutability may be rather straightforward. Dollar assets and euro assets are different because they give returns in different currencies. By contrast, many financial economists and macroeconomists work under the convenient hypothesis that financial markets are complete or that uncovered interest parity holds. Financial institutions like equity funds should be able to swap exchange rate risk internationally.
After all, the exchange rate gain of a UK fund in US equity is the exchange rate loss of a US fund in UK equity. In such a world, rebalancing would not exist. It would be replaced by ex ante risk trading in exchange-rate derivatives. But with respect to complete global risk trading, the senator's remark might be to the point: "It is an interesting theory". This theory, however, is not supported by the data on international equity fund investment.
Equity fund managers rebalance their portfolios when home and foreign asset returns differ, and they do so along the lines predicted by portfolio balance theory. In short, there is a limit to foreigners’ holdings of US assets and thus a limit to how long the US current account deficit can remain deeply in the red. That conclusion in turn means that the US will either have to run a trade surplus in the future, or the dollar must fall to deflate the value of foreigner’s holdings.
1 See Gourinchas and Rey, 2007, "International Financial Adjustment," Journal of Political Economy, Vol. 115 (4), 665-703.
2 See Blanchard, Giavazzi and Sa, “International Investors, the US Current Account, and the Dollar”, Brookings Papers on Economic Activity 1:2005, pp. 1-49.
3 Indirect evidence for portfolio rebalancing can be inferred from the ''equity parity condition". See Hau and Rey, 2006, "Exchange Rates, Equity Prices and Capital Flows," Review of Financial Studies, Vol. 19, 273-317.
5 Hau and Rey, 2008b, "Home Bias at the Fund Level," American Economic Review P&P, Vol. 98(2), 333-338.
6 For evidence on rebalancing between risky and riskless domestic assets by Swedish households see Calvet, Campbell and Sodini, "Fight or Flight? Portfolio Rebalancing by Individual Investors," Quarterly Journal of Economics, forthcoming.
7 See "Economic Focus, The Domino Effect," The Economist, 2008, July 5-11, 81.