Higher U.S. Investment Yields Won't Allow Continued Large U.S. Current Account Deficit

By isolating composition, return, and timing effects, the researchers determine that essentially no positive differential exists for portfolio securities. That is to say, the positive and negative differentials that do appear from year to year balance themselves out, so that the average differential is virtually zero.

Conventional wisdom holds that the United States consistently earns more on its foreign investments than foreigners earn on their U.S. investments - that is, that the United States enjoys a positive returns differential with the rest of the world. It is further believed that such a situation contributes to overall economic stability, or what has been called a "relatively benign continuation of global imbalances." But a quite different view emerges in The Stability of Large External Imbalances: The Role of Returns Differential (NBER Working Paper No. 13074) by Stephanie Curcuru, Tomas Dvorak, and Francis Warnock.

Studying data from the Treasury International Capital Reporting System compiled between January 1994 and December 2005, a period of rapid financial globalization, Curcuru, Dvorak, and Warnock analyze the country and asset class composition of U.S. portfolio claims (U. S. investments abroad) and liabilities (foreign investments in the United States). By isolating composition, return, and timing effects, the researchers determine that essentially no positive differential exists for portfolio securities. That is to say, the positive and negative differentials that do appear from year to year balance themselves out, so that the average differential is virtually zero.

The researchers find a positive composition effect, as global equities had higher returns than bonds, U.S. investments favor equities, and U.S. liabilities are weighted toward debts. But this composition effect is offset by a negative return effect. The U.S. equity markets performed well over the last dozen years, while bond returns were essentially equal around the developed world. So, in sum, U.S. securities yielded higher returns than foreign securities. Given that the negative return effect almost exactly offsets the positive composition effect, the researchers say that it is surprising that they find any return differential at all.

Their explanation for this is that investors in foreign countries, and especially in developed countries, have shown a lack of aptitude in shifting between U. S. bonds and equities. Curcuru, Dvorak, and Warnock find that foreigners tend to have a relatively high equity weight when U.S. equity prices have already peaked and a relatively low equity weight when U.S. equity prices are about to rise. This finding strongly suggests that foreign investors have exhibited poor timing in their portfolio decisions.

For example, in 1994 and 1995, U.S. equities outperformed U.S. bonds, so the buy-and-hold weight for January 1996 was considerably higher than the actual weight from January 1994 and the actual weight for January 1996. In fact, actual equity weights are lower than the buy-and-hold weights for most of the second half of the 1990s. Putting a low weight on U.S. equity during the late 1990s, the researchers point out, proved to be a poor decision, as U.S. equities performed spectacularly during that period. When U.S. equities peaked in early 2000, foreigners' actual equity weights were higher than the buy-and-hold weights, indicating that foreign investors were buying stocks (or selling bonds), which proved profoundly unwise. The same kind of poor timing shows up yet again during 2003 and 2004. Curcuru, Dvorak, and Warnock estimate that foreign investors might well have improved their returns by 70 basis points per year had their timing in reallocating between stocks and bonds been better.

Understanding why foreign investors consistently fail to anticipate shifts in relative returns on different classes of assets, the researchers say, is an important subject for future study. Factors like the accumulation of dollar reserves, or fluctuations in foreign currency values, do not seem to apply. It is important to know if the poor timing shown by foreign investors is more of a permanent phenomenon than a temporary one. "Should foreign investors improve their timing," they write, "the U.S. external position would worsen at a faster pace. Our estimate of poor foreign timing is stable over our 12-year sample, but we have no confidence in its permanency. Increasing financial integration, cross-ownership of financial institutions, as well as improving information flows suggest that any skill advantage is unlikely to persist." This means that the returns differential experienced by the United States would no longer be insignificant, but would in fact turn negative. U.S. investors therefore could no long count on earning more on their foreign investments than they pay on their foreign liabilities.

Another area that needs analysis, the researchers add, is foreign investors' reallocations within each asset class. The current assumption is that foreigners invest in market indices for both equity and bonds; that is, their allocations within each asset class matches that of the benchmark index for each asset class. But if over time foreign investors' timing within asset classes is as poor as it seems to be between asset classes, then the true magnitude of timing and trading effects may well be underestimated.

-- Matt Nesvisky