The new conventional wisdom is that the return foreigners obtain on U.S. assets is less than the return U.S. residents obtain on foreign assets. And that this means that the U.S. can build up a bigger foreign debt than traditional analyses; I’ve been skeptical , . Now, we have more reason to ask how robust is the finding of a durable earnings differential in favor of U.S. investors?
A new preliminary paper (updated link 5/5) by Stephanie E. Curcuru (Fed), Tomas Dvorak (Union College) and Frank Warnock (Darden/UVA), in “The Stability of Large External Imbalances: The Role of Returns Differentials” argues to the contrary, at least for bonds and equities.
Were the U.S. to persistently earn substantially more on its foreign investments (“U.S. claims”) than foreigners earn on their U.S. investments (“U.S. liabilities”), the likelihood that the current environment of sizeable global imbalances will evolve in a benign manner increases. However, utilizing data on the actual foreign equity and bond portfolios of U.S. investors and the U.S.
equity and bond portfolios of foreign investors, we find that the returns differential of U.S. claims over U.S. liabilities is essentially zero. Moreover, were it not for the poor timing of investors from developed countries, who tend to shift their U.S. portfolios toward (or away from) equities prior to the subsequent underperformance (or strong performance) of equities, the returns
differential would be even lower. Thus, in the context of equity and bond portfolios we find no evidence that the U.S. can count on earning more on its claims than it pays on its liabilities.
How is it that Curcuru et al. obtain such different results from Gourinchas and Rey? From the introduction.
We analyze the returns differential in a manner that differs from previous analysis in at least two ways. First, we use information from actual bond and equity portfolios. Specifically, we
utilize data on the country and asset class composition of U.S. portfolio claims and liabilities. By matching precise country and asset class weights to corresponding total market returns (and by
being careful with the currency composition by asset class and country), we are able to obtain accurate estimates of the returns differential. Using data on bilateral positions also has the
advantage of allowing us to distinguish differentials vis-a-vis developed from those vis-a-vis developing countries.
The second way our study differs is methodological. Ascertaining the extent of the America’s privileged status in having a positive returns differential vis-a-vis the rest of the world
is essentially an exercise in performance evaluation. Therefore, we utilize portfolio performance evaluation techniques inspired by Grinblatt and Titman (1993). Specifically, we decompose the
return differential into its three components: composition, return, and timing effects. The first two — the composition and return effects — capture average characteristics of U.S. claims and liabilities. The composition effect would be positive were U.S. claims on foreigners weighted toward asset classes with higher average returns. The return effect would be positive were U.S. investors to earn higher average returns within each asset class. The third effect, timing, is driven by reallocations across different asset classes. The timing effect captures investment skill as given by the covariance between current weights and subsequent returns. A positive covariance between current asset weights — themselves the outcome of a buy-and-hold strategy or active trading — and subsequent asset returns would mean that portfolios were correctly positioned to capture subsequent returns. If portfolio weights and subsequent returns covary more positively in
U.S. claims than in U.S. liabilities, U.S. investors display more skill abroad than foreign investors in the U.S. and the timing effect would be positive.
Our findings can be easily summarized: Overall, we find no returns differential. For our sample (January 1994 through December 2005), the returns differential is 5.6 basis points per month but statistically insignificant. Were we to end the sample a year earlier (end-2004), the returns differential would be negative. In either case the differential is statistically indecipherable from zero.
These returns are depicted in Figures 3 and 4 from the paper.
Figure 3: US holdings of foreign portfolio assets. Source: Curcuru et al. (2007).
Figure 4: Foreign holdings of US portfolio assets. Source: Curcuru et al. (2007).
It is important to recall that these findings do not overturn the results pertaining to FDI (see the CBO report), although the authors express some skepticism about this differential as well. All this sugests that what differential exists might only pertain to a small segment of cross-border of assets and liabilities.
All this leads me to reinforce my view that we can’t let policy go on automatic pilot — the current account deficit remains a threat to economic stability.
current account deficit,