Validating Your U.S. and Non-U.S. Equity Exposure

For most of the past 10 years, U.S. equities have meaningfully outperformed their non‑U.S. counterparts. This extended period of relative outperformance has led some investors to increase their allocations to U.S. stocks or, at least, question whether they should do so.

In this paper, we consider portfolio allocations to U.S. and non‑U.S. equities and determine what an allocation split between the two markets implies in terms of investors’ forward‑looking return expectations. We do not attempt to identify an optimal allocation between U.S. and non‑U.S. stocks but do suggest that it is important for investors to recognize the degree of forward‑looking return spread implied by their allocation decisions.

U.S. vs. Non‑U.S. Equity Performance History

The latest period of U.S. equity market outperformance has been prolonged but not unprecedented. Viewed over rolling historical 10‑year windows, we have seen multiple multiyear performance cycles for both U.S. and non‑U.S. stocks (Figure 1).

While it may be tempting to tilt a portfolio increasingly toward the outperforming region, this approach can carry considerable risk, in our view. Figure 2 compares the difference between U.S. and non‑U.S. equity returns over rolling 10‑year periods, rolled monthly, with relative returns over subsequent 10‑year periods. We see a strong negative relationship, meaning that the outperforming market historically has tended to underperform over the following decade.

What Does Your Equity Allocation Imply?

Capitalization weights in broad market indexes essentially represent investors’ consensus views about company valuation. Many investors, therefore, look to these indexes as reference points when designing their own portfolios. For example, as of December 31, 2022, the country weights in the MSCI All Country World Index suggested that a “neutral” global equity allocation would have consisted of 60% U.S. equities and 40% non‑U.S. equities.

There are reasons why investors’ global equity allocations may deviate from the market benchmark. Some investors may wish to build a home country bias into their portfolio or tactically over‑ or underweight a region based on a forward‑looking view of its relative prospects. Alternatively, some investors simply may fail to rebalance their allocations over time.

Whatever the reason, it is important to recognize that an investor’s asset allocation embeds a future return expectation. For example, an investor who allocates 100% of their equity portfolio to U.S. stocks is positioned as if they believe that the U.S. market will meaningfully outperform over their investment time horizon. We can estimate this implied forward‑looking return assumption using two pieces of information:

• An investor’s current allocation between U.S. and non‑U.S. equities and
• the historical covariance of those assets as measured by the joint variability between the two equity regions.¹

We can use this information to solve for the return assumptions that would make an investor’s portfolio optimal. Figure 3 shows the results of our analysis based on the above factors. For example, we can estimate that an investor who allocates 100% to U.S. equities implicitly assumes that U.S. stocks will outperform non‑U.S. stocks by approximately 2.5 percentage points, annualized.

A 2.5 percentage point premium may or may not be consistent with an investor’s forward-looking view. For this reason, we believe investors would be wise to compare the return differential implied by their U.S. equity allocation with other forward-looking estimates.

To do this, we compared the relative return expectation implied by a 100% U.S. equity allocation with T. Rowe Price’s own estimate—as shown in the firm’s current capital market assumptions (CMAs)—as well as with a valuation-implied premium based on the earnings differential between the Russell 3000 Index and the MSCI All Country World Index ex USA.

These comparisons revealed a significant discrepancy. As of January 2023, T. Rowe Price estimated that U.S. equities, as represented by the Russell 3000 Index, could underperform non‑U.S. equities, represented by the MSCI All Country World Index ex USA, by 1.80 percentage points annualized over the five years ending in 2027. The forward one-year valuation-implied premium for U.S. stocks was even more negative: -2.71 percentage points (Figure 4).

While this return differential is only an estimate, we believe it is important for investors to recognize when there are potential variations across the return differentials implied by their own allocations, their explicit expectations for cross‑market returns, and external return estimates.

A large spread between implied and forecasted returns could suggest that a reexamination of one’s allocation would be prudent, as there may be a disconnect between an investor’s explicit views and the implied return forecast embedded in their equity allocations.

Regional Allocations of U.S. Investors Vary Significantly

As noted, the capitalization weights in the MSCI All Country World Index suggested that a “neutral” equity portfolio would have had roughly a 60% allocation to the U.S. market as of December 31, 2022. But how were U.S. investors actually positioned?

To find out, we examined the U.S. equity allocations of two major groups of U.S. investors: wealth managers and public defined benefit (DB) plans (Figure 5). As can be seen, there was considerable dispersion in U.S. equity exposure among both investor groups. On average, wealth managers had 78% of the equity allocations in their model portfolios invested in U.S. stocks, while public DB plans averaged 67% in their plan portfolios.

Wealth managers also skewed more significantly toward a U.S. home country bias, with 76.7% of all wealth managers having 70% or more of the equity allocations in their model portfolios invested in U.S. stocks, compared with less than a third of public DB plans.

That said, there is no one “correct” level of U.S. equity exposure. The key, especially for investors with a significant U.S. equity bias, is to be sure that any tilt is intentional and well understood. Our analysis of the forward‑looking return assumptions embedded in relative U.S. versus non‑U.S. allocations highlights the potential for those assumptions to conflict with more explicit return expectations and may provide a starting point for investors who want to validate their allocation decisions.

Conclusion

Our results are not intended to prescribe a given level of U.S. or non‑U.S. equity exposure. Rather, our aim has been to highlight the forward‑looking return assumptions embedded in portfolio allocations and to advocate that investors be intentional in their portfolio positioning.

We believe that viewing the return differentials implied by allocation weights in the context of other forward‑looking return estimates (such as T. Rowe Price’s capital market assumptions) can be helpful and that a significant gap between a portfolio’s embedded view and more explicit return estimates can warrant further assessment to revalidate positioning.