International Diversification: Genuine Hedge or Speculative Bet?

Jul 4
6 min read

Updated: Jul 15

What Fourteen Years of VOO, VEA, and VXUS Data Actually Show

Key Takeaways

VEA and VXUS have never risen in a year the U.S. market fell, across fourteen years of data.
VOO, VEA, and VXUS move as a single correlated cluster, not as separate pairs.
Shared channels — Fed policy, risk sentiment, dollar flows, and trade exposure — explain most of that correlation.
Correlation rises, not falls, during market crises, exactly when a hedge would need to work.
The case for holding VEA or VXUS rests on valuation and long-run structural bets, not downside protection — and even that valuation case is contested.

International diversification is one of the most repeated pieces of investing advice in existence: hold some non-U.S. stocks, the thinking goes, and your portfolio will be cushioned when the U.S. market has a bad year. It's intuitive, it's in every target-date fund prospectus, and it's rarely tested against actual data. This article tests it, using the Vanguard FTSE Developed Markets ETF (VEA), the Vanguard Total International Stock ETF (VXUS), and the Vanguard S&P 500 ETF (VOO) as the case study, and the results argue for a more honest framing of what international diversification actually does.

The belief versus the data

The table below shows the annual total return for VOO, VEA, and VXUS for every full calendar year from 2012 through 2025. VXUS launched in late January 2011, so its first full calendar year is 2012; all three funds are compared over the same fourteen-year window for consistency.

Twelve of the fourteen years show all three funds moving in the same direction. Two years, 2014 and 2015, show VOO rising while both VEA and VXUS fell, and notably, VEA and VXUS diverged from VOO in the identical two years, which is itself further evidence that the two international funds behave as one unit relative to the U.S. market. In neither divergent year did the pattern run the other way: there is no year in this fourteen-year record in which the U.S. market fell while either international fund rose.

The two down years for VOO tell a more nuanced story than a simple 'they fall together' headline. In 2018, both VEA and VXUS fell by more than VOO did. In 2022, both VEA and VXUS still fell, but by less than VOO. Both are consistent with the correlation data: all three funds moved in the same direction in both crashes, but the relative severity flipped between the two episodes, which is exactly what a 0.79 correlation predicts rather than a hedge relationship would.

What the correlation number says

The measured correlation between VEA and VOO sits at roughly 0.79, with a range of about 0.75 to 0.82 depending on the measurement window. Squaring that figure gives an R² of approximately 0.62, meaning roughly 62 percent of VEA's year-to-year movement can be statistically attributed to the same forces moving VOO, and the remaining 38 percent is independent variation. A correlation of 0.79 is meaningfully below 1.0, and it is also nowhere close to zero or negative. It describes two assets that mostly move together with different amplitude, not two assets that offset each other.

This is where the picture widens rather than staying a two-fund story. VXUS, Vanguard's broader international fund that adds emerging markets on top of VEA's developed-markets universe, correlates with VEA at 0.98 to 0.99, essentially indistinguishable. And VXUS correlates with VOO at 0.79 to 0.81, statistically the same figure as VEA's correlation with VOO. The three funds are not two independent pairs; they behave as a single correlated cluster. Layering emerging markets on top of developed markets, as VXUS does relative to VEA, does not meaningfully change the fund's relationship to the U.S. market, because VEA and VXUS are themselves near-substitutes for one another. An investor choosing between VEA and VXUS is choosing between two ways of expressing the same underlying exposure to the same three-fund correlation cluster, not choosing a materially different relationship to U.S. equities.

Why they move together

Four identifiable channels explain most of this correlation, and none of them are coincidental.

Federal Reserve policy sets a global cost of capital. Rate changes reprice the discount rate applied to future earnings everywhere, not just in the United States, because global borrowing costs key off U.S. Treasury yields to varying degrees.
Global risk sentiment operates as a single switch rather than a per-country dial. When institutional investors reduce equity exposure out of fear, they typically cut it across markets simultaneously rather than selling one country's stocks and buying another's.
Dollar flows link U.S. and international funds mechanically. A stronger dollar tends to draw capital toward U.S. assets while simultaneously compressing the reported returns of foreign holdings for U.S.-based investors, and both effects move in the same direction at once.
Trade and multinational earnings genuinely overlap. Many companies held in VEA, including major European and Japanese firms, generate substantial revenue from U.S. consumers and businesses, so a U.S. slowdown reduces their earnings directly, not just through investor sentiment.

These four channels dominate during the exact moments when a hedge would be expected to work. In April 2025, a tariff-driven selloff sent international and U.S. equities lower together. In late March 2026, a separate bout of market stress saw international funds give back a meaningful share of their earlier 2026 gains alongside the U.S. market. Correlation did not fall during either episode; if anything, it tightened.

When they actually diverge

The clearest counterexample sits in January through early March 2026, when the S&P 500 was down roughly 1.4 to 3.5 percent year-to-date while VEA and VXUS were up on the order of 10 to 11 percent over the same stretch. This was a genuine divergence, but it was not driven by fear or by a hedge mechanism. It was driven by specific, identifiable country-level catalysts: a multi-year corporate governance reform push in Japan that has driven record share buybacks and the unwinding of decades-old cross-shareholdings, a memory-chip cycle lifting South Korean semiconductor names, and a European fiscal pivot toward defense spending. These are real, structural, country-specific stories, not evidence that international stocks systematically zag when the U.S. zigs. They explain why correlation is 0.79 and not 1.0, without implying that the 0.21 of independent variation behaves like insurance.

So why invest in VEA or VXUS at all?

Once the hedge framing is set aside, several narrower but genuine reasons remain, and they apply to VXUS as much as VEA given how tightly the two move together.

A correlation below 1.0 still reduces the volatility of a combined portfolio, even when both assets typically move in the same direction, which is basic portfolio mathematics rather than a hedge claim.
VEA provides direct access to global industry leaders that are structurally difficult to reach through a U.S.-only portfolio, including companies in advanced semiconductor equipment, pharmaceuticals, and luxury goods.
International developed markets currently trade at a valuation discount to the United States, and JPMorgan projects developed international equities could return roughly 7.5 to 8.1 percent annually over the next ten to fifteen years, ahead of a 6.7 percent projection for the S&P 500, based on that valuation gap closing over time, though Fidelity's research counters that cheap international valuations have historically predicted lower, not higher, odds of subsequent outperformance.
For an investor whose income, home equity, and business are already concentrated in the U.S. economy, adding international equity reduces concentration across total wealth, even though it does little to offset a portfolio-level U.S. selloff specifically.
Periodic rebalancing between two imperfectly correlated assets produces a modest but real long-run return benefit, independent of whether either asset ever moves opposite the other.

None of these five points depends on VEA rising when VOO falls. They are arguments about valuation, access, and total-portfolio construction, and they stand on their own regardless of the correlation figure.

Naming the speculation directly

A meaningful share of the case for holding international equity is, in plain terms, a forecast rather than a measured property of these funds. The forecast takes one of two forms. The first is a bet that the four shared channels described above weaken over time: that the Federal Reserve's role as a global rate-setter diminishes, that the dollar's reserve-currency dominance erodes, or that global capital markets fragment into blocs that no longer move together. The second, more modest bet is that country-specific catalysts like Japan's governance reforms continue to outweigh the shared channels for extended stretches, as they did in early 2026. Both are reasonable positions supported by credible institutional voices, including Morgan Stanley and Bank of America on the diversification side. They are also, without exception, forecasts about the future rather than descriptions of the historical record, and the historical record described in this article shows fourteen years in which neither international fund has ever once risen in a year the U.S. market fell.

The bottom line

International diversification, as expressed through VEA or VXUS against VOO, does not behave like a hedge in any full calendar year on record, and the correlation data explains exactly why: all three funds are exposed to the same Federal Reserve policy, the same global risk sentiment, the same dollar flows, and substantially overlapping corporate earnings. Choosing VXUS over VEA for a supposedly different relationship to the U.S. market does not change this picture, since the two international funds correlate with each other at 0.98 to 0.99. What remains is a real but narrower case built on valuation, access to non-U.S. industry leadership, and long-run structural speculation about whether today's tight coupling between markets loosens over time. Investors should hold VEA or VXUS, if they hold either, for that reason, and not for a downside-protection story the data does not support.