Decomposing International Returns: What Actually Drove the Gain

An international fund posts a strong year and the fact sheet credits the manager's "disciplined stock selection process." Often the real driver was a currency swing or a lucky country bet that had nothing to do with picking better companies, and an investor who cannot tell the difference cannot judge whether the fund is worth its fee.

By Karanvir Singh, founder of Quantix Invest

The core mechanism: three sources of return

When a U.S.-based investor buys a fund holding, say, German and Japanese stocks, the return reaching their brokerage account is the product of two separate outcomes: how the underlying stocks performed in their local currency, and how that local currency moved against the dollar over the holding period. A German stock can rise 8% in euro terms and still hand a U.S. investor a loss if the euro falls far enough against the dollar in the same stretch. This is the currency effect, and it is entirely independent of anything the fund manager decided about which stocks to buy.

Layered on top of currency is the manager's actual decision-making, which itself splits into two further pieces. Country (or market) allocation captures whether the manager put more money in countries that subsequently outperformed and less in countries that lagged, a bet about macroeconomic and market-level conditions rather than individual companies. Security selection captures whether, within any given country, the manager's specific stock picks beat that country's own market average. A manager can get the country call right while picking mediocre stocks within it, or vice versa, and a fund's total outperformance can mask either pattern completely unless the return is formally taken apart.

Performance attribution is the discipline of doing exactly that: taking a portfolio's total return relative to a benchmark and assigning the difference to currency, allocation, and selection, using the portfolio's and benchmark's weights and returns in each country as the raw inputs. It answers a specific question a plain total-return number cannot: not "did the fund beat its benchmark," but "why."

It helps to think of attribution as answering a chain of increasingly specific questions rather than a single calculation. The first question is whether the fund even beat its stated benchmark once currency is handled consistently, meaning both the fund and benchmark are measured in the same currency terms. The second is whether the excess return, if any, is concentrated in a handful of countries or spread broadly, since concentrated outperformance is more likely to reflect a single macro call than a repeatable process. The third, and most diagnostic, is whether the manager's selection effect is positive within the specific countries where the fund has meaningful weight, since a positive overall selection number can still hide poor stock-picking in the fund's largest positions, offset by strong picking in smaller, less consequential ones.

The math: two worked decompositions

Worked example 1: isolating the currency effect. An investor holds a Japan-focused fund. Over the year, the fund's local-currency (yen) return is 8%. Over the same period the yen depreciates against the dollar by 6%, meaning each yen of gain converts into fewer dollars. The dollar return is not simply 8% minus 6%; the two effects compound multiplicatively: USD return = (1 + local return) x (1 + currency return) − 1 = (1.08) x (1 − 0.06) − 1 = (1.08 x 0.94) − 1 = 1.0152 − 1 = 1.52%. A local market gain of 8% arrives as a dollar return of just 1.52%, with currency depreciation consuming roughly 6.5 percentage points of the local gain. An investor reading only the fund's dollar-denominated 1.52% return, without knowing the local market actually rose 8%, would badly misjudge how the underlying companies performed.

Worked example 2: allocation, selection, and interaction in a two-country portfolio. Consider a benchmark split 60% Country A (which returns 10% for the year) and 40% Country B (which returns 4%). The benchmark's total return is (0.60 x 10%) + (0.40 x 4%) = 6.0% + 1.6% = 7.6%. A manager overweights the stronger country, holding 75% in A and 25% in B, but their stock picks within A return only 9% (slightly below A's own 10% market average, weak selection) while their picks within B return 6% (well above B's 4% average, strong selection). The portfolio's total return is (0.75 x 9%) + (0.25 x 6%) = 6.75% + 1.50% = 8.25%, beating the benchmark's 7.6% by 0.65 percentage points.

A full attribution splits that 0.65 points three ways. The allocation effect for each country is (portfolio weight − benchmark weight) x (benchmark country return − benchmark total return): for A, (0.75 − 0.60) x (10% − 7.6%) = 0.15 x 2.4% = 0.36 points; for B, (0.25 − 0.40) x (4% − 7.6%) = −0.15 x −3.6% = 0.54 points; combined allocation effect = +0.90 points, reflecting a smart overweight of the stronger country and underweight of the weaker one. The selection effect for each country is benchmark weight x (portfolio country return − benchmark country return): for A, 0.60 x (9% − 10%) = −0.60 points; for B, 0.40 x (6% − 4%) = 0.80 points; combined selection effect = +0.20 points. The remaining interaction effect, which captures the overlap between the allocation and selection bets, is (portfolio weight − benchmark weight) x (portfolio country return − benchmark country return): for A, 0.15 x −1% = −0.15; for B, −0.15 x 2% = −0.30; combined interaction = −0.45 points. Summing all three: 0.90 + 0.20 − 0.45 = 0.65 points, matching the total excess return exactly. The manager's real edge here was country allocation, not stock picking; the selection contribution was modest and nearly offset by a poor call within Country A itself.

What the evidence shows

Long-run studies of international equity return volatility consistently find that currency movements account for a large, and often the largest, share of short-term return variability in unhedged international portfolios, at times contributing more month-to-month noise than the swings in the underlying local equity markets themselves. Over multi-decade horizons the picture softens somewhat, since currencies do not trend indefinitely and tend to mean-revert around long-run levels tied to relative inflation and interest rate differentials, but over any single investor's typical holding period of a few years, currency can dominate the return an investor actually experiences.

Systematic reviews of attribution results across large samples of actively managed international funds tend to find a fairly humbling pattern: the average selection effect, the piece that reflects genuine stock-picking skill, clusters close to zero once fees are subtracted, while the allocation effect is on average mildly negative, consistent with a tendency to chase countries that have recently performed well only to see that performance mean-revert. The subset of funds that do show a persistently positive selection effect tend to be concentrated in less efficiently covered markets, frontier and small-cap international equities, where less analyst coverage leaves more genuine informational edge to be found, mirroring a pattern seen in domestic markets as well.

A separate strand of research on currency-hedged versus unhedged share classes of the same underlying fund has found that, over long horizons, the hedged and unhedged versions produce similar average returns, since currency movements roughly wash out over sufficiently long periods, but with meaningfully different volatility along the way, hedged share classes generally showing lower standard deviation. This matters because most investors do not hold a single fund for multiple decades without touching it; the volatility experienced during the actual holding period, not the theoretical long-run average, is what determines whether an investor stays invested through a rough stretch.

Research decomposing emerging-market fund returns specifically has tended to find an even larger currency contribution than in developed markets, since emerging-market currencies have historically exhibited larger and less predictable swings against the dollar, tied to differences in inflation, current account balances, and capital flow volatility that developed-market currency pairs do not experience to the same degree. This means the case for isolating currency effect before crediting manager skill is, if anything, stronger for emerging-market funds than for a developed-market fund like the Japan example above, even though the underlying attribution mechanics are identical.

Applying this in a real portfolio

For a professional building an internationally diversified portfolio, the practical use of attribution is diagnostic rather than backward-looking trivia. Before crediting a fund's outperformance to manager skill, and paying an active fee for it going forward, it is worth checking whether that outperformance survives a rough attribution: was it mostly currency, was it a country bet that will not repeat, or was it genuine stock selection within countries. A fund whose entire edge over five years traces to being overweight one strongly performing country during a period that country happened to rally is a different bet than a fund whose edge comes from consistently picking better companies within each country it holds, and only the latter has a plausible case for repeating.

The currency decision itself deserves separate, deliberate attention rather than being an accidental byproduct of which fund an investor happens to buy. A U.S. dollar-based investor holding unhedged international equities is making an implicit bet on the dollar's direction in addition to a bet on foreign equity markets, whether or not that was the intent. Some investors, particularly those already carrying significant foreign currency exposure through international business income or foreign-denominated debt, may prefer currency-hedged international funds to avoid stacking correlated currency risk on top of an existing exposure; others, seeking currency diversification precisely as a hedge against a weakening dollar eroding domestic purchasing power, may prefer to leave the currency exposure unhedged on purpose. Either choice is defensible, but it should be a choice, not a default inherited from whichever share class happened to be recommended.

It is also worth checking a fund's attribution consistency across multiple years rather than a single strong period. A manager whose selection effect is positive in most years, even modestly, across different market regimes offers a more credible case for repeatable skill than one whose single standout year came entirely from an allocation call that happened to pay off, since allocation bets on entire countries are closer to a macroeconomic forecast than to fundamental company research, and macro forecasting has a considerably weaker track record of persistence across the investment industry broadly.

A related, and often overlooked, practical detail is how a fund's benchmark itself is constructed, since attribution results are only as meaningful as the benchmark comparison underlying them. Two international funds can both claim to beat "the international benchmark" while using benchmarks with meaningfully different country weights, one closer to a broad global index and another tilted toward developed markets only, making a direct comparison between the two funds' attribution results misleading unless the benchmarks are first reconciled to a common standard. A careful investor checks not just the attribution numbers but the specific benchmark index cited, including its country and sector weighting methodology, before comparing one fund's reported skill against another's.

Actionable breakdown

• Reading a fund's stated outperformance
  • Ask whether returns are reported in local or home currency.
  • Check the fund's currency hedging policy in its prospectus.
  • Compare a hedged and unhedged share class if both exist.
• Judging the source of past outperformance
  • Look for multi-year selection-effect consistency, not one year.
  • Discount outperformance concentrated in a single country bet.
  • Weight selection effect more heavily than allocation effect.
• Deciding on currency exposure deliberately
  • Consider existing foreign currency exposure before choosing.
  • Match hedging choice to your holding period and goals.
  • Revisit the choice only on a schedule, not after big currency moves.

Common pitfalls

Crediting skill for a currency tailwind: a fund that outperformed during a period of dollar weakness may simply have benefited from unhedged currency exposure, not analysis.

Judging a manager on one standout year: a single well-timed country bet can dominate a multi-year track record without indicating any repeatable edge.

Ignoring the currency decision entirely: defaulting into whichever share class a platform offers means accepting an unexamined currency bet.

Comparing hedged and unhedged funds head to head: without adjusting for currency, this comparison conflates a currency call with a stock-picking call.

The bottom line

Before paying for international outperformance, decompose it into currency, allocation, and selection, because only the last one reflects a skill worth paying for repeatedly.

All articles · Performance attribution procedures · Risk factors in international investing · International diversification potential · International investing guide