Since 1900, shares have beaten bonds and cash in every market with an unbroken record — but nobody could have picked the winning country, and two markets went to zero

Most arguments about investing abroad are made from the last ten or twenty years, which is too short to separate a pattern from a run of luck. The longest consistent record is the database built by Elroy Dimson, Paul Marsh and Mike Staunton of London Business School and Cambridge, published each year as the UBS Global Investment Returns Yearbook: stocks, bonds, bills, inflation and currencies in 35 markets, most of them from 1900¹. This article reads what that record says about spreading money across countries. The companion piece, home bias and currency risk, covers why investors over-own their home market and how to measure in the currency you will spend; this one is about the history.

The headline record

Shares have rewarded patience almost everywhere. In every one of the 21 Yearbook markets with a continuous investment history, equities were the best-performing asset since 1900, and bonds beat cash everywhere except Portugal¹. In the US, one dollar put into shares in 1900 had grown to $124,854 by the end of 2025, against $284 in long-term government bonds and $69 in Treasury bills, while prices rose 38-fold¹. Dividing out that inflation, US shares returned roughly 6.6% a year in real terms over 126 years¹.

The ride was rough. Since 1900, equities and bonds have each lost more than 70% in real terms on several occasions, whereas a 60:40 blend of the two never fell more than 50%¹. In the US, recovering from the 1929 crash took until February 1945 — fifteen and a half years — and shares bought before the 1973–74 bear market were under water in real terms for more than ten years³.

The spread between countries

The average hides an enormous range. The table uses the authors’ own published figures for the 17 markets with complete records over 1900–2005, from a paper that sets out each country in detail; later Yearbooks extend the same series to 2025.

Two things stand out. First, an investor in 1900 had no way of knowing which column they were in: the countries at the bottom included some of the richest economies of the day. Second, the losses that matter most were not dips but decades. German shares lost 88% in real terms over 1939–48 and Japanese shares 96%, and investors in Germany, France and Japan waited more than half a century before their real returns turned positive from a 1900 start².

Over the same 106 years, the world index returned 5.75% a year in real terms with a standard deviation of 17.2%, lower than every single market except Canada; the US returned 6.52% with 20.2%, and Germany 3.09% with 32.5%². A spread of countries did not deliver the best market’s return, but it delivered most of the average return with a good deal less of the worst.

Markets that went to zero

Long-run studies are sometimes accused of looking only at survivors. The authors addressed it directly. Russia was the largest market missing from their early database, at just under 5% of world equity value in 1900, and China about 0.4%; shareholders in both lost everything when companies were nationalised after the revolutions of 1917 and 1949². Exchanges in Czechoslovakia, Hungary and Poland also ended in disaster after the Second World War².

Their estimate is that the markets left out made up about 10% of world equity value in 1900, and that leaving them out overstates the long-run world equity premium by at most about a tenth of a percentage point a year². For the world as a whole, the bias is small. For an investor whose whole portfolio sat in one of those markets, it was total. That asymmetry is the plainest argument for not holding everything in one country.

The US century, and the risk of extrapolating it

The US is the best-documented market and, today, by far the largest: about 62% of world equity value at the end of 2025, up from a far more evenly spread market in 1900, partly because of strong returns and partly because of heavy share issuance¹. It was not the obvious choice then. In 1900 London was the world’s leading financial centre, and the British equity market was worth about 50% more than the New York Stock Exchange⁴. The US record is excellent but not the best: over 1900–2005, Sweden, Australia and South Africa all beat it in real terms².

The more recent evidence cuts both ways. Over the 50 years to 2024, investing globally rather than at home gave a better return for the risk taken in the vast majority of countries — with one prominent exception, the US, whose investors would have done better staying at home. The authors call this a cautionary tale: good decisions, made on sensible grounds, can still have disappointing outcomes³.

Two cautions follow. The market that looks safest is the one that has done best, which is exactly how the largest markets of 1900 looked. And concentration is now unusually high: by the end of 2025, US equity market concentration was at its highest level for at least 100 years¹. Industries change too — railways were 63% of the US market in 1900 and are now under 1%, and some 80% of the value of US firms listed in 1900 was in industries that are small or extinct today¹.

Emerging markets and India

Emerging markets are often sold as the higher-return option. Over the full period the record says the opposite: developed markets returned 8.5% a year from 1900 to 2025 against 6.9% for emerging markets. From 1960 the order reversed, at 10.9% a year for emerging markets against 9.6% for developed ones¹. These are composite indexes; for an investor who spends in another currency, the exchange rate is also part of the result.

India illustrates both halves. It had a stock market in 1900, but in 1913 it belonged to a group of small markets — with Egypt, Finland, Greece, Hong Kong, New Zealand and Sri Lanka — that together made up less than 1% of world equity value, and a reliable return series from that era was not then available². It is now one of the 32 markets in the Yearbook’s world index¹. The lesson for an Indian investor is not that India will repeat the US century or will not; it is that India today is a large economy but a small share of the world’s listed companies, and the same history that rewarded patience punished concentration.

Correlation and currency

Diversification works only when things do not all fall together. Over the very long run, the correlation between shares and bonds has averaged 0.33 across countries and 0.19 in the US; it was mostly negative from the late 1990s until 2021, a valuable cushion, and that era ended in 2022 when both fell sharply together³. Correlations between developed and emerging markets, and between shares and bonds, have risen recently, which makes diversification harder but has not removed its benefit¹.

Between countries, the same trend shows up. Vanguard measured the ten-year correlation between US and non-US shares rising from 0.51 at the end of 1989 to 0.86 by September 2020, before levelling off⁵. A high correlation still leaves room for long stretches in which one region leads: non-US shares led in the mid-1980s and for most of the 2000s, and US shares for most of the 2010s⁵. A diversified investor owns whichever region is leading, without having to know in advance which one it will be.

The authors add that, since 1900, most large peacetime falls in shares were set off by economic causes rather than geopolitical ones⁶ — which is why spreading across economies, not only across headlines, is what reduces the damage.

Currency is the second layer. On average, exchange-rate movements added around 6 percentage points to the volatility a US-dollar investor experienced in a foreign market, whether in shares or bonds¹. When that currency risk was hedged, investors in the vast majority of markets were better off investing globally than domestically over 1974–2025¹. How much exchange-rate risk to carry depends on where you will spend — the subject of the home bias article.

What a global investor gets — and does not

• Gets: protection from the one outcome that history shows is unrecoverable — a single market collapsing for decades or for good.
• Gets: a smoother path. The world index was steadier than all but one of the 17 national markets over 1900–2005².
• Does not get: the best market’s return. A spread portfolio will always trail whichever country turns out to win.
• Does not get: protection in a global crash. In 1929–31 the world index fell 54% in real terms and in 2000–02 it fell 44%²; only bonds and cash cushioned those years.
• Takes on: currency movements, which can help or hurt and are worth matching to where the money will be spent.

None of this sets a right share abroad for any household; that depends on costs, taxes, the currency you will spend in and how soon you need the money. It is information about the historical record, not financial advice. For how the mix between shares and safer assets usually shifts as a horizon shortens, see asset allocation by age.