Did Stocks Double During the Civil War? The Index Rose 129%. Purchasing Power Rose 24%.

Yes, and by more than the usual telling. The best reconstruction of nineteenth-century New York Stock Exchange prices, assembled by Goetzmann, Ibbotson and Peng from month-end newspaper quotations, has the index at 0.73 at the end of 1861 and 1.68 at the end of 1865. That is a gain of 129% across the four calendar years of the war.¹ Over the same window, consumer prices rose 85%.² Deflating one by the other leaves a real price gain of about 24%.

That gap between 129% and 24% is the whole subject of this guide, and it is not a curiosity of reconstructed 1860s data. A century later, with data nobody has to reconstruct: between the Gulf of Tonkin resolution and the fall of Saigon, the S&P 500 returned 81.2% including dividends, while consumer prices rose 79.6%. The real total return over those twelve years was 0.9%.¹⁸ Not 0.9% a year. Nine-tenths of one percent, in total, over twelve years.

Before going further, one thing should be said plainly and then not repeated: an index return measures the market value of claims on surviving listed companies. It is not a measure of national welfare, lives, or destruction, and nothing below should be read as suggesting any war was good for anyone.

What the Civil War numbers actually show

The index behind the 129% figure is price-weighted and built from at least two month-end prices for each of more than 600 securities, taken from The New York Shipping List and later the New York papers. Its authors are explicit that their equal-weighted variant is badly upward-biased, so the price-weighted series is the one to use.¹ Year by year, the war looks like this: 1861 −3.7%, 1862 +49.2%, 1863 +41.0%, 1864 +10.5%, 1865 −1.3%. The peak was the end of 1864, not the end of the war.

Much of that rise was the measuring stick shrinking. The Union suspended gold convertibility and issued paper “greenbacks” to help pay for the war. Gold’s official price was $20.67 an ounce; on July 11, 1864 it was quoted at 285 greenbacks per 100 gold, which is $58.91 an ounce, with a greenback dollar worth about 35 gold cents. By April 1865 gold had recovered to around 150, still a 50% premium to its official price.³ A share price quoted in a currency that had lost a third of its value against gold was going to rise whether or not the underlying business improved.

Two details are worth correcting, because both circulate widely. The New York Stock Exchange did not close during the Civil War. What happened on May 11, 1861 was a ban on trading Confederate securities, which is a listing restriction. The exchange’s long shutdown came later, in 1914.¹ And the market of the 1860s was not a diversified economy in miniature: transportation, chiefly railroads, made up 30 of the 56 securities in the index in 1861 and 37 of 82 in 1865, with banks accounting for most of the rest. Mining went from 3 securities to 17 at the 1864 peak, a wartime speculative boom of its own. Those are counts rather than weights; no capitalization weights exist for this period.

What actually paid investors was dividends

Here the story turns, and in a direction the popular version misses entirely. Nineteenth-century American companies paid out most of their earnings, and most shares traded near a $100 par value, so capital appreciation was small and income was large. Over 1815 to 1870, the same researchers put price appreciation at 0.84% a year against a total return of 4.72% a year on their conservative dividend estimate. Dividends were the overwhelming majority of the return, and among known regular payers the yield averaged 9% to 11% in an era when high-grade corporate bonds yielded 5% to 7%.¹

Applying their own low and high dividend estimates to the war years gives a nominal total return of roughly 160% to 195%, and a real total return of roughly 41% to 60%. So the honest answer to the question is layered: the price index rose 129% and bought 24% more goods, and shareholders nonetheless did meaningfully well, because they were collecting 4% to 8% a year in cash the whole time. The headline doubling was mostly the currency. The actual payment was the dividend.

The inflation-adjusted figures in the last two paragraphs are arithmetic performed here on two published series, not numbers published by anyone else.

What the Confederate side was pricing

The most interesting evidence from this period is not about stocks at all. Confederate gold bonds traded in Amsterdam, and their prices imply what European investors thought was going to happen. Weidenmier and Oosterlinck estimate those prices carried roughly a 42% probability of Confederate victory before Gettysburg and Vicksburg, falling to about 15% by the end of 1863.⁴ That is a functioning prediction market, and it is doing exactly what markets do: pricing a distribution of outcomes and revising it on news.

Which events counted is itself instructive. Willard, Guinnane and Rosen tested the daily greenback series for statistically significant breaks and found exactly seven across the war. Gettysburg and Vicksburg are among them. The largest single move of the entire war was not either of those: it was July 12, 1864, when Jubal Early’s army retreated from the outskirts of Washington, at nearly double Gettysburg’s effect. Several events that historians treat as decisive produced no measurable break at all.⁵ The fall of Atlanta, often listed as a market-moving event, does not appear in the greenback results; it shows up instead in London-traded Confederate cotton bonds.⁶

And one result that should complicate anyone’s intuition: Confederate cotton bond prices more than doubled between December 1863 and September 1864, after Gettysburg and Vicksburg, because the bonds were convertible into cotton and the cotton price was rising.⁷ A security can appreciate while the entity that issued it is losing a war. Security prices and national outcomes are related, but they are not the same measurement.

The same illusion, with data nobody has to reconstruct

The Civil War figures rest on a careful reconstruction, which means reasonable people can argue about them. Vietnam does not have that problem. Using annual total returns and CPI for 1928 onward:

The nominal column overstates the real column in every row, by 10.8 points for the Gulf War and 200.9 points for Afghanistan. Vietnam is the extreme case, and it is worth sitting with: an investor who held through the entire American involvement, reinvesting every dividend, ended with essentially the purchasing power they started with. A chart of the nominal index over those years slopes cheerfully upward.

Every war table depends on dates somebody else chose

Now hold the war constant and vary only the endpoints. World War II, S&P 500, real total return:

Same war, same index, same source. The answer ranges from a doubling to almost nothing, and every one of those rows is defensible. An author who wants the war to look good starts in 1942. An author who wants it to look bad starts in 1937 and runs through 1948. Neither is lying, and that is the problem with the format.

The bill arrives after the armistice

The 1939–1948 row is low for a specific reason. In 1946, the first full year of peace, the S&P 500 returned −8.4% while consumer prices rose 18.1%, for a real return of −22.5%. Price controls came off, wartime saving met civilian production that had not yet restarted, and the purchasing power that had been borrowed during the war was repaid by savers.

The same pattern appears at global scale after the First World War. Global Financial Data’s reconstruction has the world index falling less than 6% between 1912 and the armistice in November 1918, then 36% over the next three years. Measured from August 1912 to July 1921 the index fell 40% in nominal terms and 69% after inflation, which their analyst describes as worse than the 1929 to 1932 decline.⁸ A war-return table that stops at the armistice stops one year before the story.

Bonds did not hedge the wars that mattered

“Hold bonds for geopolitical risk” is common advice. Over the same windows, in real total return:

Long Treasuries lost real value through World War II, Korea, Vietnam and the Ukraine window. Those are precisely the inflationary conflicts, which is the point: nominal bonds hedge a growth scare, and a war that impairs supply is not primarily a growth scare. The Ukraine column is the cleanest modern illustration, with bonds down 22.2% in real terms while equities gained.

There is a mechanism behind that. Work at the Bank for International Settlements building a geopolitical risk index from European rather than American news sources finds that shocks to it are both recessionary and inflationary, which is the combination nominal bonds handle worst, though the estimated effect of any single shock is small.¹⁷ How a war is financed matters as much as how it is fought. Le Bris, studying French markets from 1870 to 1945, found equity behaviour depended materially on the financing method, with the closed and financially repressed World War II economy producing a paradoxical and short-lived rise in nominal share prices.¹⁵ Taxes, borrowing, money creation and financial repression divide the cost of a war among citizens in different proportions, and a bondholder can receive every promised payment and still be the one who paid.

Gold looks like the hero of this table and mostly is not. Its two huge numbers sit in the Afghanistan and Iraq windows, which overlap the 2000s gold bull market and a currency story that had little to do with either war, and it lost 19% in real terms across the Gulf War. Two of these windows also contain the dot-com collapse and the 2008 financial crisis, which is a general warning about this whole exercise: wars rarely happen in isolation, and attributing a decade of returns to the war inside it is not credible. The inflation-hedge guide covers what gold has and has not done across inflationary periods.

Volatility during World War II was lower than average

This is the finding that surprises people most, and it holds up. Using daily US market total returns from 1926 to 2026, annualized volatility across the full century is 17.1%. Between Pearl Harbor and VJ Day it was 10.3%. Korea ran at 10.0% and the Vietnam window at 12.5%. Pooling every war day in the sample gives 16.4% against 17.6% for every other day.¹⁸

This replicates published work. Cortes, Vossmeyer and Weidenmier find US stock volatility roughly 25% lower during wartime across 1890 to 2017, and attribute a large part of it to the military demand channel: government procurement is large, contracted in advance, and predictable, which makes the cash flows of exposed companies easier to forecast even while everything else is uncertain.⁹

Three limits, stated immediately. This describes a distant, victorious belligerent whose markets stayed open and whose territory was never invaded; it generalizes to nobody else. Realized volatility of listed share prices does not measure national risk, and the years in question were not calm. And in the same dataset the post-2001 windows show higher volatility than average, because they contain the dot-com bust and the financial crisis. That window is measuring a credit crisis, not a war.

The decline usually happens before the war starts

Here is the mechanism that explains most of the confusion in this subject. Using exact dates rather than calendar years:

In all three cases the market fell while the war was becoming likely and rose once it had started. A table that begins on the day of the first air strike reports +17.8% for the Gulf War and +149.4% for the Second World War. That is not a lie about the data; it is a claim about when the relevant event happened, and it is usually wrong.

Attempts to map war chronologies onto returns more systematically tend to come back empty. Hudson and Urquhart tested the British market across the major events of the Second World War and found limited evidence of strong links between those events and returns, with only some support for bad news mattering more than good.¹⁶ Events that fill the history books are not necessarily the events that changed what investors expected.

The reason is that prices reflect expectations, so what moves them is the difference between what happens and what was already assumed. Suppose investors think a market is worth 100 under peace, 90 under a limited war, and 50 under a catastrophic one, and they assign those outcomes probabilities of 50%, 30% and 20%. The market trades at 87. War then breaks out and is visibly limited, so the price moves toward 90. Prices rose, on unambiguously bad news, because the news was better than the distribution the price already contained. Nothing about that requires the war to have been good for anyone. The stock market and the economy makes the same argument for recessions.

The wars that are missing from the table

Every table in this guide shares a defect it cannot fix: it contains only markets that kept publishing prices. When the First World War began in 1914, every major stock exchange in the world closed, some for months as in the United States, some for years as in Germany and Russia.⁸ The New York Stock Exchange shut from July to December 1914, the longest closure in its history and far longer than anything the Civil War produced.¹ There is no continuous American series through the outbreak of that war, because for five months there was no market.

A closed market is a different failure from a falling one. A quoted price you cannot transact at is not a price, and an investor who cannot sell, convert, or repatriate does not hold what the index says they hold. Investors in Russia after 1917 did not experience a bad return; they experienced the end of the asset class. Stocks Are Always Risky covers those cases, and Optimists, But Not Naive covers why series that begin after the turmoil make history look tamer than it was.

Distance matters enormously, and it is measurable. Studying 66 economies in a four-week window around the 2022 invasion of Ukraine, Federle, Meier, Müller and Sehn find that a country bordering the conflict suffered a 23.1% equity decline, an effect that fades by about 2.6 percentage points per 1,000 kilometres of distance. Ukraine’s fifteen near neighbours fell more than 20% on average while the other 51 markets fell less than 6%. Controlling for trade links cuts the distance effect to about 1.0 point per 1,000 km, leaving a residual the authors attribute to the risk of being drawn in.¹⁰ Research covering more than 150 wars since 1870 points the same way: real GDP in the country where the fighting happens falls around 30% within five years, with costs declining as distance grows.¹¹

A global index can therefore look almost undisturbed while investors in one region are destroyed. Both things are true at once, and only one of them appears in a chart of world equities.

Run the numbers yourself

The tool below holds annual returns and inflation for 1928 to 2025. Pick a conflict, then drag the start and end years and watch how much of the headline was the deflator and how much was the date range.

Does buying defense stocks work?

The cleanest available evidence is a pair of event studies on the same set of large arms manufacturers. Around the February 2022 Russian invasion, they averaged about 10 percentage points of cumulative abnormal return. Around the October 2023 Hamas attack on Israel, abnormal returns were indistinguishable from zero.¹²

The authors call this paradoxical, and it is worth seeing why. The Hamas attack was a genuine surprise and moved defense stocks not at all; the Russian invasion had been visibly building for months and moved them a great deal. Their explanation is the policy response: European governments committed to rearmament after February 2022, and no comparable procurement shift followed October 2023. What the shares were pricing was future government purchasing, which is not readable from the fact that a war has started.

The proximity study points the same way from another angle: aerospace and defense companies listed in Ukraine’s neighbouring countries rose sharply during the invasion window even as their home markets sold off, while defense companies elsewhere showed no such gain.¹⁰ So the trade required knowing not just that war was coming, but which governments would respond with money, and which listed companies would capture it. No study establishes a tested, cost-adjusted rule that maps public conflict news to defense-sector returns, and the two events above would have produced opposite outcomes from the same signal.

Where the evidence disagrees with itself

Honest reporting requires noting that the research does not speak with one voice on predictability. Hirshleifer, Mai and Pukthuanthong analysed roughly seven million New York Times articles across 160 years and found that elevated war discourse positively predicts subsequent market excess returns, with an out-of-sample R² of 1.35%, which they call a war return premium.¹³ Dimson, Marsh and Staunton, working with 126 years of global data, report no relationship between geopolitical threat indices and equity returns one month or one year ahead, and conclude that economic risk has historically mattered more to investors than geopolitical risk.¹⁴

Both can be defended. They use different measures, samples and horizons, and an out-of-sample R² of 1.35% is a real statistical result that is also a thin foundation for a personal trading strategy after costs and taxes. What the two agree on is more useful than where they differ: the same authors find that World War I, World War II and the 1973–74 oil shock produced three of the six worst episodes for large global equity markets since 1900. Wars can be economically ruinous for investors. The difficulty is knowing in advance which one is which, and whether the price already says so.

What this means for a DIY portfolio

Six habits follow from all of this, and none of them requires a view on any conflict.

• Judge everything in real, after-tax terms. This is the single highest-value habit the guide can leave you with. Vietnam returned 81% and bought nothing.
• Hold a globally diversified portfolio. The proximity evidence is an argument for it: geography determined outcomes in 2022 more than almost anything else, and you do not know which geography will matter next. The global diversification guide covers the case, including its limits.
• Size liquidity around your spending. Cash and short-term bonds should cover near-term needs, job security and known liabilities. Raise that number when your circumstances change, rather than when the news gets worse.
• Rebalance on a written rule. If a shock pushes an 80/20 portfolio to 74/26, a policy band gets you buying at lower prices without requiring a forecast about when the conflict ends.
• Avoid leverage into this kind of risk. War risk includes gaps, closures and capital controls, which is exactly the environment where leverage converts a temporary decline into a permanent loss.
• Do not sell because the news is horrifying. The severity of an event and the amount of unpriced information in it are different quantities. Missing the Best Days covers what exiting tends to cost.

Who should adjust, and who should not

Some investors have real, non-diversifiable exposure and should act on it. A household living in or near a threatened country reasonably prioritises emergency liquidity, accounts at more than one institution, some foreign-currency exposure and up-to-date documents. That is resilience planning rather than market timing. A business owner whose revenue, property and portfolio all sit in the same country or industry has a genuine concentration to fix. Someone spending from a portfolio within a few years may discover in a crisis that their allocation was always too aggressive, but the problem there is the mismatch, not the war.

Trading around wars is a poor idea for anyone without a written exit rule, anyone using margin, anyone reacting mainly to social media, anyone who will need the money soon, anyone already concentrated in the affected country or sector, and anyone who cannot distinguish a nominal return from a real one. Wars generate intense moral clarity and very little tradable information, and the two feel similar from the inside.

Frequently asked questions

Did the stock market really double during the Civil War?

The best available price index rose about 129% between the end of 1861 and the end of 1865, so it more than doubled in greenbacks. Consumer prices rose 85% over the same span, leaving a real price gain near 24%. Including dividends, which were the bulk of nineteenth-century returns, the real total return was roughly 41% to 60%.

Do stocks usually go up during wars?

For the United States, nominal returns have usually been positive, and that is the least informative way to state it. Real returns have ranged from strongly positive in World War II to essentially zero across Vietnam. For countries where the fighting happened, outcomes have included closed exchanges, expropriation and total loss. There is no stable wartime return because wars are not one kind of event.

Why do stocks sometimes rise when a war breaks out?

Because the price already contained a probability-weighted set of outcomes, and the outbreak often resolves uncertainty or turns out narrower than the worst case investors were pricing. The market fell 10% during the Gulf War buildup and rose 18% after the air war began. The decline had already happened.

Are bonds a good hedge against geopolitical risk?

Against a growth scare, high-quality government bonds often help. Against a war that impairs supply and raises inflation, they have not. Real 10-year Treasury returns were negative across World War II, Korea, Vietnam and the 2022 to 2025 window.

Should I buy defense stocks when a conflict starts?

The evidence does not support a rule. The anticipated 2022 invasion moved large arms manufacturers about 10 percentage points; the surprise 2023 Hamas attack moved them approximately zero. The difference tracked which governments committed new procurement money, which the headline does not tell you. A concentrated sector bet also carries valuation, contract and political risk regardless of how much a government spends.

Does a globally diversified portfolio protect me from a world war?

Not fully, and it is important to say so. Diversification spreads country, currency and political risk, which is precisely what the 2022 proximity evidence shows mattering. It does not help against a genuinely global correlated catastrophe. It lowers the probability of regret more than it lowers the probability of loss.

Why does the stock market rise while a country is being destroyed?

Because an index measures the market value of claims on surviving listed firms, and those firms can receive government contracts, benefit from scarcity, or simply be priced in a currency that is losing value. Real wages, consumption, infrastructure and lives are not inputs to that calculation. The two quantities are measuring different things and can move in opposite directions without either being wrong.

Key takeaways

• The Civil War doubling was mostly the currency. The index rose 129% while prices rose 85%, leaving about 24% of real price appreciation. Dividends did the actual work.
• Vietnam repeated it with clean data. +81.2% nominal, +0.9% real, across twelve years.
• Endpoints decide the answer. World War II gives +105% or +10% real depending only on which years you count.
• The market usually falls before the war starts. The buildup is where the decline happened in 1990, 2022 and 1941 to 1942.
• Nominal bonds did not hedge the inflationary wars. They lost real value in four of the seven windows here.
• Distance to the fighting dominated 2022 outcomes. A bordering country fell 23.1%; the effect faded by 2.6 points per 1,000 kilometres.
• Build resilience rather than forecasts. Diversify globally, size liquidity to spending, rebalance on rules, avoid leverage, and measure in real terms.

Related guides

• What Real Return Should You Assume for Stocks? for the return assumptions that should replace nominal history in a plan.
• Stocks Are Always Risky for the markets that went to zero, and why the surviving record flatters equities.
• The Four Deep Risks for inflation, deflation, confiscation and devastation as separate hazards.
• Optimists, But Not Naive on why data series that begin after the turmoil understate how bad bad can get.
• Missing the Best Days for what selling into frightening news has historically cost.
• Inflation Is the Greatest Enemy of Retirees for what an inflationary decade does to a portfolio being spent from.
• Do 200 Years of Stock Returns Still Matter? for how the pre-1926 record was reconstructed and what it changes.

Sources and method

1. Goetzmann, William N., Roger G. Ibbotson, and Liang Peng. “A new historical database for the NYSE 1815 to 1925: Performance and predictability.” Journal of Financial Markets 4, no. 1 (2001): 1–32. Source of the price-weighted index levels quoted here (0.73 at year-end 1861, 1.71 at year-end 1864, 1.68 at year-end 1865), the low and high dividend estimates, the 0.84% and 4.72% annual figures for 1815–1870, the security counts by industry, and the statement that the NYSE was open during the Civil War and closed from July to December 1914. Working paper PDF. Separately, the phrase that Civil War share prices got “a big lift” comes from Goetzmann’s personal Yale page rather than from this paper, which contains no Civil War narrative; it is not quoted here for that reason.
2. Officer, Lawrence H., and Samuel H. Williamson. “The Annual Consumer Price Index for the United States, 1774–Present.” MeasuringWorth. 1861 = 8.540 and 1865 = 15.790 on the 1982–84 = 100 base, giving 84.9% cumulative inflation; confirmed directly against the dataset. Other reconstructions of 1860s prices differ and should not be mixed with this one. measuringworth.com
3. Mitchell, Wesley C. Gold, Prices, and Wages under the Greenback Standard. University of California, 1908. Records the greenback’s maximum depreciation on July 11, 1864, with gold quoted at 285 and the greenback worth 35.09 gold cents. The $20.67 official price was fixed by the Coinage Act of 1834. Full text
4. Weidenmier, Marc D., and Kim Oosterlinck. “Victory or Repudiation? Predicting Confederate Bond Prices from the Civil War.” NBER Working Paper 13567 (2007). Confederate gold bonds traded in Amsterdam imply roughly a 42% probability of Confederate victory before Gettysburg and Vicksburg, falling to about 15% by the end of 1863. nber.org
5. Willard, Kristen L., Timothy W. Guinnane, and Harvey S. Rosen. “Turning Points in the Civil War: Views from the Greenback Market.” American Economic Review 86, no. 4 (1996): 1001–1018. Identifies seven statistically significant breaks in the wartime greenback series, the largest being July 12, 1864. The underlying daily data are public. Greenback series data
6. Brown, William O., Jr., and Richard C. K. Burdekin. “Turning Points in the U.S. Civil War: A British Perspective.” The Journal of Economic History 60, no. 1 (2000): 216–231. Finds two turning points in London-traded Confederate cotton bonds: Gettysburg and Vicksburg, and the fall of Atlanta.
7. Weidenmier, Marc D. “The Market for Confederate Cotton Bonds.” Explorations in Economic History 37, no. 1 (2000): 76–97. Documents cotton bond prices more than doubling between December 1863 and September 1864. See also Weidenmier, “Turning Points in the U.S. Civil War: Views from the Grayback Market,” Southern Economic Journal 68, no. 4 (2002): 875–890, which finds Confederate currency turned on different events from the Northern greenback.
8. Taylor, Bryan. “Wars and Financial Panics: Global Bear Markets in the Twentieth Century.” Finaeon / Global Financial Data. Source of the statement that every major stock exchange in the world closed in 1914, and of the figures that the world index fell less than 6% between 1912 and the November 1918 armistice, 36% over the following three years, and 40% nominal or 69% real between August 1912 and July 1921. This is proprietary vendor research on a reconstructed index rather than peer-reviewed work, and the page carries no publication date. finaeon.com
9. Cortes, Gustavo S., Angela Vossmeyer, and Marc D. Weidenmier. “Stock Volatility and the War Puzzle: The Military Demand Channel.” NBER Working Paper 29837, revised May 2024. Reports US stock volatility about 25% lower during wartime over 1890–2017. An earlier version of the same paper reported 33%, a figure still widely repeated in secondary coverage. Unpublished working paper. nber.org
10. Federle, Jonathan, André Meier, Gernot J. Müller, and Victor Sehn. “Proximity to War: The stock market response to the Russian invasion of Ukraine.” CEPR Discussion Paper 17185 (2022); published in the Journal of Money, Credit and Banking 58, no. 3 (2026). The 23.1% figure, the 2.6 points per 1,000 km decay, the 1.0 point residual after controlling for trade, the 66-economy sample and the 10 February to 10 March 2022 window, and the finding on defense companies in neighbouring countries, are taken from the authors’ own summary of the discussion paper; published figures may differ slightly. Authors’ summary
11. Federle, Jonathan, André Meier, Gernot J. Müller, Willi Mutschler, and Moritz Schularick. “The Price of War.” Kiel Policy Brief 171, Kiel Institute for the World Economy, February 2024. Based on more than 150 wars since 1870; real GDP in the war theatre falls around 30% within five years of onset, with costs declining with distance. A policy brief rather than a peer-reviewed article.
12. Kaja, Fatjon, and Kevin Foster. “The market dynamics of military conflict: financial returns and strategic considerations in the global arms industry.” Defence and Peace Economics 37, no. 4 (2025). From the abstract: “Arms companies averaged 10 percentage points of cumulative abnormal returns (CAR) after the Russian invasion, but after the Hamas attack, CAR was indistinguishable from zero,” a result the authors describe as paradoxical and attribute to “the different reaction of policymakers.” tandfonline.com
13. Hirshleifer, David, Dat Mai, and Kuntara Pukthuanthong. “War Discourse and Disaster Premium: 160 Years of Evidence from the Stock Market.” The Review of Financial Studies 38, no. 2 (2025): 457–506. Roughly seven million New York Times articles; the war topic predicts market excess returns with an out-of-sample R² of 1.35%. A related NBER working paper covers bonds and a separate 2025 Journal of Finance paper covers the cross section.
14. Dimson, Elroy, Paul Marsh, and Mike Staunton. UBS Global Investment Returns Yearbook 2026, public summary edition. Source of the finding that World War I, World War II and the 1973–74 oil shock produced three of the six worst episodes for large global equity markets since 1900, and of the finding of no relationship between geopolitical threat indices and equity returns one month or one year ahead. The full Yearbook is restricted to UBS clients, so no specific war-period return figures from it are quoted here. Public summary
15. Le Bris, David. “Wars, inflation and stock market returns in France, 1870–1945.” Financial History Review 19, no. 3 (2012): 337–361. Concludes that equity behaviour depended materially on how each war was financed, with the closed and financially repressed World War II economy producing a paradoxical short-lived rise in nominal stock prices. A published erratum corrects a figure caption only.
16. Hudson, Robert, and Andrew Urquhart. “War and stock markets: The effect of World War Two on the British stock market.” International Review of Financial Analysis 40C (2015): 166–177. Finds limited evidence of strong links between individual war events and market returns, with some support for a negativity effect. Cited here as a caution against reading war chronologies into return series. Open-access preprint
17. Bondarenko, Yevheniia, Nayeon Kang, Vivien Lewis, Matthias Rottner, and Yves Schüler. “Geopolitical risk in the euro area: measurement and transmission.” BIS Working Paper 1348 (2026), also Deutsche Bundesbank Discussion Paper 14/2026. Finds euro-area geopolitical risk shocks are both recessionary and inflationary, though the estimated magnitudes per shock are small. The index follows the method of Caldara and Iacoviello (2022) but uses European sources, because the US-based index misses these effects. bis.org
18. Computed figures and method. All return tables labelled as computed here use two datasets already stored in this project. Annual nominal total returns and US CPI for 1928 to 2025 come from Aswath Damodaran’s historical returns dataset; real returns are derived as (1 + nominal) ÷ (1 + inflation) − 1, and the derivation was checked against the stored real series to machine precision. Because those data are annual, every war window is a calendar-year approximation of a conflict that began and ended mid-year, which is one of the measurement problems this guide describes. The volatility figures and the exact-date episode returns use daily value-weighted US market total returns from the Kenneth R. French Data Library, 1 July 1926 to 29 May 2026 (26,253 trading days); volatility is the annualized standard deviation of daily returns, and those figures are nominal, with no inflation adjustment. The two datasets are not interchangeable and their figures should not be compared directly.

Editor’s note

Educational content, not investment advice, and not a recommendation on any security, market or course of action. Index returns measure the value of claims on surviving listed companies and are not a measure of human welfare; nothing here should be read as a judgement about any conflict. Several widely circulated claims were checked and left out because they did not survive: that the New York Stock Exchange closed during the Civil War, that the fall of Atlanta moved the greenback market, and a frequently cited estimate of the return drag from international political crises whose source could not be opened for verification. Where a superseded figure is still in wide circulation, such as an earlier estimate of the wartime volatility reduction, the current one is used and the discrepancy noted in the sources. Figures attributed to Dimson, Marsh and Staunton are limited to the publicly available summary edition. Sources verified on August 2, 2026; market data cutoffs are stated in the method note above.