Articles · Great investors
Keynes as an investor: from market timer to value stock picker
How John Maynard Keynes ran King’s College’s endowment from 1921 to 1946: his early macro timing, the losses of 1929–30, his switch to value stock picking and the record.
On this page
John Maynard Keynes (1883–1946) is remembered as the economist whose ideas reshaped government policy after the Great Depression. He was also an active investor for most of his adult life, and for 25 years he ran the endowment of King's College, Cambridge. Thanks to archival work by the financial historians David Chambers and Elroy Dimson, his investment record is now one of the best documented of any investor of his era.
That record has an unusual shape. For his first decade Keynes tried to forecast the economy and move money between stocks, bonds and cash, and he did not do it well: he missed the crash of 1929. Around 1932 he changed his method completely and became a patient, concentrated stock picker who bought shares he judged cheap relative to their value. From then on his results were far better. This article covers his investing only; his economic theory is in Keynesian economics: demand, the multiplier and the evidence.
The bursar and the Discretionary Portfolio
Keynes became a Fellow of King's in 1909. Appointed Second Bursar just after the First World War, he had primary responsibility for the college's investments from 1921 and became First Bursar, with full charge of its finances, in 1924. He held the post until his death in April 1946, while also writing his major books, advising the British Treasury and negotiating at Bretton Woods.
Oxford and Cambridge colleges were then bound by trustee laws and their own statutes, which limited most investments to high-grade fixed income such as government and railway bonds, plus land. In 1921 Keynes persuaded the Fellows to set part of the endowment outside those rules. Chambers and Dimson call it the Discretionary Portfolio. It began as a single account known as the Chest in the financial year ended August 1921; a second account, Fund B, was added in 1933 to pool many small endowed funds. The discretionary money grew from 8% of the college's securities in 1921 to 68% by 1946. The rest, the Restricted Portfolio, stayed largely in bonds and real estate.
The record, year by year
Keynes reported to the college each year on every account. From those reports, Chambers and Dimson rebuilt the annual total returns of the Discretionary Portfolio and compared them with an equally weighted index of UK stocks. Over the 25 financial years from 1921–22 to 1945–46:
| Measure (annual, before costs) | Discretionary Portfolio | UK equity index |
|---|---|---|
| Average (arithmetic) return | 15.97% | 10.37% |
| Standard deviation | 19.08% | 17.11% |
| Sharpe ratio | 0.73 | 0.49 |
| Years behind the index | 6 of 25 |
The Restricted Portfolio, held mostly in bonds, averaged 6.81%. Keynes's tracking error against the index, the spread of his yearly return around the market's, was 13.9%, several times what modern endowments run. His portfolio looked nothing like the market.
Compounding the table's annual returns shows the shape of the record. £100 placed in 1921 would have grown to about £2,893 by August 1946 against £855 in the UK index, roughly 14.4% a year against 9.0%. The college itself did not accumulate this way, since it spent its investment income each year, but the series shows how the two strategies compare.
The figures are estimates from Keynes's own year-end valuations, and they are gross of trading costs. Chambers and Dimson also note that Keynes, a well-connected public figure, may at times have had access to information other investors lacked. Even with those caveats the margin is large, and the authors' statistical tests attribute it to stock selection, not to luck in a few years.
The 1920s: forecasting the cycle
Keynes's first approach followed naturally from his profession. He believed that an economist who understood the credit cycle could anticipate when stocks, bonds and cash would do best, and shift between them. In the 1920s the Discretionary Portfolio's weight in UK ordinary shares averaged about 75%, and he moved in and out of sectors and asset classes according to his reading of the economy. He also speculated personally in currencies and commodities. In 1920, according to his biographers, a currency speculation he ran with friends was nearly wiped out before he recovered the losses.
The results were mediocre. The portfolio trailed the UK index in 1926, 1927 and 1928, three strong years for British stocks. Keynes did not foresee the sharp fall in share prices that began in late 1929. The portfolio fell 14.2% in the year to August 1930, by which time it was, by the authors' count, 12.6% behind the index since inception and 40.3% behind over the trailing five years. When the UK market began to fall in October 1929, he sold about a fifth of his UK equities over the following twelve months and moved into government bonds, a classic market-timing response.
Chambers and Dimson tested his asset-allocation moves statistically and found no evidence of timing skill in any period. The great economist could not reliably forecast the market. The same lesson appears in the 1929 history (The crash of 1929: causes, the 1929–32 decline and lessons) and in nearly every later study of market timing.
The change: a value investor
Between 1930 and 1932 Keynes rethought his approach. The archive shows him corresponding with a close investing friend, Francis Scott, chairman of the Provincial Insurance Company, about individual stocks dozens of times in 1933–34, after rarely doing so before. In an August 1934 letter to Scott he wrote that he was increasingly convinced the right method was to put fairly large sums into a few businesses he understood and whose management he trusted. He later attributed his success to concentrating on a few core holdings that he considered cheap relative to their intrinsic value and holding them for several years.
The new approach had several features:
- Bottom-up, not top-down. He judged companies one by one rather than forecasting the economy.
- Value. He favored shares with high dividend yields and low prices relative to assets and earnings. This was close in spirit to Benjamin Graham's method, developed independently at the same time in New York (see Benjamin Graham and value investing: margin of safety explained).
- Concentration. His top holdings made up a large share of the portfolio. Mining companies (tin in the 1920s, gold in the 1930s) and industrial firms, including the carmakers Austin and Leyland, featured heavily, while banks, about a fifth of the index in 1927, and oil stocks were absent from his portfolio in the years the authors sampled.
- Smaller companies. He tilted toward mid-sized and small firms, a pattern later research associates with higher long-run returns (see Small-cap investing and the size effect: a 26-year test).
- Patience in falls. When the UK market dropped sharply again from August 1937, he did the opposite of 1929: he added modestly to his equities and kept the stock weighting above 90%.
The patience did not prevent bad years. The portfolio lost 22.6% in 1937–38, more than the market's 8.7% fall, and Keynes defended himself to the college's committee in a memorandum arguing that a long-term holder should not sell good companies merely because their prices had dropped. After 1930, the 1937–38 year was the only one of substantial underperformance, and the portfolio was never behind the UK index over any rolling three- or five-year period.
Why the record matters
Equities for institutions. Keynes put most of King's discretionary money into ordinary shares when trustee investors in Britain and America regarded them as too risky. In the 1920s the largest US university endowments held less than 10% in common stocks; no other Oxford or Cambridge college made a substantial equity allocation until well after the Second World War. Over the following century the equity risk premium rewarded that choice. The idea that a perpetual institution can bear short-term volatility to earn long-run returns runs from Keynes to the modern endowment model of David Swensen (see David Swensen and the Yale model of endowment investing).
Learning from failure. Few investors with Keynes's reputation have so clearly abandoned a method that was not working. His shift from forecasting the economy to valuing companies is a documented case of an expert discovering the limits of his own expertise.
Concentration cuts both ways. A 13.9% tracking error means long spells of looking very different from the market. Keynes had the authority to ride out his bad years; most professional managers would have been replaced in the late 1920s, before the good years arrived.
How Finabase helps
Finabase does not reproduce King's College's portfolio, but it supports the habits Keynes adopted. The Screener can filter for high dividend yields and low price-to-book ratios, the measures his later portfolio favored. A symbol's Dividends tab, for example Coca-Cola's, shows the payout history that a yield-minded investor checks. A portfolio tracks a concentrated set of holdings, so you can see how far it strays from the market over time.
Frequently asked questions
Was Keynes a good investor?
By the measure of Chambers and Dimson's reconstruction, yes: the King's College Discretionary Portfolio averaged about 16% a year before costs from 1921 to 1946, against about 10% for UK stocks. But nearly all of the outperformance came after he changed his method around 1932. Before that he trailed the market.
What was the King's College Chest Fund?
The Chest was the account Keynes ran with full discretion from 1921, outside the rules that limited college investments to bonds and property. From 1933 it was joined by a second account, Fund B, and the two together form what researchers call the Discretionary Portfolio.
Did Keynes lose money in 1929?
Yes. He failed to foresee the late-1929 crash, and the Discretionary Portfolio lost 14.2% in the financial year to August 1930 and another 11.5% in the year to August 1931, though it fell less than the UK index in both years.
What was Keynes's investment strategy?
In the 1920s he moved money between stocks, bonds and cash based on economic forecasts. From the early 1930s he picked a small number of companies he considered cheap relative to their value, often with high dividend yields, and held them for years, including through market falls.
Summary
Keynes ran King's College's discretionary money from 1921 to 1946 and earned, by Chambers and Dimson's estimate, an average return about five and a half percentage points a year above UK stocks. The path matters as much as the result: a decade of economic forecasting that missed the 1929 crash and trailed the market, then a deliberate switch to concentrated, patient value investing that produced most of the gains. He was also among the first institutional investors to commit heavily to stocks, a choice that later shaped how endowments invest.
Check your understanding
4 questions on this lesson. Each answer comes with the reason, and nothing is scored.
What was unusual about the King’s College Discretionary Portfolio that Keynes ran from 1921?
- It held only government bonds
- It was freed from the trustee rules that kept college money in bonds and property, so Keynes could put most of it in stocks
- It was managed by an outside bank
Show the answer
B. It was freed from the trustee rules that kept college money in bonds and property, so Keynes could put most of it in stocks Keynes persuaded the Fellows to exempt part of the endowment. Few institutions anywhere held much in common stocks at the time.
How did Keynes’s results compare before and after his change of method around 1932?
- They were strong before and weak after
- He trailed the market in the late 1920s and missed the 1929 crash; afterward he never trailed the UK index over any rolling three- or five-year period
- They were the same in both periods
Show the answer
B. He trailed the market in the late 1920s and missed the 1929 crash; afterward he never trailed the UK index over any rolling three- or five-year period Nearly all of his long-run outperformance came after he moved from economic forecasting to picking a few undervalued companies.
Chambers and Dimson tested Keynes’s moves between stocks, bonds and cash. What did they find?
- Strong market-timing skill
- No evidence of market-timing skill in any period
- Timing skill only in the 1940s
Show the answer
B. No evidence of market-timing skill in any period Even an economist of his stature could not reliably forecast the market. His edge came from stock selection.
Why should the 15.97% average annual return be read with some caution?
- It is after all costs and taxes
- It is estimated from Keynes’s own valuations, is gross of trading costs, and he may sometimes have had privileged information
- It covers only one year
Show the answer
B. It is estimated from Keynes’s own valuations, is gross of trading costs, and he may sometimes have had privileged information The researchers state these caveats themselves. The margin over the market is large enough that the conclusion survives them, but the exact figure is an estimate.
Sources and further reading
- Chambers and Dimson — John Maynard Keynes, investment innovator, Journal of Economic Perspectives 27(3), 2013, pp. 213–228
- Chambers and Dimson — Keynes the stock market investor, working paper, Cambridge Judge Business School and London Business School, September 6, 2012 (Table 2: King’s College performance 1921–46)
- Chambers, Dimson and Foo — Keynes the stock market investor: a quantitative analysis, Journal of Financial and Quantitative Analysis 50(4), 2015
- Moggridge (ed.) — The Collected Writings of John Maynard Keynes, vol. XII: Economic Articles and Correspondence: Investment and Editorial, 1983
- Skidelsky — John Maynard Keynes, vols. 1–3, 1983–2000
