Claim check: Dollar-cost averaging vs lump sum
“Dollar-cost averaging beats investing a lump sum.”
What the data says
Does not hold Not supported by the data: after twelve months, dollar-cost averaging ended ahead of a lump sum on a median 50% of start dates across 10 markets, and lump sum was ahead on more than 60% of start dates in 3 of them; the 5th-percentile outcome was -19% for DCA and -30% for the lump sum (medians over markets). It does not win more often than a coin flip, though it softens the worst outcomes.
No effect we can tell from chance, or the opposite effect.
Who says it
Vanguard's research says lump sum wins 2/3 of time; DCA is nonetheless an everyday recommendation.
Origins are as commonly cited and were not independently verified.
How we tested it
For every possible start date: a lump sum buys everything on day 0; dollar-cost averaging buys one twelfth on day 0 and then one twelfth every month for 11 more months, with the rest in cash at 0%. Both are valued twelve months after the start. We report the share of start dates on which DCA ends ahead, and the 5th-percentile outcome of each. Start dates overlap heavily, so they are not independent trials.
Results
| Market | Start dates | DCA ahead | First half | Second half | Lump sum, 5th pct | DCA, 5th pct | Lump sum, median | DCA, median |
|---|---|---|---|---|---|---|---|---|
| S&P 500 | 3988 | 16% | 15% | 18% | -6% | -6% | +16% | +9% |
| Gold | 3988 | 37% | 50% | 24% | -16% | -10% | +6% | +3% |
| Silver | 3988 | 48% | 62% | 35% | -28% | -18% | +2% | 0% |
| Oil | 3988 | 52% | 60% | 45% | -63% | -39% | 0% | 0% |
| Natural gas | 3988 | 71% | 77% | 65% | -64% | -43% | -24% | -12% |
| Commodities | 3988 | 52% | 59% | 45% | -31% | -19% | 0% | +1% |
| Agriculture | 3988 | 53% | 68% | 38% | -19% | -11% | -2% | -1% |
| EUR/USD | 4124 | 54% | 52% | 57% | -15% | -9% | -1% | 0% |
| Bitcoin | 4078 | 32% | 30% | 33% | -59% | -41% | +69% | +39% |
| Ether | 2929 | 43% | 41% | 45% | -73% | -46% | +11% | +11% |
Share of start dates on which dollar-cost averaging ends ahead
Stability across the two halves of history
A saying that only works in one half of the history is not much of a rule. Each row shows the effect in the full history, the first half and the second half; green means the claimed effect was present.
| Market | DCA ahead, full | First half | Second half |
|---|---|---|---|
| S&P 500 | 16% | 15% | 18% |
| Gold | 37% | 50% | 24% |
| Silver | 48% | 62% | 35% |
| Oil | 52% | 60% | 45% |
| Natural gas | 71% | 77% | 65% |
| Commodities | 52% | 59% | 45% |
| Agriculture | 53% | 68% | 38% |
| EUR/USD | 54% | 52% | 57% |
| Bitcoin | 32% | 30% | 33% |
| Ether | 43% | 41% | 45% |
The honest caveat
Cash earns zero here, which flatters lump sum a little, and we ignore investor behaviour.
Descriptive statistics from historical data, before costs. Not a forecast and not investment advice.
Questions
Dollar-cost averaging beats investing a lump sum.
Not supported by the data: after twelve months, dollar-cost averaging ended ahead of a lump sum on a median 50% of start dates across 10 markets, and lump sum was ahead on more than 60% of start dates in 3 of them; the 5th-percentile outcome was -19% for DCA and -30% for the lump sum (medians over markets). It does not win more often than a coin flip, though it softens the worst outcomes.
How was this claim tested?
For every possible start date: a lump sum buys everything on day 0; dollar-cost averaging buys one twelfth on day 0 and then one twelfth every month for 11 more months, with the rest in cash at 0%. Both are valued twelve months after the start. We report the share of start dates on which DCA ends ahead, and the 5th-percentile outcome of each. Start dates overlap heavily, so they are not independent trials.
Markets used
S&P 500 · Gold · Silver · Oil · Natural gas · Commodities · Agriculture · EUR/USD · Bitcoin · Ether