Dollar-cost averaging vs lump sum: the evidence

You have a sum of money that is not yet invested — an inheritance, a bonus, the proceeds of a sale — and a plan for where it should go. The question is whether to put it in today or in equal pieces over the coming months. The second is called dollar-cost averaging, and it feels safer. The evidence on which one has actually ended with more money is unusually clear, and it is not the one that feels safer.

What Vanguard found, twice

In 2012 Vanguard's research group compared the two methods on U.S. data from 1926 to 2011, for a 60% stock, 40% bond portfolio, investing a lump sum at once versus spreading it over twelve months, measured over every rolling ten-year period. The lump sum came out ahead 67% of the time, with an average ending value 2.3% higher. The United Kingdom gave 67%; Australia gave 66%. The paper's title is its finding: "Dollar-cost averaging just means taking risk later."

In 2023 a second Vanguard team re-ran the question on global stock data from 1976 to 2022, with a three-month averaging period and a one-year horizon. Investing at once won 68% of the time for an all-stock portfolio; by market, 66.4% in the U.S., 68.1% in the U.K., 67.2% in Canada, 67.5% in Australia and 61.6% in emerging markets. The median advantage after one year was 2.2% of wealth for an all-stock portfolio, 1.8% for 60/40 and 1.2% for 40/60.

Northwestern Mutual's own study, updated in 2025, measured rolling ten-year periods and found the lump sum ahead 75% of the time for an all-stock portfolio, 80% for 60/40 and 90% for an all-bond one. CNBC's report of the 2021 version of that study gives its data as starting in 1950.

Why the lump sum wins

Markets have risen more often than they have fallen, over most stretches of the last century. Money sitting in cash while it waits to be averaged in is money not earning that. Averaging over twelve months means that, on average, half the sum spends six months out of the market, and the expected cost of those six months is positive. That is the entire mechanism. Averaging does not reduce the risk of the investment; it postpones it — which is what the 2012 title means. At the end of the averaging period you hold exactly the portfolio you would have held on day one, having paid, on average, a little more for it.

Why averaging still has a case

Two-thirds is not all. The one-third of periods in which averaging won are the ones where the market fell soon after the lump sum went in, and those are precisely the outcomes people fear most — not because of the arithmetic but because of what it feels like to watch a lifetime's savings drop in the first month. The 2023 Vanguard paper is explicit that cost averaging "might be considered for investors with very high aversion to both risk and losses," because lowering the early risk "could limit drawdown and the accompanying investor regret." If the alternative to averaging is not investing at all, or investing and then selling at the first decline, averaging is the better outcome by a wide margin. The evidence says the lump sum has the higher expected value; it does not say that expected value is the only thing a person is allowed to care about.

One case where the question does not arise: money that arrives in paychecks is averaged whether you like it or not. Dollar-cost averaging is only a choice when you have a sum in hand.

The cousin of this question: the order returns arrive in

Once money is being withdrawn rather than added, the order of returns starts to matter in a way it never did while saving. Two retirements that earn the same average return over thirty years can end very differently if one of them gets its bad years first, because the early withdrawals are taken from a diminished pile and the recovery happens on less money. This is sequence-of-returns risk, and it is the reason retirement planning is not simply "what is the average return."

The classic work is William Bengen's 1994 study of U.S. data from 1926 onward: withdrawing 4% of the starting portfolio in the first year, and that dollar amount adjusted for inflation thereafter, had lasted at least thirty years in every historical period he tested, with a portfolio of 50% to 75% stocks. "I counsel my clients to withdraw at no more than a four-percent rate." The Trinity study by Cooley, Hubbard and Walz, published in 1998 and updated in 1999 with data through 1997, found the same 4% rule over thirty years succeeded in 95% of historical periods for a 50/50 portfolio and 100% for 75/25, and that stock-heavy portfolios made 3% and 4% withdrawal rates "exceedingly conservative" in most periods.

These are historical U.S. results, not laws, and the authors say so; a future that is worse than the past is the case they cannot cover. The point for this page is narrower: the moment money starts coming out, the timing of returns becomes a risk in its own right, and a plan that only looks at averages has not looked at it.

What this has to do with a trading desk

Only this: the question people ask the desk most — "is now a good time?" — is the same question in another costume, and the evidence here is the best available answer to it. Waiting for a better moment has, on average, cost money. What has not cost money is deciding first how much risk you can carry, and then taking exactly that much, whatever the calendar says. Position sizing from the stop is the desk's version of the same rule.

• The three-fund portfolio and why it is hard to beat
• Rebalancing — how often, and what it is really for
• Taxes: what the account decides before the market does

← Long-term portfolios · All of Learn