Lump Sum vs DCA: What Historical Distributions Show (Not a Recommendation)

The choice is simple to state, and still hard to make

The decision is simple to state. You have a large sum in cash, and you either put it into your target allocation in one go, or you feed it in over a set number of months (dollar-cost averaging, DCA). Which approach produced better outcomes in the past is well studied. Why the choice still feels so hard is less often discussed.

When I had a large amount to invest, the math was not what stopped me. What stopped me was not knowing where the market stood. Buy near a high and the early stretch can show a loss right away; buy near a low and that early pain is less likely. For a lot of beginners, that is the whole dilemma at the start: the first decision feels like it already needs a correct call on market position. I kept trying to answer that call first, as if the rest of the choice depended on it.

It doesn't, and it can't. Nobody answers "where is the market now?" reliably, including the people who sound most certain. So this piece takes a different route: it shows how to make the lump-sum-or-DCA decision without needing that answer. That means looking at the full range of historical outcomes rather than the average, choosing a DCA window on purpose, and deciding in advance which kind of regret you can live with.

What the published research measured

The most cited study is Vanguard's February 2023 paper Cost averaging: Invest now or temporarily hold your cash? by Megan Finlay and Josef Zorn. The setup matters more than the headline, so here it is.

They compared investing a sum all at once against splitting it into three equal parts invested one month apart. Each strategy was judged by wealth one year after the start, on a rolling basis, using MSCI World Index returns from 1976 to 2022. The base case was 100% equity, with the uninvested cash earning nothing.

Under those assumptions, lump sum ended ahead of the three-month split 68% of the time. The split ended ahead of simply staying in cash 69% of the time. When the authors let the waiting cash earn the 3-month U.S. Treasury bill rate, lump sum still came out ahead 65% of the time for the all-equity portfolio.

The mechanism is unglamorous. Money waiting in cash is not exposed to the stock market's risk premium during the waiting months. Over most historical windows, markets rose more often than they fell, so the waiting cost something more often than it saved something. That is the whole engine behind the result. It is not a claim that anyone can tell when the market is high.

The distribution, not the average

"68% of the time" also means roughly one time in three the split came out ahead. The paper's percentile table (Figure 3) shows what those cases looked like. Starting from $100,000 in a 100% equity portfolio, one-year ending wealth looked like this:

95th percentile (best years): lump sum $139,453; 3-month split $131,012.

50th percentile (median): lump sum $111,940; 3-month split $109,580.

5th percentile (worst years): lump sum $82,947; 3-month split $85,906.

Source: Vanguard, Finlay & Zorn (2023), Figure 3. MSCI World Index, 100% equity, $100,000 start, no interest on waiting cash. Historical, not a forecast.

Read across the bottom row. In the worst 5% of starting points, splitting the money left about $3,000 more than going all in. Across the middle and the top of the range, lump sum finished ahead, and further ahead at the top. Splitting narrows the range at both ends. It gives up some upside to soften the downside, and over a typical year that trade has cost money.

That is the honest version of the "buy at a high" fear. The fear points at something real: the bottom tail. It just overestimates how often you land there.

A histogram of those gaps across every historical start month would make the same point visually. When our /tools/lump-sum-vs-dca/ (shipping this biweek) calculator ships, it will show that distribution for a chosen amount and window. Until then, the percentile figures above are the public reference.

Why "is it a high right now?" is the wrong first question

A market at an all-time high feels like a special situation. Historically it is a common one. In a market that rises over long stretches, new highs keep arriving. That means "invest windfall all at once at all time high" is the same decision as investing it on any other day, just with louder feelings attached.

What matters for regret is not the label "high" but what happens next: how deep a drawdown follows, and how long the portfolio sits below its starting value.

Depth and duration both matter. A short 10% dip feels different from a 30% drop that takes years to recover. Using dividend-adjusted daily closes as of 2026-10-30 for SPY (1993-01-29 through 2026-10-30), only 10.5% of trading days were at a new all-time high, while 55.5% of days sat within 5% of the prior high. Declines of 10% or more from a prior high happened 12 times; the deepest was −55.2% from the 2007 peak to the 2009 trough, and getting back to that prior peak took hundreds of trading days. On QQQ (1999-03-10 through 2026-10-30) the picture is harsher still: 47.6% of days spent 30% or more below the prior high, almost entirely the long stretch from the 2000 peak until that level was regained in 2015.

The useful question stays personal: if the portfolio fell 20% or 30% in the first stretch, would the written rule still hold?

A separate check on the same SPY series compared investing all at once against splitting over 12 months with waiting cash earning 0%. Lump sum finished ahead of the 12-month split in 78.9% of start months when judged 12 months later, and in 76.5% when judged 10 years later. That is consistent with the Vanguard direction for shorter windows, and it is still not a forecast for any particular starting day.

A table like that turns a vague fear into something you can size. "If the portfolio fell 20% in the first year, would I still hold it?" is a question you can answer today, about yourself. "Is the market at a peak?" is not.

DCA over 6 months or 12 months

If you choose to split, the length of the window is a choice of its own, and it is usually made by gut feel.

The Vanguard paper tested windows from three to six monthly installments (Appendix 1). For the global index in U.S. dollars, the share of one-year periods where lump sum came out ahead was 66.4% with a three-month split and 73.7% with a six-month split. The longer the window, the more time cash sits out of the market, and the more often the split falls behind. The paper did not test a 12-month window. Our own SPY check above (lump sum vs a 12-month split, cash earning 0%) found lump sum ahead most of the time at both the one-year and ten-year marks; the mechanism is the same: more months in cash, more expected drag, a narrower range of outcomes.

So the window is really a dial. Turning it up buys a narrower outcome range and pays for it with expected return. There is no "correct" setting. The goal is a setting you picked on purpose and will not change halfway because of a headline.

When the calculator ships, it will also compare 3-, 6-, and 12-month windows side by side. The mechanism above already tells you what to expect: longer windows spend more months in cash and fall behind lump sum more often.

A way to decide without knowing where the market is

Here is the workflow I wish I had used. None of it requires a view on the market.

Write the rule before you look at prices. Decide lump sum or DCA, and if DCA, the number of months, the amount per installment, and the date each one goes in. Put it in writing. Automate the transfers if your broker allows it. A rule written on a calm day is the point; a rule rewritten after a bad week is just market timing in a different outfit.

Compare distributions, not anecdotes. Look at the median gap and the worst-5% gap between the two approaches, not at the one year a friend remembers. When our tool ships, /tools/lump-sum-vs-dca/ (shipping this biweek) will show those ranges for your amount and window. Until then, the Vanguard figures above are a reasonable public reference with their assumptions stated.

Pick the regret you can live with. There are two failure modes. With a lump sum, the market drops right after and the whole amount takes the hit. With DCA, the market rises steadily and part of the money goes in at higher prices. Both are uncomfortable. Ask which one would make you abandon the plan. The Vanguard authors make a similar point from the other side: a split can make sense for someone with very high loss aversion who would otherwise leave the money in cash indefinitely, and even then they favor a short window. Staying in cash forever is the outcome both strategies beat.

Decide what the waiting cash earns. If the money waits, park it somewhere with a stated yield and note that yield as an assumption. It changes the comparison, as the T-bill version of the study shows.

Then check the result honestly. Once the money is invested, /tools/investment-return/ can show what the contributions actually earned, so the review rests on numbers rather than memory.

You are not the only one stuck on this

The question comes up constantly. In the 90 days to 2026-10-30, our scan of eight personal-finance subreddits matched 443 posts on this theme (regex count). Two representative threads: one reader describing how it's mentally getting harder to invest without timing the market, and another weighing whether to DCA or lump sum €350K. The replies mostly repeat the headline result. What readers seem to want is the part above: the range, the worst case, and a rule they can keep.

What this does not cover

Taxes, account types, and whether part of a windfall belongs in an emergency fund or debt repayment first are separate decisions. They often come before this one. The research above uses broad index data. A single stock, a sector fund, or a non-U.S.-dollar investor's currency exposure can behave very differently. The historical record is one path through markets, and the next decade is not obliged to resemble it.