Lump sum or spread it out at a record high? 95 years of data

From December 2000 to October 2025, a sum invested in MSCI World at once in euros beat the same sum spread over 12 monthly instalments in 69% of starting months, and starting at a record high barely changed that (65%). The question is current: MSCI World's euro price index closed October 2026 at a month-end record, its eighth of the year, and the AMF counted 394,000 new retail ETF investors in the first half of 2026, the most since 2018.

Investing at once usually compounded more, but in the other third of cases it lost more, sometimes far more. The useful question is less "now or later" than how large a fall the whole sum can sit through.

Key takeaways

1. Investing at once won more often in this sample. It beat a twelve-month spread in 206 of 299 euro starting months, or 68.9%, with an average gap of 3.1% of the original sum before trading costs.
2. Record highs did not supply a clear timing rule. The lump sum won in 28 of 43 euro record starts, or 65.1%; this small, overlapping sample does not establish the odds for the next year.
3. An entry schedule cannot fix the final allocation. In the 1929 replay, spreading softened the first year but was still down 56.4% after three years; decide how much risk the whole sum can carry once every instalment is invested.

Is it better to invest a lump sum or spread it at a record high?

In both datasets we examined, investing at once came out ahead about two times in three, whether or not the market was at a record. We compared two ways of investing the same sum. The lump sum goes in at one month-end. The spread invests a twelfth at that month-end and at each of the next eleven, keeping the rest in cash at the euro overnight rate (or the US Treasury bill rate). Both are compared twelve months after the start, ignoring costs and taxes.

We ran this for every starting month in two series: MSCI World with net dividends reinvested in euros, from December 2000, and the S&P 500 with dividends from Robert Shiller's data, from December 1927. A start counts as a record when the price index is at or above every earlier month-end.

MSCI World in euros: 68.9% of 299 starts, 65.1% of 43 record starts. S&P 500 in dollars: 67.8% of 1,135 starts, 69.3% of 251 record starts. Source: Copio calculation on MSCI World index levels (MSCI), EONIA and the euro short-term rate (ECB), Robert Shiller's S&P Composite data and Damodaran's T-bill returns; summary (CSV), euro windows (CSV), US windows (CSV).

The euro record sample is small and recent. MSCI World's euro price index peaked in August 2000 and did not close a month above that level again until February 2015, so every euro record start falls between 2015 and 2025. The US series adds records before 1929, 1973, 1987, 2000 and 2007, and the result holds there too.

The most recent completed case is October 2025, also a record month. Invested at once, the sum was up 15.9% a year later; spread over 12 months, it was up 11.1%.

How much does spreading cost on average?

On average, spreading over 12 months left a gap equal to 3.1% of the original sum after one year. That denominator is the money available at the start, not the lump-sum investor's ending balance. Dividing each window's gap by its lump-sum ending balance gives an average relative shortfall of 1.6% across the 299 starts, our calculation.

Once both portfolios are fully invested, with no further external flows or differing costs, each window's ratio of balances stays the same as the index grows. It is that relative ratio which persists, not 3.1% of starting capital reinterpreted as 3.1% of future wealth.

Average +3.1%, median +4.5%, 5th percentile −13.8%, 95th percentile +16.0% of the sum; worst −25.9% (August 2008 start), best +27.1% (March 2009 start). Source: Copio calculation on MSCI World Net Total Return EUR and ECB overnight rates; euro windows (CSV), monthly inputs (CSV).

Through the copio lens, the measure is compounded wealth after costs, and here the average favours investing at once. Markets rose in most years, and a spread keeps part of the sum out of them while they do. The spread also adds costs the index does not show: twelve purchases instead of one, unless the broker's savings plan is free.

The distribution is wide on both sides. The worst start for the lump sum, August 2008, was followed by the steepest part of the financial crisis; the best, March 2009, came one month after the index's month-end low. Neither was knowable at the start.

What happens if a crash follows straight after?

Spreading cut the loss in the first year, but when the fall lasted longer than the spread, the spread investor lost heavily too. The chart shows five starts chosen after the fact, at or near the peak before a fall, which is the case spreading is meant for.

After three years: September 1929 −67.7% at once against −56.4% spread; August 2000 −30.3% against −19.3%; October 2007 −19.2% against −1.2%; January 2020 +27.1% against +34.3%; December 2021 +32.1% against +41.3%. Source: Copio calculation; replays (CSV), US monthly inputs (CSV).

Source: Copio calculation; replays (CSV).

In 1929, the spread investor had made the last instalment by August 1930, and the market kept falling for almost two more years. Spreading changed the entry price on part of the sum; it did not change what the whole sum was exposed to once invested. Across all euro starts, the worst 12-month result was −39.4% for the lump sum and −25.8% for the spread: less, but still a large loss.

This is the crash-aware point. A sum that would have to be sold after a 40% fall is too exposed whether it goes in at once or over a year. Our study of recovery time after crashes shows why a forced sale during a fall turns a temporary loss into a permanent one, and our model of withdrawals during a crash shows what a separate cash reserve changes.

Is spreading a sum a form of diversification?

Spreading temporarily combines equities with cash, but does not add a lasting asset-class diversifier after the last instalment. Over a 12-month spread, an average of 54% of the sum is invested during the year and 46% waits in cash. Cash does not fall with equities, which is why the spread loses less early in a crash. A lasting diversification decision concerns the completed portfolio; a spread holds a cash buffer for a year, then holds the same index as the lump sum.

The cash leg used to earn little. From January 2001 to October 2026, the euro overnight rate averaged 1.37% a year, and it was negative from October 2014 to August 2022; on 2 November 2026 the euro short-term rate was 2.44%. A higher cash rate narrows the gap slightly, by a fraction of the 46% of the sum held in cash, but it does not change which strategy had the higher average.

If what you want is a lasting buffer, that is an allocation decision, not an entry schedule: a share of the portfolio held in assets that behaved differently in falls, kept after the first year. Our diversification study compares which asset classes did that and when.

Is a record high a reason to wait?

This sample does not provide a clear reason to wait merely because the index is at a record. The euro record-high subset contains only 43 overlapping starts, so similar observed win rates do not prove equal odds. Records are common in a rising market: from January 2015 to October 2026, 36% of month-ends of MSCI World's euro price index were records. A record can also be followed by a long wait. The August 2000 peak of the euro price index was not regained until February 2015.

None of this is a forecast for the next twelve months. It says only that "the market is at a record" was not, on its own, a reliable signal in the past.

When can spreading still make sense?

Spreading can still make sense when it keeps you invested. The 31% of euro starts in which it won include the cases most investors fear most, and abandoning a plan after an early fall can change the outcome much more than the average entry-price gap. Three situations are worth separating:

• Money needed within a few years. That is not a timing question; it belongs in cash or short-term savings, not in an equity spread.
• A sum that is large compared with what you already hold. The worst first-year loss falls on new money you have not yet seen move, and the regret is larger.
• A schedule you will keep. A spread fixed in advance, with set dates, behaves like the one we modelled; a spread that waits for a dip becomes market timing.

What does copio conclude?

Choose the lasting allocation before the entry schedule. Investing at once won in 68.9% of the euro windows, but spreading sometimes left substantially more money after a falling first year. Once fully invested, both approaches face the same market exposure. The historical win rate cannot choose a risk level or identify the next losing start for you.

Three checks before investing

1. Separate the money by purpose and date. List foreseeable spending and the liquid assets that will fund it, then identify the portion intended for long-term investment. Apply a 40% equity-loss scenario to the fully invested target allocation and calculate the loss in euros. If that would force a sale to meet spending, revisit the allocation before choosing instalment dates.
2. Compare the two entry plans after costs. For investing at once and for your chosen schedule, record purchase fees, the return on waiting cash and the first-year loss you could face. Use the same final allocation and horizon in both cases. The 3.1% average historical gap is a share of the original sum, not a guaranteed return or a permanent 3.1% relative wealth difference.
3. Write the schedule and the rule after it ends. If you choose instalments, record amounts and dates before the first purchase, along with how cash will be held until invested. Define what happens if markets rise or fall, and set the subsequent rebalancing rule for the completed portfolio. Repeatedly postponing the next instalment is a different strategy from the fixed schedules studied here.

Use the portfolio checker to explore the final fund mix and the cash-withdrawal study to examine its funding needs. Copio provides educational analysis and does not choose an allocation or place trades for you.