Cash and withdrawals during a market crash: what a reserve changes

Money needed for spending has a deadline that a market recovery does not. A cash reserve can fund withdrawals while shares and bonds are down, but holding it also changes exposure before the crash and forgoes returns in a rising market. A useful comparison must count both effects.

For identical withdrawals, the comparison measures remaining wealth across complete paths. It does not treat cash stability or lower volatility as goals in themselves, and the scenarios have no probabilities from which to estimate expected log growth.

The asset-class crash study provides historical context. The model below uses invented return paths, not a forecast, a retirement success-rate estimate or a reconstruction of 2008 or 2022. Its purpose is to show exactly where the money for withdrawals comes from.

Key takeaways

• A recovered market price does not restore investment units already sold to fund withdrawals.
• Cash changes both the initial amount exposed to a crash and the timing of later sales.
• The hypothetical crash and growth paths show different ending balances; neither establishes an optimal reserve.

Historical cash was stable relative to risky assets, but not constant

In the parent's fixed September 2007–September 2010 window, the Solactive Euro Overnight Return cash index gained 6.2% and showed no month-end drawdown. In January 2020–January 2023 it returned −0.7%, with a maximum drawdown of 1.2%.

Across its longer December 2005–July 2026 history, that series lost 2.8% from September 2014 to July 2022 and regained the old nominal peak in October 2023. The wait was 109 months, including 94 to the trough and 15 to recover. Low volatility did not guarantee either an unchanged nominal balance or preserved purchasing power.

These are overnight-rate index observations, not returns on banknotes, an insured savings account or a particular money-market fund. Those holdings have different interest, credit, access and fee arrangements. The Solactive index history via Curvo supplies these later cash observations. The dot-com window in the parent uses a separate ECB EONIA compounding calculation; no stitched continuous series is assumed.

Define the hypothetical shock and the spending rule

Start with €100,000 total, divided between cash and a risky investment sleeve. Compare initial cash amounts of €0, €12,000 and €24,000. The sleeve is initially 60% equities and 40% government bonds.

At time zero, assume equities fall 30% and bonds 15% simultaneously. The sleeve therefore falls 24%: 0.60 × 30% + 0.40 × 15%. These invented, synchronous losses are essential. Combining the historical worst losses of different assets on different dates would not produce an observed portfolio return.

After that shock, model the sleeve as a single unit price earning the same monthly return for 24 months, chosen to regain its pre-shock price exactly: (1 / 0.76)^(1/24) − 1, approximately 1.15% per month. There is no further modelling of the component assets or rebalancing.

Withdraw €1,000 at each month-end, after that month's return, spending cash first and then selling sleeve units. Cash earns zero. There are no taxes, fees, inflation, external income or reserve refills. Total withdrawals are €24,000.

Price recovery does not restore units already sold

The summary CSV and 144-row monthly ledger reproduce these rounded results.

With no cash, withdrawals require sales from month one. Only 72.5% of the original sleeve units remain at the end, so the full unit-price recovery leaves €72,541. That is €3,459 below the original €100,000 less €24,000 spent. Early sales surrendered participation in the rebound.

With €12,000 cash, selling begins in month 13. With €24,000 cash, all 24 payments come from cash, and the untouched sleeve recovers to its original €76,000 value.

This is not free insurance. Immediately after the shock, the three total balances are €76,000, €78,880 and €81,760 respectively. Cash already cushioned the initial loss by reducing the capital exposed. The ending differences reflect that initial exposure choice and the subsequent withdrawal sequence together.

A reserve also has an opportunity cost

The separate growth case in the table assumes no initial shock and a smooth 20% cumulative sleeve gain over 24 months, using 1.20^(1/24) − 1 each month. Spending and cash rules are identical.

Here, keeping €24,000 in zero-yield cash leaves €2,573 less at the end than starting fully invested: €91,200 versus €93,773. These scenarios were chosen to reveal the trade-off. They have no assigned probabilities and cannot establish that any reserve maximises expected geometric growth or makes withdrawals safe indefinitely.

For a real spending liability, the relevant inputs include payment dates, reliable external income, cash yield and the consequences of a recovery taking longer than expected. Cash held for a scheduled bill also has a different purpose from cash awaiting an investment opportunity, as discussed in the Klarman article.