The Emergency Fund Sweet Spot
"Keep 3–6 months of expenses" is a blunt rule that ignores your second income and how fast your industry rehires. Size the fund to the shortfall you'd actually face and the 95th-percentile time to get back to work. Then see what over-stuffing cash costs you in forgone market returns.
All figures are in today's money: every rate on this page is real, i.e. above inflation. Why?
The Household
Risk & Returns
| Metric | P10 | P50 | P90 |
|---|---|---|---|
| Optimal Fund Size 12 mo × €410 shortfall | – | €4,920 | – |
| Worst-Case Fund (both jobless) your fund already covers it | – | €12,240 | – |
| Annual Opportunity Cost on the excess cash, /yr | – | €504 | – |
| Over-saved Wealth Cost (25 yr) forgone growth on excess | €2,911 | €24,054 | €79,090 |
Coverage by Fund Size
- Coverage %: Share of the emergency-need window this fund size covers.
- Sweet spot: The fund-size range that balances coverage against opportunity cost.
- Both-jobless fund: Fund size needed to cover both incomes being lost at once.
Cash Drag Over Time
- Cash drag: The gap between the two lines.
The gap is what your whole buffer forgoes by sitting in cash. The portion up to the optimal fund is deliberate insurance, not waste; only the excess above it is truly idle.
Opportunity Cost of the Idle Surplus
Forgone market growth on the €10,080 you hold above optimal, if it were invested instead of held as cash. The solid line is the deterministic median path; the band is the P10–P90 spread of the modelled return distribution.
- Forgone growth P10–P90 · 80% of scenarios: The middle 80% of simulated futures: 1 run in 10 ends above this ribbon and 1 in 10 below it.
- Forgone growth P25–P75 · middle half: The middle half of simulated futures: a quarter end above it, a quarter below.
- solid = median (P50) path: The solid line is the median (P50) path. Because returns compound, the arithmetic mean sits ABOVE the median: a few very good runs pull the average up.
The fund is above the modelled need
It holds 305% of what the 95th-percentile scenario needs. In the model the surplus is safe but idle: at the assumed market return, the same money invested would be expected to grow.
305%
of optimal held
Our baseline is the realistic case: one earner loses their job while the other keeps working. Of the two possible job losses we size for the one that needs the bigger fund, here earner 1's. The monthly shortfall is expenses, minus the costs that vanish when you stop working (commute, lunch), minus unemployment benefit, minus the surviving income: €4,500 − €150 − €2,040 − €1,900 = €410/mo while benefit is paid. Benefit, vanishing costs and a second income all work together so you burn cash slower than the "months of expenses" rule assumes.
The optimal fund covers that shortfall for the 95th-percentile time your industry takes to rehire (12 months): €410 × 12 = €4,920. The coverage curve is an S-shaped logistic on re-employment time that starts at 0% for an empty fund (nobody is rehired before losing the job) and is calibrated so the optimal fund lands at 95% coverage. Past it, each extra euro buys almost no safety, while it still sits idle earning nothing. See the Cash Drag chart below for what holding your whole buffer as cash costs against investing it instead, though only the slice above optimal is genuinely wasted; the rest is the insurance you're paying for.
The true worst case is both earners jobless at once: no income survives, but both earners' vanishing work costs and unemployment benefits still cut the burn to €1,020/mo while benefit is paid , so the fund needed becomes €12,240 (the red "both jobless" marker). It is the belt-and-braces number, but sizing your cash for it instead of the realistic case costs an extra €366 per year in forgone growth (€7,320 of extra cash × 5%). Which case you fund is a risk-appetite call; the verdicts above anchor on the realistic one-income case.
Cash beyond optimal earns nothing while it could compound at 5%. Holding €10,080 too much costs €504 in the first year and €24,054 of forgone growth over 25 years.
How the scenarios work: the €10,080 you hold above optimal compounds at a lognormal return with median 5%/yr and volatility σ = 15% (set in the header); the shaded band is the resulting P10–P90 range of forgone growth, and the solid line is the median (P50) path. Only the idle excess is modelled this way. The buffer up to optimal is protection, not an investment, so it never enters the bet.
Realistic monthly shortfall
S = max(0, 4,500 − 150 − I_kept − B_lost)
- S: realistic monthly shortfall: what the household can't otherwise cover
- E: monthly household expenses
- J: costs that vanish when jobless (commute, work lunches etc.)
- I_kept: net income of the earner who keeps working, per month
- B_lost: unemployment benefit of the earner who loses work (their income × own benefit %, clamped 0-100). Of the two possible job losses, the one needing the bigger fund is used
Optimal fund size
Fopt = S × min(12, 12) + S′ × max(0, 12 − 12) S′ = max(0, E − J − I_kept)
- Fopt: optimal fund size
- S: realistic monthly shortfall while benefit is paid (see above)
- S′: monthly shortfall once benefit has run out
- m: months to 95th-percentile re-employment (Industry Stability)
- b: months unemployment benefit is paid
Worst-case fund (both earners jobless)
Fworst = max(0, 4,500 − 2J − B1 − B2) × min(12, 12) + max(0, 4,500 − 2J) × max(0, 12 − 12)
- Fworst: worst-case fund size, with both earners jobless at once
- E: monthly household expenses
- J: costs that vanish per earner when jobless
- B1: earner 1 unemployment benefit (earner 1's net income × own benefit %, clamped 0-100)
- B2: earner 2 unemployment benefit (earner 2's net income × own benefit %, clamped 0-100)
- m: months to 95th-percentile re-employment
- b: months unemployment benefit is paid
Idle excess above optimal
X = max(0, 15,000 − Fopt)
- X: excess cash held above the optimal fund, the only slice modelled as an investment
- F: current emergency fund held, in cash
- Fopt: optimal fund size (see above)
Forgone growth on the idle excess
Wealth(25) = X × ((1 + r/100)^25 − 1)
- Wealth: forgone growth over N years, had X stayed in cash instead of invested
- X: idle excess (see above)
- r: market return, %/yr (real)
- N: wealth-cost horizon, in years
Simplifications: the realistic case assumes whichever single job loss needs the bigger fund, while the worst case assumes both lose work simultaneously and stay jobless for the full rehire window. Every figure here is real (inflation-adjusted): expenses and incomes are held in today's money, and the global Market Return the excess cash forgoes is itself a real return, so nothing here needs a separate inflation input, and the comparison stays like-for-like. Each earner's unemployment benefit is a percentage of their own prior income; in the worst case both benefits are drawn at once. The fund itself is assumed to earn 0% real; opportunity cost is charged only on the excess above optimal. Re-employment risk is a single logistic curve, not a full labour-market simulation.
Assumptions behind every figure: how this site models the market →
Every model leaves things out. Here is what this one does not see:
- The "both earners jobless" worst case is offered as an alternative scenario to pick, not weighted by its actual likelihood: no probability is attached to either case. Anchoring on the realistic single-job-loss case, as the verdicts above do, understates the true risk for a household whose two incomes come from correlated industries (same employer, same sector, same region), where a recession is more likely to take both at once.
- Only unemployment is priced as an emergency. Medical bills, urgent home or car repairs and family support all draw on the same cash buffer, but none of them enters this calculator's sizing, so a fund labelled "optimal" here can still be inadequate for a non-job-loss shock.
- No correlation between market conditions and job-loss timing. Recessions tend to destroy jobs and depress markets together, but the modelled P10-P90 fan treats the market path as independent of the job-loss event the fund exists to survive. The forgone-growth opportunity cost of holding cash could look worse on paper than it would be in the exact recession that actually triggers the need for the fund.
- The fund is assumed instantly accessible at its full nominal value with zero real return and no access friction: no early-withdrawal penalty, tax hit or delay for money sitting in a retirement account or illiquid asset. Where the fund is actually held could make it slower or costlier to tap than the model assumes.