The Myth Ledger · Myth 07 · Wrong risk
“Cash and bonds are safe investments.”
Cash is stable in nominal units and high-quality bonds can match dated liabilities; neither guarantees long-run purchasing power or a fully funded retirement.
What survives scrutiny
Risk is failure relative to a goal. Stocks make account values visibly unstable. Cash hides inflation loss in a smooth nominal line. Long nominal bonds add duration and inflation risk; short bonds add reinvestment risk. Equities add severe drawdown and sequence risk. There is no safe asset independent of liability, horizon, currency, and behavior.
Evidence stack
What the literature can—and cannot—say
Tags distinguish measured evidence, economic foundations, model-dependent results, and important limits.
Economic foundation
The hedge depends on the liability.20
A short Treasury bill can closely match a near-term nominal bill. Inflation-indexed bonds can match real spending more directly. A long nominal bond can fluctuate sharply when yields or inflation expectations change.
Empirical evidence
Long horizons do not make any asset invulnerable.19
Broad international history reveals long real-loss episodes in stocks, bonds, and bills and warns against treating the unusually successful U.S. record as the only possible path.
Model result
A contested all-equity result is evidence—not a default.18,21
A 2025 working paper’s bootstrap lifecycle model favors globally diversified equities over conventional stock/bond glide paths on its objectives. The result challenges conventional age-based guidance; its working-paper status, assumptions, adherence demands, and treatment of household risks must travel with the conclusion.
Technical lens
Name the unit and horizon of safety.
Nominal volatility, real terminal wealth, interim drawdown, liquidity, and probability of funding consumption are different risk measures. Optimizing one can worsen another. Asset-liability matching starts by specifying cash amount, date, inflation linkage, and flexibility.
real return ≈ (1 + nominal return) ÷ (1 + inflation) − 1
Concrete example
A stable $100,000 can become a shrinking grocery cart.
At 3% annual inflation, unchanged cash buys about 74% as much after ten years. Yet putting next year’s rent in equities can force a sale after a crash. The solution is not one “safe” asset; it is matching layers of the portfolio to layers of the plan.
Decision checklist
A decision rule, not a slogan
Optional checkmarks stay in this browser. JavaScript enables saving; the decision rules below are always readable.
- Keep near-term required spending in instruments whose maturity and currency match the bill.
- Use inflation-linked assets for real liabilities where available and appropriate.
- Hold enough growth exposure for long liabilities, diversified to the level the household can actually keep through a crash.
- Stress-test inflation, rate shocks, bad early returns, longevity, and an inability to rebalance—not just annual volatility.
Optional lab: explore the assumptions
The default illustration is readable without JavaScript. Changing its inputs requires JavaScript; outputs are not forecasts.
Counterfactual lab · illustrative
Stable dollars, unstable purchasing power
Translate a smooth nominal account into the amount of today’s consumption it can fund.
Statement value
$155.8K
Today’s purchasing power
$100K
Annual real return
+0.0%
The account never shows a red year in this toy path. Yet its real outcome can still miss the goal. Conversely, volatility can be unacceptable when the bill is near even if long-run expected purchasing power is higher.
Model boundary · Constant yield and inflation, annual compounding, no tax, default, duration, reinvestment, or deposit-insurance limits.
Chapter sources: evidence and limits
Original source numbers are retained across chapters. Each finding travels with its limitation.
- 20Who Should Buy Long-Term Bonds? ↗
Campbell, J. Y. & Viceira, L. M. (2001). American Economic Review 91(1), 99–127.
Finding: Bond risk depends on the investor’s horizon and inflation exposure; inflation-indexed and nominal bonds hedge different liabilities.
Limit: The model abstracts from taxes, default, transaction costs, and behavioral responses to losses.
- 19Stocks for the Long Run? Evidence from a Broad Sample of Developed Markets ↗
Anarkulova, A., Cederburg, S. & O’Doherty, M. S. (2022). Journal of Financial Economics 143(1), 409–433.
Finding: International history reveals meaningful long-horizon loss risk that is obscured by focusing only on successful U.S. market history.
Limit: Historical country data have measurement challenges and cannot enumerate every future political or market regime.
- 18Beyond the Status Quo: A Critical Assessment of Lifecycle Investment Advice ↗
Anarkulova, A., Cederburg, S. & O’Doherty, M. S. (2025). SSRN working paper, revision dated 2025.
Finding: Historical block-bootstrap simulations across developed markets favor globally diversified all-equity allocations in the authors’ modeled lifecycle outcomes.
Limit: This is a working paper, not a universal prescription; results depend on historical resampling, utility and ruin definitions, adherence, and omitted real-world frictions.
- 21The Safe Withdrawal Rate: Evidence from a Broad Sample of Developed Markets ↗
Anarkulova, A., Cederburg, S., O’Doherty, M. S. & Sias, R. (2025). Journal of Pension Economics & Finance 24(3), 464–500.
Finding: A 38-country sample produces more severe retirement-spending risk than conventional U.S.-only backtests.
Limit: A constant real withdrawal and a binary ruin threshold are only one retirement objective; flexible spending, annuities, pensions, taxes, and bequests alter the problem.