Week 38 Day 6: The Retirement Stress Test: What Happens When Things Go Wrong

Your plan: retire at 60 with $1.5 million, withdraw $60,000/year. In a normal market, this works fine. Now stress test: what if the market drops 40% in year 1? Your $1.5M becomes $900,000. Your $60,000 withdrawal is now a 6.7% rate -- dangerously high. Can you cut spending? Do you have a backup income source? Stress testing forces you to plan for bad luck.

Stress test scenarios: (1) Market crash in year 1. Your portfolio drops 40%. Withdrawal rate spikes. Question: can you reduce spending by 20-30% for 1-3 years while the market recovers? If not, you need a larger portfolio or lower baseline spending. (2) Inflation spike. Prices rise 7-8%/year for 5 years. Your $60,000/year spending becomes $84,000 in real terms. Are your investments indexed to inflation (VTIP, SCHD dividend growth, Social Security COLA)? If your portfolio is all nominal bonds, inflation eats your purchasing power. (3) Health emergency. A major surgery costs $50,000 out of pocket. A long-term care event costs $100,000+/year. Do you have a health savings account (HSA), long-term care insurance, or a dedicated healthcare reserve? (4) Divorce. Assets are split. Your $1.5M becomes $750,000. Your income sources are reduced. Can your plan survive at half the asset level? (5) Longevity. You retire at 60 and live to 100 -- 40 years of withdrawals. At a 4% withdrawal rate, your portfolio needs to survive 40 years, not 30. Historical data: the 4% rule survived every 30-year period but fails in some 40-year periods. Stress test response: for every scenario, identify (a) the probability it occurs, (b) the financial impact, and (c) your mitigation strategy. If you cannot identify a mitigation strategy for a plausible scenario, your plan has a vulnerability that needs to be addressed before retirement.

Formal retirement stress testing uses Monte Carlo simulation (detailed in Week 45) to model the joint probability distribution of returns, inflation, longevity, and spending shocks. The standard outputs: (1) probability of success (portfolio lasts through the planning horizon), (2) expected bequest (residual portfolio value at death), and (3) conditional probability of ruin (probability of running out of money, conditional on living to age X). Blanchett, Kowara, and Chen (2012) showed that the standard retirement planning assumption (fixed 30-year horizon, 4% withdrawal, 7% nominal returns) produces dramatically different outcomes when the variables are modeled stochastically: the 95th-percentile outcome (good luck) leaves a $3M bequest, while the 5th-percentile outcome (bad luck) depletes the portfolio in year 18. The range of outcomes -- not the average -- is what matters for stress testing. Finke, Pfau, and Blanchett (2013) showed that in a low-return environment (bond yields below 2%, equity premiums below 5%), the historical 4% rule fails in approximately 50% of simulated 30-year horizons. This suggests that stress testing should include a 'low-return world' scenario that is increasingly plausible given current valuations and yields. The comprehensive stress test matrix: (returns: low/medium/high) x (inflation: low/medium/high) x (longevity: 25/30/35/40 years) x (healthcare: normal/catastrophic) x (spending: fixed/flexible) = 180 scenarios. A retirement plan that works in 85%+ of these scenarios is robust.