When Bonds Stop Hedging: 150 Years of Stock–Bond Correlation Regimes

The standard advice for the defensive part of a portfolio is "hold government bonds: when stocks fall, bonds rise". That advice is built on 2000–2020, when the correlation between US stocks and Treasuries was negative in 96% of months. Over 1875–2026 it was positive in 58% of months; in December 2024 the 36-month correlation reached +0.72, the highest reading in the sample. The paper documents the stock–bond correlation from 1875 to 2026, shows that since the 1950s its sign has followed inflation along a gradient, and tests which real-time signals indicate that the hedge is broken. In months when stocks fell more than 3%, Treasuries beat cash in 65% of cases when the trailing correlation was negative, but in only 33% when it was positive. A rule that holds Treasuries only while the trailing correlation is negative would have avoided most of the bond losses of 2022–2026, but over 70 years it improved the risk-adjusted return only slightly (Sharpe 0.54 vs 0.51), and the improvement is not statistically significant. The paper also reviews 12 S&P 500 sell-offs since 1955 and runs 2,000 placebo tests. Practical conclusion: know which regime you are in, and do not let one asset carry the whole defensive job. A defensive sleeve combining Treasuries, cash and gold was more robust across regimes than any single asset or switch. InvestWealth White Paper No. 1, version 1.1 (data to 30 September 2026). Backtested results are hypothetical. Educational research, not investment advice.