How Withdrawal Behavior Destroys Leverage Capacity: Evidence from Leveraged FIRE Portfolios

AbstractIn leveraged retirement portfolios, the method of withdrawal determines not merely the sequence of asset liquidation, but whether the portfolio retains its capacity to sustain leverage over time. We formalize Leverage Capacity (LC_t = C_t × h / CR_min) as an observable state variable in the context of leveraged retirement portfolios, representing the maximum margin position supportable by the collateral base at time t. Under Physical-First withdrawal (Strategy A), each withdrawal permanently shifts the future path of leverage capacity downward, progressively reducing the excess returns that leverage generates. Using a half-Kelly leveraged S&P 500 portfolio ($1,000,000 physical, $140,000 margin) with inflation-adjusted annual withdrawals of $40,000 (Bengen, 1994) and two independent simulation methodologies—stationary block bootstrap and GARCH(1,1)—we compare three strategies: Physical-First (A), Margin-First without Rebuild (C), and Margin-First with Rebuild (B). The liquidation-order effect (C vs. A) accounts for only 7% of the median terminal wealth difference, while the leverage-maintenance effect (B vs. C) accounts for 104%. Strategy B reduces portfolio depletion risk from 10.84% to 0.07% over 30 years. Results are robust to interest rate variation (2%–6%) and transaction costs up to 20 basis points.