An Empirical Evaluation of Tax-Loss-Harvesting Alpha (Summary)

Retail investors can increasingly access tax-loss harvesting, an investment strategy that sells losing positions in securities, generating a tax credit that can then be reinvested. Because short-term capital gains are often highly taxed compared with longer-term gains, the strategy crystallizes losses earlier and reduces the overall tax bill.

Until recently, tax-loss harvesting was available only to investors with large portfolios where the benefits of the strategy outweighed the costs. The strategy is increasingly available to retail investors, thanks to falling transaction costs, narrowed bid–ask spreads, and advances in technology, which have engendered a raft of fintech startups offering automated tax-loss services.

Now that retail investors can access tax-aware strategies, they wish to know the potential returns under different market conditions. This article explores how much value can be gained from tax-loss harvesting under current US tax regulations and in a variety of markets.

Each month, the authors’ simulated tax-loss-harvesting strategy sells all securities showing losses and repurchases the securities immediately after. The strategy is applied from July 1926 to June 2018 to a portfolio of the 500 largest companies by market capitalization.

The authors calculate the performance of the tax-loss-harvesting strategy over four 23-year periods: 1926–1949, 1949–1972, 1972–1995, and 1995–2018. Each period illustrates a different type of market, ranging from recession and high volatility to expansion and low volatility.

The authors start by assuming negligible transaction costs and ignoring the “wash-sale rule,” which prohibits investors from immediately repurchasing shares that have been sold at a loss. They assume that the tax credit created by harvesting losses can be immediately reinvested, and they limit their analysis to long-only strategies. The authors apply a 35% marginal tax rate to short-term capital gains and a 15% rate to long-term capital gains and dividends.

The authors then incorporate various practical realities into their analysis, such as the potential impact of the costs of transactions, turnover, and liquidation and of imposing the wash-sale rule on the strategy to determine how these and different tax rates affect a retail tax management strategy.

From 1926 to 2018, the tax-loss-harvesting strategy delivered an average annual alpha of 108 bps over and above a passive buy-and-hold portfolio.

Tax-loss harvesting improves equity returns during each of the four discrete periods. The strategy performs particularly well when stock markets are highly volatile and more opportunities to harvest losses are available. For example, for the 1926–49 period, which encompasses the Great Depression, the average alpha is approximately 2.13% a year. Given that few investors would have made any gains to offset against losses, tax credits could have been carried forward to offset future gains.

The strategy matches the passive benchmark only during periods of reduced volatility and economic expansion. During 1949–1972, which was punctuated by post–World War II economic expansion, average alpha is just 0.51% a year. The 1972–95 and 1995–2018 periods fall between these two extremes, with average alphas of 1.08% and 0.81% a year, respectively.

When applying the wash-sale rule, the authors found that yearly alpha fell from 1.08% to 0.82% across the full time series. Likewise, including transaction costs would subtract 13 bps from the strategy, reducing alpha to 95 bps a year.

Substantial alpha is available to retail investors by using tax-loss-harvesting strategies, even after costs.

Tax-loss-harvesting strategies are more profitable in certain market environments: lower market return (i.e., with more losses to harvest), higher market volatility, and greater cross-sectional dispersion of returns for individual securities.