Vladimir de Vassal and Alex Atanasiu of Glenmede Investment Management wrote a special report for FactSet Insight titled Managing Equity Portfolios with Tax Efficiency. The report, published January 14th, highlights how quantitatively-based active strategies using a disciplined approach to harvesting losses can produce after-tax returns superior to traditional “passive” strategies.
Principles for a tax-aware portfolio strategy
Atanasiu and de Vassal recommend following three basic principles to develop a tax-aware portfolio strategy:
First — Make an effort to avoid taking short-term gains.
Second — Always take advantage of tax loss harvesting opportunities (ie, bear markets).
Third — Prioritize the selling of higher cost basis stocks.
In order to test different tax-sensitive strategies over time, Atanasiu and de Vassal modeled portfolios of randomly selected large cap stocks of the Russell 1000 Index over the last 20 years.
The first scenario included 1,000 portfolio simulations with the assumption that turnover would only occur from acquisition, trimming of oversized positions (>2% of total portfolio weight) or selling positions that have fallen to an insignificant weighting below 0.1% of the total portfolio. The second scenario made the assumption that all losses (short or long-term) would be harvested monthly. The third scenario simulated portfolios that harvested all losses with greater than one month holding periods (which would avoid the negative effects from one-month stock price reversions). The fourth scenario modeled portfolios that only purchased stocks ranked in the top three deciles of proprietary multi-factor stock ranking models (buy universe) and harvested losses for stocks that were no longer highly ranked.
Active loss harvesting strategy produces highest returns in historical modeling
The portfolio simulations in the first case with no tax harvesting had average gross and after-tax annualized adjusted returns of +10.2% and +9.6%, respectively. The annual turnover (21%) and taxable gains/credits were due to cutting oversized positions >2% (reduced to 2%) or selling positions that were less than 0.1% weighting in the portfolio. The proceeds from sales are proportionately reinvested in new positions (with total stock positions maintained at a constant 100 level) monthly.
The simulations were changed to harvest all losses on a monthly basis in the second case. In this strategy, the average gross return dropped to just +10.1%, but the average after-tax return jumped up to +11.2%. Of note, portfolio volatility ticked up to 16% versus 15.3% in the base case. The average turnover was 61% over the two decades.
In the third case, the simulations harvested all losses save for one-month losses that could see short-term price reversal effects. The average portfolio turnover was cut to 33% and the average gross return climbed to +10.4%. That said, the average tax-adjusted return was only +10.7%.
Atanasiu and de Vassal used a multifactor stock ranking screen that reflected “smart” beta exposures to valuation, fundamental and technical factors in the fourth case. Of note, only stocks in the top three deciles could be purchased. Any holdings that have a loss and are not ranked in the top three deciles of the stock screen were sold on a monthly basis. The average gross and tax-adjusted returns for the fourth case were +11.1% and +11.6%, respectively.
They note: “This strategy outperformed the harvest all losses strategy on an after-tax return basis by about 0.5% and should have lower transaction costs with an average turnover of 41% (versus 61% for harvest all losses).”