Tax Matters

Tax-Efficient Portfolio Transition, or How to Rejuvenate Ossified Equity Portfolios

${ numberSection } ${ text }
Tax-Efficient Portfolio Transition, or How to Rejuvenate Ossified Equity Portfolios

In a recent paper “Tax-Efficient Portfolio Transition: A Tax-Aware Relaxed-Constraint Approach to Switching Equity Managers,” we discuss portfolio transition techniques that could alleviate the tax costs of switching managers for a taxable investor. 1 1 Close Sosner, Nathan, and Stanley Krasner. 2021. “Tax-Efficient Portfolio Transition: A Tax-Aware Relaxed-Constraint Approach to Switching Equity Managers.” The Journal of Wealth Management 23 (4): 31-57.   We find that transition to a tax-aware “relaxed-constraint” strategy 2 2 Close A relaxed-constraint portfolio construction relaxes the long-only constraint by allowing the strategy portfolio to hold short positions whose value amounts to a fixed percentage of the portfolio’s net asset value (NAV). To keep the sum of portfolio weights equal 1.0, the long leg of the portfolio is levered up by the corresponding percent of the NAV. For example, a relaxed-constraint 130/30 strategy holds 130% of the NAV in long stock positions and 30% of the NAV is short stock positions.   yields a meaningful pre-tax alpha and high tax efficiency both during and after the transition. This allows an investor utilizing such an approach to achieve high after-tax post-liquidation returns.

We focus on transitioning an appreciated Russell 1000 benchmark portfolio to an actively managed  strategy. 3 3 Close In the paper, we modeled hypothetical actively managed strategies as a value and momentum styles-based strategies. All the strategies we considered used identical hypothetical value-momentum signals.   We model transition for a more and a less appreciated pre-transition portfolio, with a built-in gain amounting to 60% and 40% of the portfolio value, respectively, and track the pre-tax and after-tax performance for 10 years following the transition.

We assume that all the stock positions of an appreciated hypothetical portfolio are taken over by a new manager who implements either a tax-agnostic or a tax-aware transition to either a long-only (LO) or a relaxed-constraint (RC) style-premia-based actively managed hypothetical strategy. Post transition, the hypothetical strategies are managed consistently with transition—tax-agnostic transition is followed by tax-agnostic portfolio management and tax-aware transition is followed by tax-aware management.



Figure 1 below 4 4 Close Figure 1 presents the same data as Exhibit 7, Panel B, in the paper.   depicts after-tax information ratios 5 5 Close Information ratio is a standard performance statistic computed as the ratio of a strategy’s active return (its outperformance over benchmark) divided by its tracking error.   of the hypothetical strategies we considered in the paper. It shows that after taxes, both LO and RC hypothetical strategies implemented as tax-aware outperform their tax-agnostic counterparts. While not shown here, in the paper, we show that there is some pre-tax information ratio degradation due to tax awareness. However, as Figure 1 demonstrates, it is more than compensated by greater tax efficiency. Moreover, the hypothetical tax-aware RC strategy has a meaningfully higher after-tax information ratio than all the other strategies.

Further in the paper, we show that our main finding (that transition to a tax-aware RC strategy results in the best after-tax performance) is robust to liquidation taxes and to having pre-transition portfolios that are significantly more concentrated than a Russell 1000 benchmark portfolio.


Figure 1. Hypothetical After-Tax Pre-Liquidation Information Ratio (IR) during a 10-Year Period Following the Transition

Source: AQR, XPressFeed, Russell. We model transition to hypothetical long-only and relaxed-constraint value-momentum strategies. The pre-transition portfolio is the Russell 1000 index. We model two levels of realized built-in gains, 60% and 40% of the hypothetical pre-transition portfolio value. We use a simplifying assumption that all the positions of the hypothetical pre-transition portfolio have the same built-in gain relative to their respective value—60% and 40%, alternatively, and that all the built-in gains are long-term. Our overall sample period is a 30-year sample period from January 1988 to December 2017. We use this sample period to construct five hypothetical 10-year investment periods whose start dates are separated by five years: 1988-1997, 1993-2002, 1998-2007, 2003-2012, and 2008-2017. Our stock universe approximately corresponds to Russell 1000 index constituents.  We assume a tax rate of 20% for long-term capital gains and dividend income and 35% for short-term capital gains. Taxes are assumed to be paid using funds outside of the strategy. The transition process is modeled as follows. For each of our five 10-year samples, on the last business day of the year before the sample begins, the hypothetical Russell 1000 pre-transition portfolio is rebalanced to a hypothetical value-momentum strategy portfolio which is either long-only or relaxed-constraint. On the last business day of the 10-year periods hypothetical strategy portfolios are fully liquidated. We use two types of rebalancing—tax-agnostic and tax-aware. The tax-agnostic transition is followed by a tax-agnostic portfolio management for the next ten years until liquidation, while the tax-aware transition is followed by a tax-aware portfolio management. All the net realized capital losses are carried forward and offset future realized capital gains.  At liquidation, carryforward losses offset liquidation gains and any losses remaining after offsetting the liquidation gains are ignored. After the transition, every month, a new hypothetical value-momentum alpha model is computed, and the hypothetical portfolio is updated to reflect this new hypothetical alpha model using an optimization procedure. The hypothetical value-momentum alpha model budgets its risk equally between value and momentum signals. The hypothetical portfolio maximizes exposure to this hypothetical alpha model subject to tax and transaction cost penalties and constraints on portfolio weights and portfolio beta. All the hypothetical value-momentum strategy portfolios are benchmarked to Russell 1000 index and are constructed with a 3% tracking error to this benchmark. The hypothetical long-only strategy is fully invested such that its portfolio weights sum to 100% of the portfolio value, only has non-negative portfolio weights, and has a predicted beta of 1.0 with respect to the Russell 1000 index. The hypothetical relaxed-constraint strategy has long portfolio weights that sum to 130% of the portfolio value and short portfolio weights that sum up to 30% of the portfolio value such that all the weights, long and short sum to 100%, whereas the absolute value of weights sums to 160%. Similar to the hypothetical long-only portfolio, the hypothetical relaxed-constraint portfolio has a predicted beta of 1.0 with respect to the Russell 1000 index. All the results are shown as averages across our five hypothetical sample periods.

Show more