Portfolio rebalancing is an important practice that needs to adhere to a client’s investment policy statement, and rebalancing may serve as a tool to generate a higher growth rate. Rebalancing that aims to buy low and sell high can harvest excess return from the volatility of the underlying assets. There is the risk of underperformance as a result of differences in the growth rates of the assets; but over the longer term, rebalancing portfolios appears to have a higher growth rate.
What’s Inside?
The authors study the concept of volatility harvesting, with a focus on quantifying the potential long-term performance advantage and the risk of underperformance in the short run. They introduce a formula to decompose excess returns versus the market into three parts: volatility return, dispersion return, and drift return. The formula can be used for any portfolio strategy and does not require stochastic modeling assumptions. The formula is adjusted to acknowledge multiple factors that affect the level of concentration in the market. The concept is easier to understand and more useful when dealing with historical datasets.
How Is This Research Useful to Practitioners?
Traditionally, alpha is generated from selected securities that are expected to outperform thanks to a manager’s skill or the factor risk premium. Volatility-harvesting formulas provide investors with a practical tool for understanding a different kind of outperformance. Rebalancing can be a dynamic allocation strategy that relies on the long-term stability of the capital distribution and the presence of volatility.
Volatility return is defined as the difference between the portfolio growth rate and the weighted-average asset growth rate. The volatility return increases with higher levels of cross-sectional volatility. In volatility harvesting, asset weights are important for measuring risk. A constant-weighted portfolio close to the market weights tracks the market well, whereas an equal-weighted portfolio that is more diverse faces higher relative-return risk. For a constant-weighted portfolio, the changes in the distance between the portfolio and the market are entirely the result of market weights moving, which is caused by dispersion of asset returns; this part of the excess return is called “dispersion return.” If the portfolio is positioned well, it will have more weight in securities that have higher growth rates. If not, the active bets will be a drag. Drift returns can cause portfolio concentration. This part of the formula is the only part over which portfolio managers have significant control.
In simple terms, a constant-weight strategy may have dispersion return but no drift return, whereas a buy-and-hold strategy that does not rebalance may have drift return but no dispersion return. Taking all three returns together, the authors conclude that the volatility return will eventually overtake the dispersion return and produce a positive excess return. If a portfolio manager does not allow the drift return to be more negative than the volatility return, any dynamic strategy can become a volatility-harvesting strategy. For the strategy to work, dispersion and drift terms need to remain bounded so that the volatility term can accumulate over time.
How Did the Authors Conduct This Research?
The authors translate Pal and Wong’s (working paper 2013) framework of energy–entropy decomposition into volatility return terminology to make it easier to understand. They use the natural logarithm to turn a simple return into continuously compounded growth rates to facilitate the return decomposition process. The extra return is split into three parts:
Excess log return = Volatility return – Dispersion return + Drift return.
When the assets in the market have returns that are different from one another, the volatility return is always positive. Dispersion return is related to the long-term active bets of the portfolio, and drift return occurs when the weights are allowed to drift.
The authors illustrate the benefits and risks associated with volatility harvesting using monthly S&P Global Broad Market Index data for 20 developed and emerging market countries from March 1997 to May 2015. They construct historical market weights by drifting the weights back in time based on the total returns of the country indexes. They then compare four portfolios: capitalization weight, equal weight, equal weight drifted, and equal weight ±2% trigger. All three equal-weighting strategies outperform the capitalization-weight portfolio because of rebalancing. In addition, trigger-based rebalancing is more effective than calendar-based rebalancing and has lower transaction costs than strategies that rebalance more frequently. The authors also show that during extended asset bubbles, diversification and rebalancing can be painful. All strategies outperform a buy-and-hold portfolio with lower risk than the market. Buy-and-hold and drifting portfolios have the most negative drift returns, which offset all or part of the volatility return. Total returns and volatility returns are maximized with strategies that regularly rebalance.
Abstractor’s Viewpoint
Rebalancing is important in portfolio management, but constant rebalancing comes at a cost and is simply unrealistic. This study is an improvement in analyzing benchmark-related risks during rebalancing. Using the formula, price or mean reversals are not a requirement for a rebalancing strategy to outperform. But, as cautioned by the authors, it is crucial to remember that rebalancing does not always work because dispersion return can overtake the excess return for a long period of time, especially in extended asset bubbles. Although the theory may be sound, in practice, investors should exercise sound judgment and discipline during rebalancing.