Rebalancing Revisited: The Role of Derivatives

Portfolio rebalancing trades off tracking error against the transaction costs associated with avoiding tracking error. Prior analytical work derived optimal rebalancing strategies that minimize the expected transaction costs required to achieve a given level of tracking error. Using these strategies results in the same level of tracking error as naive strategies often observed in practice but with much lower transaction costs. Additional (substantial) reductions in expected transaction costs can be obtained by using derivatives to synthetically rebalance a portfolio. The design of an efficient synthetic rebalancing program, however, is complicated. This article describes the key elements in such a complex design.Portfolio rebalancing strategies trade tracking error off against the transaction costs associated with avoiding tracking error. Prior analytical work derived optimal rebalancing strategies that minimize the expected transaction costs required to achieve a given level of tracking error by rebalancing the portfolio back to a no-trade region. Applying Monte Carlo simulation models, we show that using these strategies results in the same level of tracking error as naive strategies often observed in practice but with much lower transaction costs.Rebalancing portfolio positions by using derivatives (synthetic rebalancing) is potentially even more efficient than using the cash market for rebalancing because the cost of trading derivatives is considerably lower than the cost of trading cash assets. The design of an efficient synthetic rebalancing program is complicated, however, because the costs and tracking-error implications of using derivatives depend on how long the derivative positions are maintained. We show the key elements of the design of an efficient synthetic rebalancing program.The expected transaction costs of using synthetic rebalancing are less than the expected transaction costs of using strategies limited to the cash market that approximate the optimal strategies derived in the literature. When the desired tracking-error target is small, synthetic rebalancing can achieve the same level of tracking error for less than half the cost of cash-market-only rebalancing. As the tracking-error target increases, the expected transaction-cost savings from synthetic rebalancing decline. Synthetic rebalancing is particularly attractive when the desired tracking error is small because derivative positions are expected to be maintained for longer periods of time when the tracking-error tolerance is large.Using Monte Carlo simulation models to analyze rebalancing strategies, we document that substantial reductions in expected transaction costs can be obtained by using derivatives to synthetically rebalance a portfolio.When a portfolio receives cash inflows or is required to make periodic cash payments, an efficient rebalancing strategy allows the portfolio to hold some cash. When cash flows arrive, gaining the desired exposure by overlaying at least some of the cash with derivative positions is more efficient than investing the entire cash flow directly into the underweighted asset classes. Investing the cash flow into cash assets generally moves the asset allocations inside the no-trade region. The overlay strategy avoids trading cash-market asset classes inside the no-trade region, which (as previous literature has shown) is inefficient. The numerical optimization procedure determines the most efficient allowable cash position and no-trade region.Editor’s Note: David Brown consults from time to time for NISA Investment Advisors, LLC, which provides investment advisory services that include services relating to the strategies discussed in this article.