Bending the return curve with rebalancing using trend-following

Rebalancing has become an important tool for portfolio management. Nevertheless, the return impact of rebalancing will be affected by market return patterns. Regular rebalancing is a mean-reverting strategy. For example, if there is a simple 60/40 stock/bond portfolio, stronger stock performance will cause the allocation to deviate from the strategic allocation and lead to a higher allocation to stocks over bonds. A rebalance will take money away from the better performing asset and give to the underperforming asset. It sells winner and buys losers on a regular basis. See “Strategic Rebalancing” by Granger, Harvey, Rattray, and Van Hemert

Rebalancing may seem like an innocent enhancement strategy under normal times, but if there are trends in markets, rebalanced portfolios will show poorer performance than a buy and hold strategy. It will affect drawdowns and volatility. Take the simple market extreme during the Financial Crisis, rebalancing would have continued to add exposure to stocks even during the extended downturn and take money away from the better performing bonds. In essence, a rebalance strategy generates negative convexity for a portfolio.

A frequent rebalancing strategy is like selling straddles. Alternatively, trend-following is similar to buying straddles. Hence, incorporating a trend component with rebalancing will change the portfolio return convexity. There may be a middle ground whereby rebalancing is done in conjunction with following trends.  This may allow the best of both strategies. If markets are trending, delay rebalancing. If they are not trending, follow a regular rebalancing schedule. 

There is a simple marriage between rebalancing and trend-following that is consistent with holding winners and not giving to losers. This will offset the impact of rebalancing drag. While the impact of strategic rebalancing will be minimal during periods of calm, there will be strong benefit if we move to extremes. Applying discipline will allow for positive portfolio convexity at the right times.