Rebalancing is a topic held near and dear to the heart of many investment advisors and often forms a key building block of what the advisor perceives as his or held value add. If diversification is viewed as a free lunch then rebalancing is often viewed as free dessert.
Unfortunately while the benefits of true diversification stand the test of time the same cannot be said about rebalancing especially when done in a highly automated way such as frequently seen in many robo-advisor offerings.
What do we mean by rebalancing? There are many reasons why investment advisors intentionally shift the weights of their portfolio components over time. The manager may have a changed view on future expected returns or they may deem a certain investment to now possess an unacceptable level of risk and thus wish to cut back on the exposure. Others may wish to more closely track a commercial index and do a wholesale portfolio rebalance to minimize deviations from benchmark weights.
In general there are two basic underlying motivations for rebalancing. The first is informational – the weighting structure of the portfolio is consciously changed to reflect an updated perspective on forward-looking estimates of return and/or risk.
The second basic motivation for rebalancing is purely driven by an automated rule that is agnostic to changing views on capital market risk and return conditions. The rebalancing rule is followed religiously regardless of the capital market environment.
This second form of rebalancing and the focus of this note we refer to as a constant mix strategy. Constant mix strategies take their cue solely from relative asset price changes over a chosen period of time and seek to bring portfolio weights back to a set of pre-defined time-invariant weights.
Why are many automatic constant mix rebalancing strategies viewed as a free dessert? Constant mix strategies trim relative winners and top up relative losers. Assuming fairly priced asset classes at the beginning of the period a constant mix rebalancing approach can be construed as a value-oriented strategy buying up cheap assets and selling expensive assets.
The working assumption is that capital market forces will correct relative asset class mis-pricings. Under this scenario markets are constantly displaying mean-reverting behavior. Asset classes that go up must come down and vice versa. Kind of a rollercoaster of relative performance.
Rebalancing programs will pick up nickels and dimes by selling relatively expensive assets and investing the proceeds into relatively cheap securities. Transaction costs and short-term capital gain taxes will be incurred but according to proponents of such rebalancing programs, such costs will be offset by gains in portfolio values and as a side benefit – the icing on the cake – a reduction in portfolio risk. Risk in this context is frequently equated with downside volatility as valuation changes are viewed as the primary cause of mean reversion in asset prices.
The idea of buying low and selling high is alluring and under a mean reverting type of market a constant mix strategy would deliver on the promise. But the romantic notion of mean reverting markets often fades away as real capital market behavior with all of its flaws sets in.
Do real capital markets behave in the idealized mean reverting manner necessary for constant mix strategies to shine? Capital markets are ever changing as asset class fundamentals, risk characteristics and investor demand fluctuate for perfectly sensible as well as irrational reasons. Capital markets are constantly calibrating supply and demand conditions and over short-term intervals changes in investor sentiment tend to overpower fundamentals.
There is also a lot of both academic as well as practitioner research on time series momentum. In general the conclusion is that over intermediate time frames such as 12 to 24 months there are trends in asset class performance. Mean reversion as a concept has been thought of more as a 3 to 5 year concept but over the typical time periods used in rebalancing programs (monthly or quarterly) the evidence is more consistent with trends in asset class prices.
Apart from taking advantage of mean reversion patterns in asset class performance how does rebalancing really work? In a recent paper titled Rebalancing Risk, Granger, Greenig, Harvey, Rattray and Zou derive analytical formulas as well as provide empirical evidence on the behavior of constant mix strategies versus an approach that allows asset class weights to drift depending on relative performance.
In the paper they neatly decompose the performance drivers of the constant mix strategy. What they show is that the constant mix strategy can be decomposed as follows:
Constant Mix Portfolio
Buy & Hold Portfolio
Selling Put and Call Options on Relative Asset Performance
The constant mix strategy carries negative convexity meaning that it outperforms when asset class return divergences are small by collecting the premium received for selling the puts and calls on the underlying portfolio.
However, when asset return divergences are large the selling of puts and calls results in losses and the buy and hold portfolio out-performs. In extreme cases when one asset class significantly lags the rest of the portfolio such as in a stock market crash the negative convexity of the strategy can magnify the drawdown.
Another way to think of the constant mix strategy is to put yourself in the shoes of the investor selling puts and calls. When volatility rises holders of the options will benefit and commensurately sellers of options will see an economic loss. That is why the constant mix rebalancing approach is often described as a short volatility strategy.
The authors of Rebalancing Risk conduct a set of empirical and simulation exercises using US stock and bond data. What do they find?
- Constant mix portfolios exhibit small positive returns when relative asset class performance (stocks versus bonds in this example) do not perform too differently from each other.
- When there are large differences in asset returns the tendency is to exhibit losses relative to buy and hold portfolios. The losses are magnified during periods of extreme relative performance.
The basic prerequisite for constant mix strategies to out-perform buy and hold strategies are reversals in asset class performance. From a risk perspective, constant mix strategies add an element of risk as the approach involves additional option related features. Justifying the use of a constant mix approach by implying risk reduction benefits is theoretically and empirically unfounded.
Access the full report here That Free Rebalancing Dessert Can be Costly!
Eric J. Weigel
Managing Partner, Global Focus Capital LLC
Feel free to contact us at Global Focus Capital LLC (mailto:firstname.lastname@example.org or visit our website at http://gf-cap.com to find out more about our asset management strategies, consulting/OCIO solutions, and research subscriptions.
DISCLAIMER: NOTHING HEREIN SHALL BE CONSTRUED AS INVESTMENT ADVICE, A RECOMMENDATION OR SOLICITATION TO BUY OR SELL ANY SECURITY. PAST PERFORMANCE DOES NOT PREDICT OR GUARANTEE FUTURE SIMILAR RESULTS. SEEK THE ADVICE OF AN INVESTMENT MANAGER, LAWYER AND ACCOUNTANT BEFORE YOU INVEST. DON’T RELY ON ANYTHING HEREIN. DO YOUR OWN HOMEWORK. THIS IS FOR INFORMATIONAL PURPOSES ONLY AND DOES NOT CONSIDER THE INVESTMENT NEEDS OR SUITABILITY OF ANY INDIVIDUAL. THERE IS NO PROMISE TO CORRECT ANY ERRORS OR OMISSIONS OR NOTIFY THE READER OF ANY SUCH ERRORS