Much of the below discussion draws from the chapter on Dynamic Portfolio Choice from Andrew Ang’s Asset Management — a book I highly recommend
Comparing the return performance of rebalanced and “buy-and-hold” portfolios says little about the economic value of one strategy or the other. In a recent blog post, Corey Hoffstein shows that returns to rebalanced portfolios underperform in times of trends, and as such finds the evidence in favor of rebalancing to be quite mixed. But because there is no explanation of why rebalancing makes sense and how it creates value, return data tells relatively little about whether you should rebalance or not.
First let’s set the stage. Suppose you’re a long term investor with a million dollars and a time horizon of 25 years. And suppose that in this world the distribution of returns is assumed to be the same in each period, (iid), that you can trade once every year, and that you do not have access to any “inside tips” that would tell you any information about the distribution of stock returns. The first assumption is reasonable because if everybody knew returns were going to be high tomorrow, they would just buy today. The second is just for simplification, and the third one reflects the situation of most individual investors without access to any kind of special research team (not that institutional investors have access to the secret sauce either)
In this world, there are two broad classes of trading strategies: static and dynamic. A static strategy just buys and holds. Set up your 60/40 stocks/bonds portfolio and wake up 25 years later, collecting whatever capital gains you get. A dynamic strategy changes how much to buy and sell in each period depending on your wealth. Rebalancing refers to the dynamic strategy of starting out at a certain portfolio allocation, say 60 percent stocks and 40 percent bonds, and in each period selling or buying to go back to this allocation. So if stocks do really well one year while bonds do poorly (say in 2013), then sell some stocks and buy more bonds.
Apriori, it would be strange if a dynamic trading strategy were dominated by static trading strategy. Under the conditions I described above, rebalancing does dominate because it allows you to reset your risk exposure every period.
To see why, change the perspective from thinking about the trades you’re doing and instead think about the decision to change your risk exposure over time. Start from the assumption that your risk aversion does not change over time. If you fail to rebalance, as the share of stocks in your portfolio increases over time, you are choosing to increase your overall exposure to your stock market. But wait! If if it’s optimal for you to hold that riskier balance in the second year, then you should have held that in the first year! Expected returns haven’t changed, nor have your preferences. So don’t let the name of “buy and hold” fool you. Just because you’re not trading under a buy-and-hold strategy, you are making an active choice about your risk exposures.
But why does rebalancing create value? The market in aggregate cannot rebalance —if you’re selling stocks to rebalance, somebody has to be buying them from you. If you’re earning a rebalancing premium, there has to be some reason why others are willing to pay you.
Here is a point that Corey does hit on — rebalancing doesn’t work if prices are trending upwards over a long period of time. Another way of putting it is that if the distribution of returns changes over time, a strategy of rebalancing will have you buying more into stocks that may be worth nothing one day. In other words, rebalancing is a strategy that is short regime changes. Rebalancing works great if tomorrow’s returns look like today’s. It’s a horrible idea if, like in 1990’s Japan, stocks collapse and don’t come back even after 20 years.
And here’s why there is a long run rebalancing premium — you’re protecting others from extreme changes in the market environment. So rebalancing is good for keeping risk exposures on time, but at the risk of black swan events that permanently change the landscape of returns.