In a prior post, I waxed philosophical about the importance of remaining skeptical of the assumptions that underlie portfolio construction.

Another commonly accepted assumption is that rebalancing is a value-add activity.  The underlying theory here is that assets tend to mean-revert over time, so by rebalancing on an annual basis, we can tilt our portfolio away from those expensive assets that have grown in relative portfolio weight and towards those cheap assets that have shrunk in relative portfolio weight.  Rebalancing institutionally annualizes the practice of "buy high, sell low."

But does it actually work?  Here's the test: we're going to use inflation-adjusted returns for the S&P 500 and a constant-maturity 10-year U.S. Treasury bond index.  We're going to construct a 50/50 portfolio and run it forward for 25 years (more or less the investment horizon for most investors).  For one portfolio, we're going to rebalance annually.  For another, we're just to just set it and forget it, letting the portfolio drift over time.  At the end of 25 years, we'll compare the total-return profile of both indices to see if rebalancing truly added value.

Note that I am not considering the tax implications or the cost of implementing rebalancing here.

Now, of course, because I am a timing luck nut, I had to perform this test for every month of the year to see if it made any difference when our annual rebalance occurred.  The results are summarized in the following graph:

Rebalancing v Drifting
Bars below 0% indicate that over the subsequent 25-year period, the drifting portfolio outperformed the rebalancing portfolio.  Positive bars indicate that rebalancing added value.

In retrospect, the results should come as no surprise: if your investment horizon occurred over a largely trend-driven period (e.g. late 1970s through 2000), rebalancing hurt you.  As equities ran, you would have wanted to overweight them: rebalancing kept them on a leash.  From 1975-2000, rebalancing may have actually cost you 100bps a year.  Note 1975-2000 was one of the best post-inflation return periods, with inflation-adjusted returns for a 50/50 portfolio eclipsing 7%.  $10,000 invested in 12/1975 in a 50/50 and rebalanced every year would have come $60,868.  Not bad – but not the $70,134 you could have ended up with if you hadn't rebalanced.

On the other hand, if you weren't lucky enough to invest from 1975 - 2000, and your investment period contained a largely mean-reversionary period – like 2000-2010 in U.S. equities were – then rebalancing would add value as you sold at highs and bought at lows.

On average, the results are pretty mixed:

Average Rebalanced v DriftedExcept for September, which seems to be anomalous and squarely in the camp of rebalancing.  Unfortunately, we haven't discovered the holy grail of rebalancing alpha – we just got lucky (damn you timing luck!).  Digging a little deeper, from 9/1986 - 9/1987, the real return for the S&P 500 was 37.3% and for our constant maturity Treasury index, it was -10.4%.  Therefore, over this period, a 50/50 portfolio would have drifted to become nearly a 60/40 portfolio.  By rebalancing back to a 50/50 before 10/19/1987 crash, we avoid a 20% loss on the excess 10% stock allocation – or about 200bps.  Now, that doesn’t necessarily explain all of the 20bps annualized – but given that September has been a historically tumultuous month, a couple more events like this throughout the years probably added up.  Or, that's my hypothesis at least.

So does the practice of annually rebalancing actually add value?  I'm squarely in the camp of "maybe."

As much as I've tried to isolate the impact of rebalancing here, I've potentially simplified too much, ignoring the fact that investors contribute to their portfolios over time and tend to move from aggressive to conservative risk tolerances as they age.  I've completely ignored the impact of other asset classes as well.  The example portfolios tested here may be sufficiently unrealistic to warrant zero practical take-away.

But it does highlight an important philosophical distinction between the two methodologies.  Annual rebalancing bears the risk that the market will mainly trend during our investment period.  Allowing a portfolio to drift bears the risk that a significant portion of our investment horizon will contain mean-reverting performance.  Risk-seeking investors may actually prefer the latter risk to the former, as missing out on returns hurts more than increased losses.  Risk-averse investors, on the other hand, would prefer the former.  Perhaps we should not only consider whether investors should move from aggressive to conservative portfolios over their investment horizon, but whether they should allow their portfolios to drift in early years and rebalance it in later years.

EDIT: In the comments below, a question was asked about the impact on risk in rebalancing versus drifting.  In the graph below, I plot the difference in annualized volatility over the 25-year periods of the rebalanced portfolio versus the drifted portfolio.  We can see that the rebalanced portfolio consistently offers a reduced volatility profile.

Rebalanced v Drifted Volatility

Corey is co-founder and Chief Investment Officer of Newfound Research, a quantitative asset manager offering a suite of separately managed accounts and mutual funds. At Newfound, Corey is responsible for portfolio management, investment research, strategy development, and communication of the firm's views to clients.

Prior to offering asset management services, Newfound licensed research from the quantitative investment models developed by Corey. At peak, this research helped steer the tactical allocation decisions for upwards of $10bn.

Corey is a frequent speaker on industry panels and contributes to ETF.com, ETF Trends, and Forbes.com’s Great Speculations blog. He was named a 2014 ETF All Star by ETF.com.

Corey holds a Master of Science in Computational Finance from Carnegie Mellon University and a Bachelor of Science in Computer Science, cum laude, from Cornell University.

You can connect with Corey on LinkedIn or Twitter.