With stock markets skidding down and bond yields dropping like a stone, many investors will likely find that the allocation to stocks in their portfolio is now below target and the allocation to bonds is above their target. Disciplined investors should regularly rebalance their portfolios to bring the allocations back to target but are likely wondering how they should go about it. Should they rebalance monthly or annually? Should they rebalance only when the current allocations are above or below some threshold? Or should they simply rebalance by directing new savings and portfolio income to the lagging asset class?

The authors of a Vanguard paper titled Best Practices for Portfolio Rebalancing attempt to answer these questions by analyzing the returns and volatility of a 60% stock-40% bond portfolio under different rebalancing strategies. The rebalancing strategies considered were based on time (monthly, for example), threshold or a combination of the two. Threshold rebalancing means a portfolio would be rebalanced only when the asset allocation has drifted from the target by a predetermined threshold. For example, a portfolio that has 20% in bonds and a 5% rebalancing threshold will be rebalanced when the allocation to bonds exceeds 25% or drops below 15%.

Perhaps unsurprisingly, the authors find that portfolios that were never rebalanced posted higher returns (with a higher SD) than portfolios that were rebalanced in some form. But it is surprising that the portfolios that were rebalanced based on time (monthly, quarterly or annually), threshold (0%, 1%, 5% or 10%) or a combination of the two produced similar returns with comparable volatility. The difference, of course, is that some rebalancing strategies resulted in a much higher portfolio turnover than others. For example, monthly rebalancing resulted in 1,008 events compared to just 58 for monthly rebalancing with a 5% threshold.

Real world investors care about such things as stock commissions, bid-ask spreads and taxes and will, therefore, prefer a strategy with low turnover. The rebalancing strategy with the lowest turnover was one in which portfolio income (such as dividends and interest payments) was invested in the underweight asset class. It may not be a practical strategy for portfolios with little or no inflows in an era when dividend yields and bond yields are so low but it should still work for investors who are regularly investing new savings.

This article has 16 comments

  1. This is a bit of a surprise to me because I was always under the impression that some re-balancing acted like a bit of buy low sell high in a relatively stable portfolio. For that reason I would have expected a better return where re-balancing is involved, particularly in a mutual fund scenario that has no fees for buying or selling.

  2. Very interesting result and I’m not sure what to make of it really.

    Difficult to determine underlying cause since they don’t show the exact magnitude of costs of rebalancing, there we can’t separate those from the timing factors.

    I wonder what the underlying correlation coefficients would look like during the rebalancing acts. For either equity or bond allocations to move >5% beyond your set allocation, markets must move significantly, and we know that correlations can temporarily increase during these times, therefore moving from bonds-equities or vice versa during volatile markets may yield no beneficial impact on portfolio returns nor reduce portfolio variation.

    Only by ‘rebalancing’ using new funds such as portfolio income, or new savings, is there a real timing benefit since in that case, you probably truely are adding to an asset class when it has dropped 5-10%, and that money has not suffered the correlated decline affecting all the asset classes.

    Maybe a further strategy would be to treat cash/savings as not being part of the portfolio, then mobilizing those funds when your allocations begin to shift. I do this to some extent, but don’t use thresholds.

    This study by Vanguard might suggest that rebalancing an exisitng portfolio is not necessary at all if you can stomach the risk, and then you use your asset allocation and deviation from the target as a timing method when adding to positions.

  3. @SavingMentor: It’s a common misconception that rebalancing boosts returns in *all* market conditions. It doesn’t. Imagine a market where stocks keep going up but bonds are more or less stable. In such a market, rebalancing means one is selling a winning asset class and buying a lagging asset class and result in lower returns than a never-rebalanced portfolio. Rebalancing, as the Vanguard article points out, is a risk mitigation strategy, not one that always boosts returns.

    @Sampson: You’re right that costs of rebalancing must be taken into account. Monthly rebalancing is therefore, probably overkill and returns net of costs are likely significantly lower.

    I haven’t seen any data supporting the theory that correlations between stocks and bonds goes up when markets move significantly. In fact, in market downturns, it is the opposite. Stock-bond correlations become even lower. Still, it does seem that a 5% threshold makes rebalancing a rare event. Between 1926 and 2009 monthly rebalancing with a 5% threshold resulted in just 58 rebalancing events or once approx. every 1.5 years.

    For me, the takeaway from the Vanguard study was to (a) rebalance to reduce risk (otherwise portfolio starts getting overweighted in stocks) and (b) to pick one method to rebalance and do it occasionally to minimize turnover.

  4. Looking through the figures and data once again, I’m really not sure to take away from it.

    At first glance, I would say, bring on the added risk for the marginal 0.6% annualized return from never rebalancing. However, the data used represents 83 years worth. “Never rebalancing” over a shorter timeframe may yield very different result due to the more significant impact of sequence of returns risk.

    It’s clear the authors want to avoid data mining, but this time frame somewhat misleads one to believe that one can extrapolate the 83 yr time frame to hold true for shorter more realistic time frames for an individual investor.

    • @Sampson: Check out pages 15 and above which look at recent rebalancing over the 21 year period from 1989 to 2009. Never rebalancing posted lower returns with higher volatility and the initial 60% allocation to stocks crept up to 70%. It would be interesting to see how different rebalancing strategies performed over 20-year rolling periods in the past and then look at the returns and volatility.

  5. Very interesting indeed. Good information for sure.

    I always rebalance by investing new money, saved money, into a lagging class. I keep my ETFs or stocks reinvested all the time, as much as possible. This way, I limit transaction costs (only one, to buy).

    @CC, how do you rebalance again?

    Rebalance seasonally?

    Only when allocation falls or rises beyond thresholds/targets? Adding new money only?
    Selling existing money?

  6. Interesting post. While rebalancing is an important part of an investing strategy, I think it is too often misunderstood. As CC points out, it is not really designed to increase returns, but to manage risk.

    The last 20 years were unusual in that bond returns and stock returns were very similar (about 8% to 9%). If you had a 60-40 portfolio in 1991 and never touched it, it would have still been about 60-40 at the end of 2010. But that’s not something one would expect over most 10- or 20-year periods.

    For people with smallish portfolios, rebalancing with new inflows is probably enough, expect perhaps after very big market moves.

  7. @ Canadian Couch Potato
    If the returns among bonds and stocks were similar over the past 20 years, why do you think the never rebalancing underperformed during that period (according to that Vanguard publication).

    As CC points out, shorter, and rolling 20-year time frames would be the most interesting. The difference in standard deviation in the never rebalance vs. any of the other groups is much smaller than I would have thought, so if one were to quantify the trade-off of ‘managing risk’ by rebalancing, and return, this study suggests to me it is much less advantageous than people make it out to be.

    • @Sampson: I somewhat disagree with you because while SD is useful in measuring volatility, it doesn’t tell the entire story. Stock returns as we all know exhibit fat tails. i.e. much more extreme events than you’d expect from a normal distribution. A 50% drop in equity prices, while rare is not uncommon. In that scenario, a 60/40 investor has a less than 30% loss, which is a significant improvement on the experience of the stock-only investor. Having said that, if one can stomach the volatility, then absolutely, a 100% stock portfolio should be a consideration. But I suspect limiting the downside will be important to most investors.

  8. Pingback: Best of Blogs – Empowering Edmonton | Retire Happy Blog

  9. Pingback: 3.18% 5-Year Fixed or P-0.75% Variable for MDJ Readers and Links | Million Dollar Journey

  10. Pingback: My Own Advisor » Weekend Reading – September 16th Edition

  11. I typically never offload previously invested dollars to rebalance, and focus mainly on investing new funds.

    An interesting post CC. When I first read the title to your posts, I didn’t expect Vanguard’s findings to yield what you have presented.

    I think CCP highlights an interesting point in that rebalancing can be very effective if it’s done properly by taking advantage of market swings.

  12. Pingback: Best Of The Blogs: In The Media Edition | The Wealthy Canadian

  13. Rebalancing is investing looking through a rear mirror, adjusting to What Is rather than what will be.