Rebalancing, Returns, and Luck in the Last Decade

Fueled by accommodative monetary policies and favorable starting points, 2019 was the year when almost everything went up. And not up by just a little, but up by a significant amount. Both bonds and stocks posted some of their best returns in over a decade.¹ It should be no surprise, then, that the standard 60/40 portfolio asset allocation mix of global stocks and core aggregate bonds posted its best returns in a decade.

Just how good those returns were, however, depended on the rebalancing schedule and frequency—not just in 2019, but throughout the decade. The difference in return patterns is something referred to as “timing luck”² and is something I have discussed in a previous post.

Luck in the last decade
We’ve all seen the headlines about “the best return in a decade” for the standard 60/40 portfolio. But would this hold true across different rebalancing schemes? We looked at monthly, quarterly, semi-annual, and annual frequencies, with the schedules differentiated by “when” they were rebalanced (e.g., semi-annual January to July versus semi-annual February to August). The naïve standard allocation consisted of just two exposures: global stocks—as measured by the MSCI ACWI IMI Index (“ACWI”)—and core aggregate bonds—as measured by the Bloomberg Barclays US Aggregate Bond Index (“Agg”).

In total, we looked at 22 different return streams. While each iteration did in fact notch the “best return in a decade” title, the magnitude at which they jumped over the previous level was different. As shown below, the difference between the 2019 return and the prior decades’ record high (2017 was the prior record high in the 2010s; 2009 was the high-water mark over the last 30 plus years) ranged from a mere 40 basis points (or 0.40%) to upward of 4%.

With differing returns compared with prior records, this naturally led me to wonder how the calendar year returns themselves differed. Did the ACWI/Agg blend rebalanced semi-annually at month end in May and November simply have a great 2017 while others didn’t? One way to quantify intradecade yearly return differences is to examine the dispersion during the different years. For 2019, the difference between the best- and worst-performing blend was 76 basis points (bps). Over the past decade, this could be considered low. Looking at the following chart, the average yearly return dispersion across the 22 different return streams was 2.3%.

Some have luck on their side
Not surprisingly, given the record-to-record calculations shown earlier, 2017 had a lot of differentiation across return streams. Why were the return streams so different in 2017? It all comes down to timing.

Below are the rolling monthly, three-month, and six-month returns for the Agg. The chart is localized on the 2016-2018 timeframe to show the differences more acutely, while the average line is based on returns from 2009 to 2019. Following the 2016 US election, the markets were gripped with reflation-euphoria and bond yields spiked – sending bond prices lower. The euphoria was short lived, and yield movements became less volatile for the rest of 2017, ending the year only 4 bps from where they started.³ Portfolios that rebalanced in the fourth quarter of 2016—more specifically, in November or December—essentially bottom-ticked the bond market—not because of any clairvoyant insight on where rates were going, but because the calendar said it was time to trade. The favorable starting point led to returns compounding at a higher rate.

Adding complexity increases the impact of luck
For a naïve, standard asset allocation mix that is the typical bedrock of a well-diversified portfolio, this is a lot of differentiation based solely on “when” to trade. Remember, this is only two exposures. We ran the same analysis by extending the asset allocation mix to break out these exposures by specific regions (US, Developed Ex-US, Emerging Markets) for equities and adding in high yield bonds on the fixed income side. When more assets are added, the impact intensifies, with the average dispersion within the calendar years going from 2.3%, as shown above, to 2.9% across all blends.

The end result is that while the 2019 figures are only 79 basis points apart, the full-decade results for all 44 portfolios (just ACWI/Agg and the regional blends) differ by 15.8% on a cumulative returns basis. For a $10,000 investment at the dawn of the 2010s, this would equate to a $1,600 difference (16% of starting principal) and a noticeable terminal wealth difference.

Dealing with luck
In the prior post, I discussed some potential remedies:

• Create overlapping portfolios that are rebalanced at different times (i.e., tranching);
• Target rebalancing to fixed percentage weights based on bands (e.g., equities can’t be more than 65%); or
• Rebalance monthly, striking a balance between exposure management and costs.

Regardless of the rebalancing method chosen, the frequency selection should always be emphasized during the due diligence and performance attribution process for any given strategy. Investors should ensure that they understand the rationale for the defined rebalancing approach. And for a multi-asset allocation portfolio implemented with ETFs, focusing on the trading cost component of total cost of ownership of the ETF is more crucial than selecting a product based on expense ratios. Trading volume, bid-ask spreads, and depth of market should all be part of the process in evaluating the implementation flexibility needed to mitigate the impact of timing luck when rebalancing.