On average, at least two sectors had negative returns over a one-year period, whether S&P 500 returns were positive or not. During sizeable market downturns, all sectors have tended to be in the red. In 249 of the 347 months analyzed, at least one sector had a negative return—that’s a whopping 72% of the time.
Expanding the frequency to daily returns, not monthly, naturally increases the probability. On 5,289 of 7,280 days analyzed, or 73% of the time, at least one sector was at a loss over the past year.
The point is, a sector replication strategy offers real potential for tax loss harvesting events—this is not just a quixotic dream or the imaginings of an ETF pundit.
Constructing a disaggregated broad market exposure
Constructing a disaggregated market-cap weighted exposure requires a reference market and associated weights. We used the S&P 500 for broad beta exposure, capturing historical sector weights on a semi-annual basis going back to 2007. This allowed us to include the events of the 2008 financial crisis in our analysis, since a 10-year lookback would be too short. Semi-annual periods were used as that will be the assumed rebalancing period for the following portfolio blend.
Rebalancing frequency is an important consideration, as we’ve discussed previously. For simplicity’s sake, we chose semi-annual (June, December). While an investor could elect to rebalance daily based on having a sector position at a loss, our focus was on replicating broad S&P 500 Index returns with the underlying sectors. More frequent rebalancing (e.g., quarterly, monthly) would reduce slippage but increase transaction costs. In addition to its obvious simplicity, semi-annual rebalancing may limit slippage while also mitigating transaction costs.
With respect to rebalancing, we did explore one other question: Why not buy and hold the sectors at their respective weights in 2007 and never rebalance?
In other words, if the sectors themselves are market-cap weighted, shouldn’t their composition change just as it would in the broad S&P 500? Wouldn’t the sector weights match the S&P 500 at all times based on the broad sector’s own return path? There are no sector constraints in the S&P 500 after all. It should all move together, right? Wrong.
First, a buy-and-hold approach wouldn’t capture changes sectors had undergone more recently with the GICS reclassification in September 2018. Second, corporate actions, as well as additions and deletions to the S&P 500, change sector weights over time.
For example, if an Industrial firm is removed from the S&P 500 and a Technology firm is added, sector weights outside of market appreciation impacts will change since there are now more tech firms and fewer industrial. We included a buy-and-hold portfolio in our analysis to show the impact of these events.
Replicating broad beta with sectors
Using sector-based indexes, we looked at the performance of two hypothetical blended portfolios of sector-based exposures over the time period December 2007 to June 2019—one buy-and-hold portfolio and one rebalanced semi-annually.
Returns on a long-term basis for both portfolios match the return of the S&P 500 Index quite closely. Beyond the returns as shown below, correlations and beta sensitivity of both portfolios were 1.