Measuring Up: Benchmarks and Your Portfolio

An investor ought to be able to get answers to two simple questions: how has my portfolio performed, and what have I paid for that performance? Though there is plenty of arithmetic that goes into producing them, the answers can and should be as straightforward as the questions.

A reasonable next question is “Why did my portfolio perform that way?” An insightful answer requires some context. We need a yardstick, reference, or “benchmark” to make an informed commentary on why a portfolio performed as it did.

bench·mark, n. a standard or point of reference against which things may be compared or assessed.

If your capital were all on deposit in a bank account or term deposit, you would have very different expectations for its performance than if it were invested in a portfolio of stocks. You might anticipate that the stock portfolio would earn a higher return than the interest rate on the deposit. For sure you would expect the month-to-month returns on the stock portfolio to be less predictable than the interest rate.

But that’s a pretty simplistic and blunt comparison. We need something a bit more informative and nuanced. For example, is the bank interest rate higher or lower than the rate of inflation? In this comparison, the inflation rate acts as a form of benchmark. The comparison to a benchmark provides some insightful context. If we’re earning less than inflation, then we know that over time the real purchasing power of the bank deposit is going to erode. We may have more money in the future. But it will buy less.

When it comes to a portfolio of securities, whether stocks, bonds or both, it would be handy to have a similarly informative benchmark – one that could help answer questions like the following. Did my portfolio have great returns simply because the market was soaring? How risky is my portfolio? Are the wild fluctuations in my portfolio’s returns warranted, given the way it’s invested? Am I getting my money’s worth for what I’m paying my investment manager? Moreover, we should be able to vary our time frame, and consider these questions over short, medium and longer time periods.

Enter the “market index benchmark”.

A market index benchmark is the hypothetical return from investing in one or a combination of market indices, such as the S&P/TSX Composite or S&P 500. It represents the return an investor could theoretically earn if they owned a portfolio comprised of, for example, all the stocks in the S&P/TSX Composite index – in precisely the same proportions as in the index, and incurred no expenses to acquire, hold or dispose of those stocks.

There are a few simple properties we should want a benchmark to have.

• First, it ought to be reasonably representative of the portfolio with which it is being compared. For example, comparing the returns on a 91-day Treasury Bill index to a portfolio of U.S. stocks is not especially informative.
• Second, a benchmark ought to be comprised of readily available, specific and recognizable underlying market indices. Even if it had significant explanatory power, an index of the 24-month moving average of the hemline lengths in 4-star Manhattan hotel cocktail bars every second Friday night would hardly qualify.
• Third, a benchmark ought to be consistent from period to period so that comparisons between it and the portfolio are not obscured by repeated changes in the benchmark. It’s difficult to plot the growth in a child’s height over 15 years if you keep changing the measuring stick every 18 months or so.

For the Nexus North American Equity Fund, we use a market index benchmark that is a blend of three underlying indices in fixed proportions:  the 91-day Government of Canada Treasury Bill Index (5%), the S&P/TSX Composite Total Return Index (50%), and the S&P 500 Total Return Index measured in Canadian dollars (45%). While these proportions – 5-50-45 – don’t match the underlying investments of the Fund perfectly, they accurately reflect the long-run allocation pattern to those three major asset classes. The Fund generally holds cash and short-term money market securities equivalent to between 0% and 10% of its portfolio, although the position has run up above 20% in the past. Canadian equities have nearly always accounted for between one- and two-thirds of the Fund’s portfolio. And U.S. stocks typically represent between 15% and 50% of the Fund’s investments. In addition to these, the Fund currently holds between 8% and 12% of the portfolio in non-North American equities.

All the component indices in the benchmark are widely published and well recognized. Moreover, each is itself “investable” in the sense that it is completely feasible to find index funds or exchange-traded funds designed to mimic each of these indices, and invest in them in the same proportions as represented in the benchmark.

Finally, the benchmark has remained the same since inception of the Fund 22 years ago. The actual asset mix of the Fund has fluctuated widely over that period, almost never matching the 5-50-45 proportions that each of the three underlying market index represents in the benchmark. And the benchmark contains no representation for international (i.e., non-North American) equities despite the Fund having long had between 5% to 10% of its assets invested there. But having maintained the same benchmark throughout, the Fund’s track record can be evaluated over any time period. It’s possible to assess whether our straying from the benchmark weightings has added to or detracted from the Fund’s performance, whether the Fund’s returns are more or less volatile (i.e., variable) than the markets it is primarily invested in, and whether the trade-off between Fund’s returns and risk (as represented by the volatility of returns) has been beneficial for the Fund’s unitholders.

There is an argument for adjusting the benchmark when there is a profound change in the Fund’s asset mix or strategy. In mid-2015, when Canadian equities fell to between 33% and 36% of the Fund’s assets – far below the 50% weighting in the benchmark – a critic might have declared the benchmark no longer representative and therefore inappropriate for comparison. But there had been no permanent shift in the Fund’s strategy to warrant a change, and, indeed, the dip in the Canadian equity allocation proved temporary. Now it’s back to between 42% and 45%, and comfortably larger than the Fund’s 35% to 40% position in U.S. stocks.

This tension about the representativeness of a benchmark, especially when the composition of a portfolio drifts or differs meaningfully from that of its benchmark, highlights an important issue. The simple comparison of portfolio-level returns with the overall returns of a benchmark merely scratches the surface, analytically speaking. A fuller understanding of why a portfolio has fared differently than its benchmark involves a much deeper, multi-layered analysis.

For example, perhaps the portfolio had a much smaller investment in Information Technology stocks than the benchmark during a period when that sector had a big influence on the benchmark’s performance. Or maybe the portfolio’s overall investment in such stocks was the same as the benchmark’s, but the portfolio manager managed to avoid the 3 worst performing ones. These are worthwhile things to identify and understand. But they may not be readily visible from the surface.

That a more nuanced and detailed analysis is required to tease out these types of insights is not a shortcoming of an imperfect benchmark. Rather, it illustrates that the usefulness of an appropriate benchmark goes far beyond simply how closely it “matches” the portfolio it is compared with, even if that additional utility comes only after applying more analytical effort and skill.