Tracking error, when referring to an ETF, is the difference between the ETF’s return and the return of the underlying index it is tracking. The ETF is usually expected to lag its benchmark index by at least its management expense ratio (MER), and sometimes more.
If we compare the 10-year annualized return of the iShares Core S&P/TSX Capped Composite Index ETF (XIC) to the returns of the S&P/TSX Capped Composite Index (its current benchmark), we end up with a very respectable annualized tracking error of -0.04% (7.93% minus 7.97%). This low tracking error appears too good to be true, since the fund paid management expenses of between 0.17% and 0.27% each year.
XIC vs. the S&P/TSX Capped Composite Index as of December 31, 2013:
Source: BlackRock Canada
The main reason for this difference is BlackRock’s choice of benchmark. Throughout the measurement period, XIC is compared to the S&P/TSX Capped Composite Index, even though it historically tracked the S&P/TSX 60 Capped Index prior to November 15, 2005. This is disclosed on their website, but can easily be missed.
BlackRock Canada Performance Disclaimer:
Source: BlackRock Canada
In order to create a more apples-to-apples benchmark for XIC, the returns of the S&P/TSX 60 Capped Index should be spliced together with the returns of the S&P/TSX Capped Composite Index. After doing this, the annualized tracking error has now increased to -0.20% (7.93% minus 8.13%). This seems much more in line with what an investor would expect to see.
XIC vs. the Spliced Canada Index as of December 31, 2013
Annualized Return (%) | 1y | 3y | 5y | 10y |
iShares Core S&P/TSX Capped Composite Index ETF (XIC) | 12.71 | 3.14 | 11.58 | 7.93 |
Benchmark* | 12.99 | 3.4 | 11.92 | 8.13 |
Tracking Error (%) | -0.28 | -0.26 | -0.34 | -0.2 |
Sources: Bloomberg, BlackRock Canada, S&P Dow Jones Indices courtesy of Dimensional Returns 2.0
*Spliced Canada Index. An index that reflects performance of the S&P/TSX 60 Capped Index through November 14, 2005; S&P/TSX Capped Composite Index thereafter.
Investors and advisors must be diligent when reviewing benchmark comparisons that have been illustrated by fund providers. In some cases the numbers may not be telling the entire story, and further analysis may be required.
You mentioned in another post:
Perhaps you could do a more in-depth post about benchmarking? :)
What is an acceptable tracking error? How would one benchmark their own portfolio? What would one compare it to?
@skube: I’ve provided examples in my rate of return white paper on why the money-weighted rate of return is not useful for benchmarking purposes: https://www.canadianportfoliomanagerblog.com/wp-content/uploads/2014/09/2015-07-10_PWL_Bender-Bortolotti_Understanding-your-portfolio-s-rate-of-return_Hyperlinked.pdf?850eac
If you’re following my model ETF portfolios diligently, you should generally be within 0.50% of the posted model portfolio returns (unless you have additional fees or a different portfolio composition).