Are All Market Indexes Created Equal? (Digest Summary)

A wide array of investment professionals use benchmarks to assess the success of
their investment process or asset optimization analysis. Such criteria as
objectivity, available data, and market coverage, among others, are used in the
selection process. Surprisingly, costs are usually not taken into account. Given
that costs have risen over the past few years and are not clearly understood by
investors, there is a need for more transparency. In this article, the authors focus
on US and global equity markets, noting that future research could include different
geographies and different asset classes.

Benchmarks are commonly used in the investment world to represent a target market,
for marketing purposes, or as part of an optimization process. Nevertheless, access
restrictions may apply, including licenses, the sharing of data, and time lags. As
the number of indexes has grown over time, acquiring the correct data has become
more complex. To increase the transparency of costs, the Spaulding Group has
developed guidelines that include a benchmark cost analysis.

The authors show that capitalization, sector exposure, and security selection are the
main drivers of performance. Nevertheless, the indexes turn out to be very similar
owing to capitalization weighting; the largest stocks are given the highest weights.
Although there can be differences in how vendors identify the country of
multinationals, these effects are offset by capitalization weighting. The same is
also true for the number of securities in a benchmark because the smaller companies
have lower weights. Using capitalization also results in similar sector weights
between indexes. The authors observe that the benchmarks have similar sector
weightings because of similar weighting and selection methodologies.

The authors advocate an increase in cost transparency. Their research could affect
the work of benchmark vendors: Although a current cost might be hidden, an increase
in transparency could force vendors to provide a breakdown of their services.

The authors consider vendors of six US equity indexes and four global equity indexes
for their analysis. When certain vendors offer several indexes, the most common
index is used. Furthermore, no indexes composed entirely of small-cap stocks are
included. Lastly, only developed and emerging markets—no frontier
markets—are considered. The authors end up with 10 years (2005–2014) of
monthly data for global indexes and 7 years (2008–2014) of monthly data for US
(domestic) indexes.

After finishing the index selection process, the authors apply three criteria to
assess the differences between the indexes. They use an analysis of variance to
assess the difference in monthly returns. But they deem this test inappropriate
because the monthly return does not answer the question whether an index represents
a certain market. Thereafter, the authors use a correlation matrix because this
method is also used for asset allocation and monitoring. The authors also use a
principal component analysis (PCA), which is a good complement to the correlation
matrix because it shows the relative influence of underlying factors.

Both the correlation matrix and the PCA show that the global and US indexes are
strongly correlated. The PCA also shows that the first component explains almost all
the variance, and the correlation matrix shows correlations close to 1 for all
indexes.

Benchmarking is used by many investment professionals, either as a standard to beat
or to test whether a passive portfolio moves in line with the overall market.
the spotlight very often, so this research is most welcome. The authors set up a
clear method whereby benchmarks of different vendors are compared, using a
correlation matrix and a PCA as test methods. Their results clearly show little
difference between the benchmark returns. Although the cost differences have not
been specifically highlighted in this research, the authors make a strong case for
increased transparency.