Morningstar Magazine - June/July 2018 - 60

Spotlight: Packaged Investing

of a time series of monthly log-returns is just
12 times the average of the monthly logreturns. For each vintage, we calculated a 3-by-3
cross-sectional covariance matrix of these
log-compound annual returns. We then
decomposed the variance of the log-compound
annual return of excess return into the
contributions of the three components. Finally,
we expressed each contribution as a
fraction of the total variance of the log-compound
annual return of excess return just as we did
in the time-series analysis.
The results, shown in E X H I B I T 2 , are quite different
from those in E X H I B I T 1 . In the time-series
analysis, the market component dominated the
variance of fund returns. In contrast, for most
vintages, market was the least important
component in the decomposition of cross-sectional
variance. In fact, its importance was roughly
inversely proportional to equity allocation,
with lesser importance among long-dated vintages
(those for younger participants), which tend to

invest more aggressively. For example, in the
2050 vintage, the market component accounts for
only 11% of the cross-sectional variance. At
the retirement date, when equity allocations tend
to be lowest across target-date funds, market
was the most important component, at 59% of the
cross-sectional variance. We attribute the
increasing importance of the market component
to the divergent glide paths across targetdate series as they approach retirement; the
spread widens in glide paths closer to retirement.
The importance of the style component decreases
as the vintage approaches its retirement
date, from 41% in the 2050 vintage down to 21%
at the retirement date, averaging 31% across
vintages. Selection was the most consistent of the
three components, averaging 46% across
vintages. To see how the three components of
return vary across series, for each target-date fund
in each series, we averaged the difference
between the log-compound annual return of each
component and the average log-compound annual

return of all funds in the fund's vintage. EXHIBIT 3
presents the results for the 10 mutual fund
series with the most assets as of Dec. 31, 2016.
We see that there is little consistency with respect
to the components of returns in excess of
vintage averages across series. For example, some
series achieved higher returns due to more
aggressive equity allocations (such as Great-West),
while others received more of a boost from
selection decisions (American Funds).
Understanding the respective return differences
across the three components is especially
important for plan sponsors seeking a certain glide
path shape. For example, a company with a
generous defined-benefit pension plan may choose
to have a more aggressive target-date series due
to the bondlike nature of pension benefits. If
the equity markets do well, the target-date series
may outperform; however, it will not necessarily be
clear if the outperformance is attributable to
the more aggressive glide path or the additional
style and selection series of the target-date series.
Introducing 'Activeness'

EXHIBIT 2

A New Picture But when we measure the relative contribution of each return
component, the importance of the market exposure is much smaller.
Contributions to Cross-Sectional Variance for Target-Date-Fund Vintages

Market

Style

Selection
100%

80

60

40

20
0
2050

2045

2040

Source: Morningstar.

60

Morningstar June/July 2018

2035

2030

2025

2020

2015

Retirement

These results demonstrate the importance of
style and selection decisions for a targetdate series. What's also important is to understand
how these risk exposures change over time.
To quantify the changes, we introduced a metric
called "activeness" in Blanchett and Kaplan
(2018). Activeness is not concerned with absolute
differences across products or whether the
changes resulted in better or worse performance.
Rather, it focuses on how much a series
changes across the return components (market,
style, and selection), providing insight
into how a provider's approach to a strategy may
be changing.
In our analysis, we found considerable differences
in activeness scores across series. We also
noted a small, but positive, correlation (0.26)
between total activeness and the average series
expense ratio. This is consistent with our
expectation: Funds with higher expense ratios
tend to have higher levels of activeness.
Understanding Risk/Return

Although equities are an important driver
of investment performance, investors need to look
Table of Contents for the Digital Edition of Morningstar Magazine - June/July 2018
