Morningstar Magazine - April/May 2018 - 40

Strategies

EXHIBIT 3

The process is repeated for the next
nonoverlapping five-year period. The variations
are averaged over time for stocks (50 years
from January 1961 to December 2010) and equity
funds (35 years from January 1981 to
December 2015).
shows the average distribution of
ranking variation for CVaR and MDD ratios for the
equity funds universe. The distribution for
both ratios is similar for the stock universe and,
thus, not plotted here. There are seven
nonoverlapping five-year periods (from January
1981 to December 2015) in which we can compute
the distribution of ranking variations, and the
seven histograms are then averaged over time and
plotted in the exhibit. The distributions are
approximately symmetrical, which indicates
that the number of funds experiencing a
downgrade is similar to the number of funds
experiencing an upgrade at different variation
levels. The ranking variation can go as high
as plus or minus 50 percentage points with a 0.2%
probability for the MDD ratio.

Ranking Variations The distribution of ranking variations between the
standard Sharpe ratio and CVaR and MDD ratios shows that most funds have
the same ranking for the ratios, but a significant amount do not.
The Distribution of Ranking Variations for the CVaR Ratio and MDD Ratio for the Equity Funds

MDD Ratio

CVaR Ratio

Frequency (Number of Funds)
80

EXHIBIT 3

Impact of Volatility on Ranking Variation

How much does the average ranking change
for a U.S. stock or U.S. equity fund with
low volatility if one switches from the standard
Sharpe ratio to one of the tail-risk-based Sharpe
ratios? This is an important practical question
because many assets are managed with
an emphasis on targeting low overall volatility.
We expect that this is where ranking downgrades
will be most pronounced.
Using the same five-year periods above, we sort
our individual stock universe and U.S. equity fund
universe into volatility quintiles and then
compute the average ranking variations for each
quintile. Our hypothesis is that the lowest-volatility
quintile (in each of the two universes) should
experience a downgrade and the highest-volatility
quintile should experience an upgrade,
because low volatility tends to exhibit negative
skewness and vice versa.
shows the sorting results. Indeed,
stocks and funds in the lowest-volatility quintile on
average experience a downgrade by about
2 to 3 percentage points for both CVaR ratios and
EXHIBIT 4

40

Morningstar April/May 2018

60

Note: The number of funds is scaled to
100 in each ﬁve-year period.

40

20

0
-40
-60%
Ranking Variations

-20

+20

0

+40

+60

Source: Morningstar. Data from January 1981 through December 2015.

EXHIBIT 4

Volatility Impact The lower the volatility quintile, the higher the downgrade
levels for stocks and funds, meaning on average they experienced more negative
tail risk than their Sharpe ratios would imply.
Ranking Variations for Volatility
Quintiles for U.S. Stocks

Lowest-Volatility Quintile

CVaR ratio ranking variation (%)

-1.55

-0.95

-0.38

0.36

2.52

(-4.14)

(-6.45)

(-3.16)

(2.40)

(9.85)

-2.35

-0.91

-0.52

0.40

3.38

(-3.71)

(-2.82)

(-2.74)

(1.94)

(5.34)

t-statistic
MDD ratio ranking variation (%)
t-statistic

Highest-Volatility Quintile

Ranking Variations for Volatility
Quintiles for U.S. Equity Funds

Lowest-Volatility Quintile

CVaR ratio ranking variation (%)

-1.65

-0.67

-0.46

0.17

2.61

(-1.46)

(-1.98)

(-1.09)

(0.35)

(2.32)

-3.32

-0.70

0.25

1.35

2.43

(-2.65)

(-1.10)

(0.22)

(2.56)

(1.64)

t-statistic
MDD ratio ranking variation (%)
t-statistic

Highest-Volatility Quintile

Source: Morningstar. Data from January 1981 through December 2015 for U.S. stock funds, January 1961 through December 2010 for U.S. stocks.
Table of Contents for the Digital Edition of Morningstar Magazine - April/May 2018
