Measuring and reducing the risk of loss is always a concern
of equity investors, but it has become virtually an obsession
for many since the 200809 market drop. The key to
reducing risk is appropriate diversification (investing in a
basket of stocks rather than a mere handful) and getting
position sizes right.
For the purpose of this article, well look at the
impact of various security-weighting schemes on the return of
one well-known and well-diversified basket of stocks, the
S&P 500, over the past 16 years. The relationships, we
find, have positive implications for actively managed as well
as indexed portfolios.
Like many capitalization-weighted indices, the S&P 500
tends to be dominated by the largest-cap stocks. Sometimes,
though, these stocks become the largest-cap because of
excessive increases in their valuations. As a result, investing
in these indices may lead to buying high and selling low: the
opposite of every investors goal.
To avoid this risk of cap-weighted schemes, some investors
have turned to equal-weighting their portfolios, but there are
risks associated with that strategy, too. Because smaller-cap
stocks tend to be hit harder in down markets and fare better in
rising markets, equal-weighted portfolios and indices tend to
be more volatile than their cap-weighted counterparts.
We looked for a portfolio-weighting approach that could
offer the best of both worlds, while mitigating downside risk
which most investors fear more than volatility.
Volatility (the standard deviation of returns) is
symmetrical: It captures upward and downward moves equally. But
most investors typically hate losing money more than they love
winning. That is, they care more about market drawdowns than
market rallies. In other words, investors want upside, but not
A variety of risk metrics focus on downside risk; for
example, the Sortino ratio measures the return achieved per
unit of downside price movement.
We used the Sortino ratio and other drawdown-sensitive risk
metrics as guides to create a capital-allocation scheme based
on expected tail loss, or ETL. A stocks ETL measures the
potential losses in the worst scenarios were likely to
see the rare and infamous black swans as well as
in down markets generally. In the ETL-weighted portfolio, we
size positions so that each stocks ETL contribution to
the overall portfolio losses would be exactly the same.
Our research results are encouraging; drawdowns for the
ETL-weighted portfolios were generally smaller than for the
other weighting schemes. In the display below we highlight the
two worst drawdowns during the 19962012 period that we
studied, as well as the average drawdown.
In the worst drawdown, which occurred during the Great
Recession, the ETL-weighted S&P 500 portfolio lost
substantially less than the equal-weighted alternative, and did
modestly better than the cap-weighted portfolio even
though investors who stayed the course in equities clung to
large-cap companies. In the second-worst drawdown, when tech
imploded, damaging large-cap names more, ETL beat the
cap-weighted alternative by a full 20 percent, and outperformed
the equal-weighted construction as well.
Fueled by results like these, the average loss for the
ETL-weighted portfolio was smaller than the average drawdown in
both a cap-weighted and an equal-weighted portfolio over the
16-year time horizon.
The ETL-weighted portfolio also provided almost all of the
additional returns of the equal-weighted index, with
substantially lower volatility. Further, volatility was
marginally lower than that of the cap-weighted index as
Given this array of results, the ETL-weighted index had a
higher Sharpe ratio (which measures return against both
positive and negative volatility) than either its cap- or
equal-weighted counterpart, as well as a higher Sortino ratio.
This last result gives us reason to be optimistic that
ETL-weighted portfolios can soften the blow during market
No risk strategy can reliably obliterate losses in down
markets, but in our view, capital allocation using ETL has the
potential to make such losses more tolerable.
In my next post, I will look at the use of expected tail
loss in mitigating risk in multi-asset portfolios. Weve
found that the benefit there is even larger than what we see
for stock-only portfolios.
The views expressed herein do not constitute research,
investment advice or trade recommendations and do not
necessarily represent the views of all AllianceBernstein
Andrew Y. Chin is Global Head of Quantitative Research
and Investment Risk at AllianceBernstein.