Download Value Investing

yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Document related concepts

2010 Flash Crash wikipedia, lookup

1 |
ETFs, ETPs & Indexing
The Voices of Influence |
Value Investing: Smart Beta
versus Style Indexes
Jason Hsu
Jason Hsu
is co-founding partner
and vice chairman at
Research Affiliates,
LLC in Newport Beach,
CA, and is a professor
in finance at UCLA
Anderson School of Management in Los Angeles,
[email protected]
Summer 2014
ost commercially available
smart beta1 indexes contain
meaning ful exposure to
the value premium. In fact,
almost all non-price-weighted index schemes
will naturally access the value anomaly,
which can be interpreted as the effect of
contrarian rebalancing against long-horizon
mean reversion in equity prices (Arnott et al.
[2013] and Chow et al. [2011]). Because
prices are prone to reverse direction in the
entire cross-section of stocks, the contratrading inherent in rebalancing—buying
what has fallen and selling what has risen
in price—can be applied to all stocks in the
equity universe. This approach to extracting
the value premium is quite different from
traditional investing, which is predicated,
at base, on buying stocks with low priceto-book ratios (P/B) at their capitalization
The first section of this article describes
how conventional, value style index construction leads to industry concentrations
and explains how price-based weighting
adversely affects returns. The second, short
section introduces mean reversion in stock
prices and presents rebalancing as the mechanism through which some of the smart beta
indexes implement a more diversified value
strategy. The third section evaluates evidence
that, just as stock prices tend to revert toward
their long-term averages, so also does the
value premium. The section further demonstrates that fundamentally weighted indexes
implicitly engage in dollar cost averaging to
capture the full effect of mean reversion in
both stock prices and the value premium.
The article concludes with observations on
the relative strengths of traditional and smart
beta approaches to index investing.
Value investing has been known since
the early 1920s to deliver superior performance. In the late 1980s and early 1990s,
with the growing popularity of consultants’
nine-box style matrices, index providers
sought to provide investors with transparent,
low cost “beta” exposure to value investing
through products tied to cap-weighted
indexes with predominately value characteristics. By and large, these first-generation
value indexes were constructed by selecting
stocks with low P/B and then weighting them
in proportion to their market capitalization.
Over time, the methodologies evolved to
include other measures of value, as opposed
to growth, and to situate stocks on a value–
growth continuum.
According to financial lore, the value
style outperforms the market in the long run.
We observe, however, that the major value
style indexes have had mixed results relative
The Journal of Index Investing to broad cap-weighted market indexes over the past 30
years (Exhibit 1). The hypothesis that value investing
reliably leads to favorable long-term results may be
wrong, or—among other possible explanations—the
traditional value style strategies may not capture the full
value premium.
The structure of the conventional value style indexes
also raises some questions. They have two important
characteristics. First, their active shares relative to the
market benchmark are dominated by industry bets. In
traditional value indexes, growth-oriented industries are
represented only to the extent value stocks have growth
characteristics. The resulting indexes are unrepresentative of the underlying economic exposure, because they
are poorly diversified from an industry-concentration
perspective. In particular, a quick examination of traditional value indexes will find large overweights in
financial and energy stocks, made possible in part by
underweight positions in technology stocks. The S&P
500 Value Index and the Russell 1000 ® Value Index
have significantly more exposure to financial and energy
stocks, and less to technology stocks, than the corresponding broad market indexes (Exhibit 2).
It is well understood today that, as value signals,
ratios such as P/B and price-to-earnings (P/E) are
more meaningful for comparing stocks within a more
homogeneous economic sector or industry (Gaudette
and Lawton [2007]). Using P/B and P/E to contrast
stocks across different sectors or industries is substantially
less useful. Therefore, the aggressive industry active
weights, resulting from favoring stocks from industries
with traditionally lower valuation ratios, are suboptimal
approaches for exploiting the value effect.
Second, although traditional value style indexes are
reconstituted once a year, the stocks that are gathered
into their value baskets are assigned index weights proportionate to their market capitalization. Consequently,
the weights of the stocks in a value index f luctuate with
prices. For example, prior to the global financial crisis,
many of the large banks became expensive relative to
their own historical valuation ratios. The cap-weighting
scheme meant that these expensive banks also took on
dominant weights within the value index just prior to
the banking sector crisis. Conversely, at the nadir of the
crisis, when banks were trading at historically low valuation multiples, their weights in the value index were
substantially reduced (Exhibit 3). Due to the effect of
stock prices on index weights, bank stocks contributed
more heavily to the value style indexes’ losses during the
crisis than they added to gains in the recovery.
It is useful to view the value investment strategy
as capturing mean reversion in stock valuation ratios
(Cohen et al. [2003]). Under this interpretation, rebal-
Value Style Indexes vs. Broad Market Indexes
Source: Research Affiliates, using data from FactSet and FRED.
JII-HSU.indd 122
5/16/14 6:57:23 PM
Index Sector Weights
*As of February 28, 2014.
Source: Research Affiliates, using data from S&P Dow Jones Indices and
Pre-Crisis and Pre-Recovery Weights
Source: Research Affiliates, using data from FactSet.
ancing against price is a critical step in exploiting
long-horizon mean eversion. Cap-weighted indexes
automatically adjust positions as constituents’ market
prices change; there is no rebalancing to exogenous
weights. Therefore, applying cap-weighting to value
stocks effectively eliminates the opportunity to profit
from mean reversion. By contrast, smart beta value
indexes offer more powerful approaches to capturing
the value premium. Let me use an index constructed in
accordance with the non-cap-weighting methodology
described in Arnott et al. [2005] as an illustration.
In this approach, stocks are weighted by fundamentals associated with economic size, such as book value
and total cash f lows. These measures are not directly
related to stock prices. Nonetheless, because they track
capitalization over time, the fundamentally weighted
strategy generally contains industry exposures that are
reasonably similar to the broad market index. As of
December 31, 2013, the fundamentally weighted index
held 21.2% in the financial sector and 11.8% in energy
stocks (compared to 26.9% and 14.7%, respectively, in
the Russell 1000 Value Index). Thus, much of the active
shares of fundamentally weighted indexes are taken up
by intra-industry bets—for instance, overweighting Ford
and underweighting Tesla. Industry-based active shares
JII-HSU.indd 123
become meaningful only if an industry precipitously
becomes substantially more or less expensive relative to
its own historical valuation level.
The fundamentally weighted strategy also rebalances annually against valuation ratio movements,
buying what has become cheaper over the course of the
year and selling what has become more expensive. The
rebalancing is effected over hundreds of stocks across
all industries. This contra-trading amplifies the weak
time-series, price mean-reversion effect through the law
of large numbers in the equity cross-section.
The long-term returns earned by smart beta
strategies are substantially higher than the traditional
value indexes’ value-added returns. The fundamentally
weighted approach to value investing results in approximately 200 bps of historical outperformance. (This outcome was first documented in Arnott et al. [2005].)
Over the 30-year period ending December 31, 2013, the
simulated return of the fundamentally weighted index
was 13.14% per annum, more than 200 bps higher than
the annualized returns of the S&P 500 Index (11.09%)
and the Russell 1000 Index (11.12%). In the same period,
the S&P 500 Value Index underperformed the S&P 500
by 18 bps and the Russell 1000 Value Index exceeded
the Russell 1000 Index by 41 bps. The outperformance
of the fundamentally weighted index is not attributable
to added risk; its hypothetical 30-year return is modestly superior to the returns of the value style indexes
on a risk-adjusted basis. The fundamentally weighted
index had a Sharpe ratio of 0.49, compared to 0.36 for
the S&P 500 Value Index and 0.41 for the Russell 1000
Value Index.
There is empirical evidence that, like the equity
risk premium, the value premium is also mean reverting.
(Asness et al. [2000] and Cohen et al. [2003]). This fact
has strategic significance.
The last two market cycles illustrate the return
impact of reversals in P/B trends. The tech boom drove
the ratio of the growth P/B to the value P/B to 14.65
in July 2000. (In other words, at that point in time
the average P/B of growth stocks was 14.65 times the
average P/B of value stocks.) In the subsequent three
years, value cumulatively outperformed growth by
60.3%. The housing and subprime mortgage bubble
drove up prices for the banking and consumer staples
5/16/14 8:40:03 PM
sectors, which are traditional value sectors. In January
2006 the growth stock P/B was 4.36 times the value
stock P/B, and value cumulatively underperformed
growth by 33.1% in the subsequent three years. The
ratio expanded again as the economy recovered from
the global financial crisis, reaching 11.5 times in March
2009; in the following three years, value outperformed
growth by 44.4% cumulatively. Exhibit 4 summarizes
these observations, and Exhibit 5 displays the ratio of
growth and value P/B at each month end with the cor-
P/B Ratios and Growth and Value Returns
Source: Research Affiliates using data from CRSP and Compustat.
Month-End P/B (1998–2013) and Subsequent Three-Year Value Premia (Annualized)
Source: Research Affiliates using data from CRSP and Compustat.
JII-HSU.indd 124
5/16/14 6:57:24 PM
Risk and Return Impact of Dollar Cost Averaging, 1963–2013
Source: Research Affiliates using data from CRSP and Compustat.
responding difference between annualized growth and
value returns for the three years then starting.
One can capture the value premium either by
investing in low P/B stocks or by rebalancing from the
last few years’ winner stocks (those whose prices have
appreciated the most) into the losers. However, when
momentum carries prices away for an extended period
of time, rebalancing can cause value stocks to underperform, perhaps substantially. The larger and more
prolonged the value underperformance, the bigger the
spread between growth and value stock P/B. The large
P/B spread is then a signal for the magnitude of the
forthcoming return reversal. There is mean reversion
in the mean-reversion effect.
When the mean-reversion effect shows evidence
of mean reversion, it makes sense to dollar cost average
contrarian bets (Brennan et al. [2005]). Let me show
this by comparing two portfolios: one that allocates a
constant tracking error to low P/B stocks, and another
that dynamically allocates more tracking error when
the gap between growth and value P/B ratios widens.
The first portfolio is akin to the traditional value strategies, which tilt toward cheap stocks in order to generate outperformance. The second portfolio is similar
to fundamentally weighted and other simpler smart beta
indexes, whose rebalancing heuristics implicitly contain
dollar cost averaging. Exhibit 6 shows the simulated
long-term results: The value portfolio with the dynamically adjusted tracking error (that is, the value portfolio
that automatically engages in dollar cost averaging)
outperforms the value portfolio whose tracking error
is held static (the traditional portfolio with a value bias)
by 49 bps with no incremental risk.
Thus the fact that the value premium is mean
reverting has strategic implications. Merely tilting
toward value stocks may leave a good part of the total
value premium on the table. By comparison, some smart
beta approaches may squeeze much more juice from the
JII-HSU.indd 125
value apple by means of a rebalancing rule that effectively carries out dollar cost averaging.
The index industry attempted to harvest the different equity risk premia long before the arrival of smart
beta indexes. However, the traditional approaches are
now understood to be relatively ineffective. The stylebased investing approach characterized by traditional
value and growth indexes neither captures the value premium effectively in individual securities nor applies the
methodology broadly enough to fully exploit the value
premium across all stocks and all industries. Through
the rebalancing mechanism, smart beta approaches
effectively sell high and buy low, profiting more efficiently from mean reversion in stock prices. In addition, the dollar cost averaging inherent in fundamentally
weighted strategies takes advantage of second-order
mean reversion (that is, mean reversion in the mean
reversion of the value premium). It bears mention, however, that the various approaches to index construction
and maintenance within the smart beta universe represent different degrees of effectiveness in harvesting factor
premia. While many of them do in fact outperform the
cap-weighted market index by reaping sources of excess
return other than bulk beta, their efficiency, as measured
by transaction costs, investability, and underlying economic exposure, can vary appreciably. The advantages
of moving away from traditional strategies such as value
style investing seem obvious. Nonetheless, investors do,
indeed, need to be smart when it comes to analyzing
smart betas.
I owe an enormous debt of gratitude to Philip Lawton
for his tireless editorial assistance. Thanks also go to Noah
Beck, Prashant Pandey, and El Winata for research support.
5/16/14 6:57:24 PM
The views and opinions expressed are those of the
author and not necessarily those of Research Affiliates, LLC.
The material contained in this document is for informational
purposes only and not intended as an offer or solicitation for
the purchase or sale of any security or financial instrument nor
is it advice or a recommendation to enter any transaction.
Towers Watson is credited with coining the phrase
“smart beta” to describe non-price-weighted indexes and
bulk beta for describing traditional cap-weighted indexes.
The term is not meant to denigrate providers or users of
the latter. Towers Watson used the word “smart” to suggest
that investors need to be smart to make the right smart beta
Almost all traditional value indexes set the stock
weights as a function of market capitalization.
Arnott, R.D., J. Hsu, and P. Moore. “Fundamental Indexation.” Financial Analysts Journal, Vol. 61, No. 2 (March–April
2005), pp. 83-99.
Arnott, R.D., J. Hsu, V. Kalesnik, and P. Tindall. “The Surprising Alpha from Malkiel’s Monkey and Upside-Down
Strategies.” The Journal of Portfolio Management, Vol. 39, No. 4
(Summer 2013), pp. 91-105.
Asness, C.S., J.A. Friedman, R.J. Krail, and J.M. Liew. “Style
Timing: Value versus Growth.” The Journal of Portfolio Management, Vol. 26, No. 3 (Spring 2000), pp. 50-60.
Brennan, M.J., F. Li, and W.N. Torous. “Dollar Cost Averaging.” Review of Finance, Vol. 9, No. 4 (2005), pp. 509-535.
Chow, T., J. Hsu, V. Kalesnik, and B. Little. “A Survey of
Alternative Equity Index Strategies.” Financial Analysts Journal,
Vol. 67, No. 5 (September–October 2011), pp. 37-57.
Cohen, R.B., C. Polk, and T. Vuolteenaho. “The Value
Spread.” Journal of Finance, Vol. 58, No. 2 (April 2003),
pp. 609-642.
Gaudette, S.C., and P. Lawton. Equity Portfolio Characteristics
in Performance Analysis. Charlottesville, VA: CFA Institute,
To order reprints of this article, please contact Dewey Palmieri
at [email protected] or 212-224-3675.
JII-HSU.indd 126
5/16/14 6:57:24 PM