Factor momentum
with Robert Arnott and Vitali Kalesnik
2022, Review of Financial Studies, forthcoming
Factors display strong cross-sectional momentum. This effect, which is distinct from time-series factor momentum, transmits into portfolio returns through variation in portfolios' factor loadings. All characteristic momentum strategies, including industry momentum, significantly correlate with each other, consistent with the profits emanating from the same source; and consistent with the factor-loadings transmission mechanism, none is profitable net of factor momentum. Factor momentum concentrates in the first few highest-eigenvalue factors. Individual stocks display short-term reversals because stocks' idiosyncratic and systematic components are negatively cross-serially correlated. Strategies that isolate and trade momentum in stocks' systematic components are profitable.
Factor momentum and the momentum factor
with Sina Ehsani
2022, Journal of Finance 77(3), 1877–1919
Momentum in individual stock returns emanates from momentum in factor returns. Most factors are positively autocorrelated: the average factor earns a monthly return of 6 basis points following a year of losses and 51 basis points following a positive year. We find that factor momentum concentrates in factors that explain more of the cross section of returns and that it is not incidental to individual stock momentum: momentum-neutral factors display more momentum and momentum in firm-specific residuals appears to capture momentum in omitted factors. Our key result is that momentum is not a distinct risk factor; it times other factors.
Media: Featured in Wall Street Journal ("
A New Way to Think About Momentum Investing”, May 5, 2019)
Award: Q-Group’s Jack Treynor 2019 Prize Winner
Long-term discount rates do not vary across firms
with Matti Keloharju and Peter Nyberg
2021, Journal of Financial Economics 141(3), 946–967
Long-term expected returns do not appear to vary in the cross section of stocks. We show that even negligible persistent differences in expected returns, if they existed, would be easy to detect. Markers of such differences, however, are absent from actual stock returns. Our results are consistent with behavioral models and production-based asset pricing models in which firms’ risks change over time. Consistent with the lack of long-term differences in expected returns, persistent differences in firm characteristics do not predict the cross section of stock returns. Our results imply stock market anomalies have only a limited effect on firm valuations.
Asset managers:
Institutional performance and factor exposures
with Joseph Gerakos and Adair Morse
2021, Journal of Finance 76(4), 2035– 2075
Using data on $18 trillion of assets under management, we show that actively managed institutional accounts outperformed strategy benchmarks by 88 (44) basis points on a gross (net) basis during the period 2000–2012. Estimates from a Sharpe (1992) model imply that asset managers’ outperformance came from factor exposures. If institutions had instead implemented mean-variance efficient portfolios using index and institutional mutual funds available during the sample period, they would not have earned higher Sharpe ratios. Our results are consistent with the average asset manager having skill, managers competing for institutional capital, and institutions engaging in costly search to identify skilled managers.
Are return seasonalities due to risk or mispricing? Evidence from seasonal reversals
with Matti Keloharju and Peter Nyberg
2021, Journal of Financial Economics 139(1), 138–161
Stocks tend to earn high or low returns relative to other stocks every year in the same month (Heston and Sadka 2008). We show these seasonalities are balanced out by seasonal reversals: a stock that has a high expected return relative to other stocks in one month has a low expected return relative to other stocks in the other months. The seasonalities and seasonal reversals add up to zero over the calendar year, which is consistent with seasonalities being driven by temporary mispricing. Seasonal reversals are economically large, statistically highly significant, and they resemble, but are distinct from, long-term reversals.
The misguided beliefs of
financial advisors
with Brian Melzer and Alessandro
Previtero
2021, Journal of Finance 76(2), 587–621
A common view of retail finance is that conflicts of interest contribute to the high cost of advice. Within a large sample of Canadian financial advisors and their clients, however, we show that advisors typically invest personally just as they advise their clients. Advisors trade frequently, chase returns, prefer expensive, actively managed funds, and underdiversify. Advisors' net returns of -3% per year are similar to their clients' net returns. Advisors do not strategically hold expensive portfolios only to convince clients to do the same; they continue to do so after they leave the industry.
Media: Featured in MarketWatch ("
Here’s why your investment adviser isn’t making you enough money”, February 26, 2018), MoneySense ("
Advisors believe their own misguided advice”, April 22, 2018), and Tuck Research & Insights ("
When It Comes to Household Finances, Are You Making the Right Decisions?”, April 29, 2019)
Earnings, retained earnings,
and book-to-market in the cross section of expected
returns
with Ray Ball, Joseph Gerakos, and
Valeri Nikolaev
2018, Journal of Financial Economics 135(1), 231-254
Retained earnings-to-market subsumes book-to-market's
power to predict the cross section of stock returns in
pre- and post-Compustat U.S. data as well as in
international data. Retained earnings-to-market's predictive power
stems entirely from accumulated earnings. Our thesis is
that retained earnings-to-market---and, by extension,
book-to-market---predicts returns because it is a good
proxy for earnings yield (Ball,1978; Berk, 1995).
Media: Featured in Chicago Booth Review ("
Why this value-investing ‘buy’ signal is out of date”, April 25, 2017) and Wall Street Journal ("
Why the Traditional Way of Measuring ‘Value’ Stocks May Be History”, September 9, 2018)
The history of the cross
section of stock returns
with Michael Roberts
2018, Review of Financial Studies 31(7), 2606-2649
Using data spanning the 20th century, we show that the
majority of accounting-based return anomalies, including
investment, are most likely an artifact of data snooping.
When examined out-of-sample by moving either backward or
forward in time, most anomalies' average returns and
Sharpe ratios decrease, while their volatilities and
correlations with other anomalies increase. The
data-snooping problem is so severe that even the true
asset pricing model is expected to be rejected when tested
using in-sample data. The few anomalies that do persist
out-of-sample correlate with the shift from investment in
physical capital to intangible capital, and the increasing
reliance on debt financing observed over the 20th century.
Our results emphasize the importance of validating asset
pricing models out-of-sample, question the extent to which
investors learn of mispricing from academic research, and
highlight the linkages between anomalies and economic
fundamentals.
Award: Marshall E. Blume Prize, First Prize for
Best Paper Published by the Rodney L White Center, 2017 (
announcement)
Decomposing value
with Joseph Gerakos
2018, Review of Financial Studies 31(5), 1825-1854
Firms move between growth and value because of either
changes in size or the book value of equity. We show that
the value premium is specific to the variation in
book-to-market ratios that emanates from changes in firm
size. A factor based on this variation earns the entire
value premium; one based on the remaining variation earns
no premium. Hence, not all high book-to-market firms earn
value premium, and some low book-to-market firms earn
value-like returns. This disconnect between book-to-market
and the value premium provides testable restrictions for
theories of the value premium. Many models price
portfolios sorted by size and book-to-market. None
distinguish firms that earn the value premium from those
that have a high book-to-market but do not earn the
premium.
Award: Second prize in the academic
competition at the Chicago Quantitative Alliance (CQA)
Fall 2012 Conference.
Accruals, cash flows, and
operating profitability in the cross section of stock
returns
with Ray Ball, Joseph Gerakos, and
Valeri Nikolaev
2016, Journal of Financial Economics 121(1), 28–45
A cash-based operating profitability measure (that
excludes accruals) outperforms other measures of
profitability and subsumes accruals in predicting the
cross section of average returns. Firms with high
accruals earn low average returns because they are less
profitable on a cash basis.
Note: Fama and French (2015) compare
our cash-based operating profitability factor against
two alternative profitability factors, and find that the
cash factor improves the description of average returns
for many left hand side sorts.
Award: First prize in the academic
competition at the Chicago Quantitative Alliance (CQA)
Fall 2015 Conference.
Retail financial advice:
Does one size fit all?
with Stephen Foerster, Brian Melzer,
and Alessandro Previtero
2017, Journal of Finance 72(4), 1441-1482
Advisor fixed effects explain considerably more variation
in portfolio risk and home bias than a broad set of
investor attributes that includes risk tolerance, stage in
the lifecycle and financial sophistication. An advisor's
own asset allocation strongly predicts the allocations
chosen on clients' behalf.
Awards: 2015 CFA Society & Hillsdale
Canadian Investment Research Award (
announcement), 2018 Amundi Smith Breeden Prize (
list of prize winners)
Media: Featured in Wall Street Journal ("
Client
Portfolios May Match Advisers’ Own Asset Allocation",
December 12, 2014), Globe and Mail ("
Make
portfolio-building a priority to justify investment
adviser fees," December 5, 2014; "
Putting
a number on the value of financial advice: 3%," June
14, 2015), Fiscal Times ("
Expensive,
one-size-fits-all advice," December 10, 2014), Booth
Capital Ideas ("
Why
financial advice isn't worth the fees," February 25,
2015), and Kellogg Insight ("
What
Good is a Financial Advisor?," November 2, 2016)
Reading the tea leaves:
Model uncertainty, robust forecasts, and the
autocorrelation of analysts' forecast errors
with Walter Torous and James Yae
2016, Journal of Financial Economics 122(1), 42–64
Analysts optimally underreact to new information if they
try to provide forecasts that are robust to model
misspecification. We estimate that analysts' concerns for
model misspecification explain approximately 60% of the
autocorrelation in analysts' forecast errors. Our model of
robust forecasting applies not only to analysts' forecasts
but to all model-based forecasts.
Return seasonalities
with Matti Keloharju and Peter Nyberg
2016, Journal of Finance 71(4), 1557–1590
We document return seasonalities in individual stock
returns, portfolio returns, anomalies, commodities,
international stock market indices, and at the daily
frequency. These return seasonalities overwhelm
unconditional differences in expected returns.
Note: To replicate the results on
seasonalities in daily returns, you need to account for
market closures due to U.S. holidays. As the lag
k
grows, the likelihood that the days go out of sync
increases; a regression at lag
k=200, for example,
is very unlikely a Monday-on-Monday (or
Tuesday-on-Tuesday, and so forth) regression. You should
"pad" the data with missing values so that there is an
observation for every stock-day even when the market is
closed.
Award: AQR Insight Award Finalist 2015 [
Announcement]
Deflating profitability
with Ray Ball, Joseph Gerakos, and
Valeri Nikolaev
2015, Journal of Financial Economics 117(2), 225–248 (lead article)
An alternative measure of firm profits, operating profits,
exhibits a far stronger link with expected returns than
either net income or gross profit.
Note: The XSGA variable in Compustat adds
to the SG&A reported by the company other items such
as R&D expenses. This issue is discussed on p. 254 of
Volume 5 of Compustat Manuals. You can recover the
reported SG&A by subtracting XRD from XSGA.
Media: Featured in Forbes ("
The
Profitability Factor Redux: Super-Duel in Space,"
June 2, 2014)
Market reactions to
tangible and intangible information revisited
with Joseph Gerakos
2016, Critical Finance Review 5, 135–163
A decomposition of book-to-market ratio into stock and
book returns creates a book return polluted by past
book-to-market ratios, stock returns, net issuances, and
dividends. Our results cast doubt on the argument that
book-to-market forecasts returns because it is a good
proxy for the intangible return.
Reverse survivorship bias
2013, Journal of Finance 68(3), 789–813
(lead article)
The distribution of estimated alphas is biased downwards
if funds tend to disappear following poor performance.
This paper estimates a structural model to correct for
this "reverse survivorship bias."
Media: Mutual fund research featured in
Time magazine ("
The Triumph
of Index Funds," September 18, 2014) and "
The
Big Question: Are successful active managers lucky or
skilled?" (Capital Ideas, August 2014).
Do investors buy what they
know? Product market choices and investment decisions
with Matti Keloharju and Samuli Knüpfer
2012, Review of Financial Studies 25(10), 2921–2958 (lead
article)
Individuals’ product market choices influence their
investment decisions.
Lack of anonymity and the
inference from order flow
with Gideon Saar
2012, Review of Financial Studies 25(5), 1414–1456
We demonstrate that broker identity is a powerful signal
about the identity of investors who initiate trades, and
that the broker ID signal is important enough to affect
prices.
IQ, trading behavior, and
performance
with Mark Grinblatt and Matti Keloharju
2012, Journal of Financial Economics 104(2),
339–362
Reprinted in Household Finance, M. Haliassos (ed.), The
International Library of Critical Writings in Economics
series, Edward Elgar Publishing (2015)
High-IQ investors are less subject to the disposition
effect and more aggressive about tax-loss trading, and
they exhibit superior market timing, stock-picking skill,
and trade execution.
Award: Runner-up for Goldman Sachs
International - Best Conference Paper Award at the 2010
European Finance Association Conference.
IQ and stock market
participation
with Mark Grinblatt and Matti Keloharju
2011, Journal of Finance 66(6), 2121–2164
Reprinted in Household Finance, M. Haliassos (ed.), The
International Library of Critical Writings in Economics
series, Edward Elgar Publishing (2015)
Stock market participation is monotonically related to IQ,
controlling for wealth, income, age, and other demographic
and occupational information.
Media: Featured in Bloomberg Businessweek
("
Smart
Money Owns More Equities Says IQ Study of Who Buys
Stocks," January 19, 2012) and New York Times ("
What
High-I.Q. Investors Do Differently," February 26,
2012)
Why do (some) households
trade so much?
2011, Review of Financial Studies
24(5), 1630–1666
When agents can learn about their abilities as active
investors, they rationally "trade to learn" even if they
expect to lose from active investing. This
learning-about-type mechanism may explain why some
households begin experimenting with day trading, lose
money, and then quit.
Jensen's inequality,
parameter uncertainty, and multi-period investment
with Mark Grinblatt
2011, Review of Asset Pricing Studies 1(1), 1–34 (lead
article)
The proper application of Jensen’s inequality to the
multi-period investment decision turns finance intuition
on its head: multi-period investments with negative risk
premia can be profitable, risk-averse investors can have
infinite demand for risky securities, settings exist in
which risk-averse investors should not diversify, and
demand for mutual funds with negative alphas may be
rational.
Do limit orders alter
inferences about investor performance and behavior?
2010, Journal of Finance 65(4),
1473–1506
Use of limit orders drives a wedge between investors'
intentions and realized trades. Limit orders are
contrarian; they are more likely to execute when there are
news or asymmetric information; and they lose money when
new information arrives to the market. When we try to
infer investors' information sets or intentions from their
realized trades, our inferences are biased. I call this
bias the "limit order effect."
Media: Featured in the Chicago
Booth
Capital
Ideas (October 2007) and the Economist
Intelligence Unit.