Journal of Financial Economics ] (]]]]) ]]]–]]]

Contents lists available at ScienceDirect

Journal of Financial Economics

journal homepage: www.elsevier.com/locate/jfec

Value versus growth investing: Why do different investors

have different styles?$
Henrik Cronqvist a, Stephan Siegel b, Frank Yu a,n
a
China Europe International Business School, People's Republic of China
b
University of Washington, Michael G. Foster School of Business, United States

a r t i c l e i n f o abstract

Article history: We find that several factors explain an individual investor's style, i.e., the value versus
Received 15 December 2013 growth orientation of the investor's stock portfolio. First, we find that an investor's style
Received in revised form has a biological basis and is partially ingrained in an investor from birth. Second, we show
28 July 2014
that an investor's hedging demands as well as behavioral biases explain investment style.
Accepted 25 August 2014
Finally, an investor's style is explained by life course theory in that experiences, both
earlier and later in life, are related to investment style. Investors with adverse macro-
JEL classification: economic experiences (e.g., growing up during the Great Depression or entering the labor
D10 market during an economic recession) or who grow up in a lower socioeconomic status
D31
rearing environment have a stronger value orientation several decades later. Our research
G11
contributes a new perspective to the long-standing value and growth debate in finance.
& 2015 Elsevier B.V. All rights reserved.
Keywords:
Investment behavior
Portfolio choice
Value investing
Growth investing

1. Introduction for these investment styles.1 For more than two decades,
Morningstar has provided a Value-Growth Score to help
The concepts of value and growth investing have a long investors choose a fund with their preferred style. Fidelity,
history in financial economics. Today, some 2,050 value funds the world's largest provider of employer-sponsored retire-
and 3,200 growth funds cater to investors with preferences ment plans such as 401(k) plans, prominently features a

☆
We are thankful for comments from Stefan Nagel (the referee), seminar participants at the Ben Graham Centre's Third Symposium on Intelligent
Investing, Cheung Kong Graduate School of Business, China Europe International Business School, China International Conference in Finance, Financial
Management Association (Asia Conference), Multinational Finance Society, Ohio State University (Finance Alumni Conference), Shanghai Advanced
Institute of Finance, Tsinghua University (School of Economics and Management), University of Miami, and for discussions with Sebastien Betermier, John
Boyd, Colin Camerer, Henry Cao, Atsushi Chino, Peter Christoffersen, John Griffin, Hai Ngo, David Hirshleifer, Kewei Hou, Chris Hrdlicka, Danling Jiang,
Samuli Knüpfer, George Korniotis, Alok Kumar, Augustin Landier, Feng Li, Juhani Linnainmaa, Jun Liu, Pedro Matos, Ed Prescott, Ale Previtero, Jun Qian, Neil
Pearson, Meir Statman, Yang Sun, Mathijs van Dijk, Tan Wang, Mitch Warachka, Mike Weisbach, Scott Yonker, and Lu Zhang. We thank Standard & Poor's
Capital IQ's China and Hong Kong teams (in particular Rick Chang) for help with various data issues. We also thank Wenqian Hu, Florian Münkel, Lucas
Perin, and Nancy Yao for capable research assistance. This project has been supported by generous research funding from China Europe International
Business School (CEIBS). Statistics Sweden and the Swedish Twin Registry (STR) provided the data for this study. STR is supported by grants from the
Swedish Research Council, the Ministry of Higher Education, AstraZeneca, and the National Institute of Health (Grants AG08724, DK066134, and
CA085739). This project was pursued in part when Henrik Cronqvist was Olof Stenhammar Visiting Professor at the Swedish Institute for Financial Research
(SIFR), which he thanks for its support. Any errors or omissions are our own.
n
Corresponding author.
E-mail addresses: [email protected] (H. Cronqvist), [email protected] (S. Siegel), [email protected] (F. Yu).
1
Source is Morningstar.com.

http://dx.doi.org/10.1016/j.jfineco.2015.04.006
0304-405X/& 2015 Elsevier B.V. All rights reserved.

Please cite this article as: Cronqvist, H., et al., Value versus growth investing: Why do different investors have different
styles? Journal of Financial Economics (2015), http://dx.doi.org/10.1016/j.jfineco.2015.04.006i
2 H. Cronqvist et al. / Journal of Financial Economics ] (]]]]) ]]]–]]]

description of value and growth funds on their Learning Finally, based on life course theory, an approach to res-
Center website.2 There are best-selling books about both earch in social psychology and neuroscience,4 which has
value and growth strategies, and countless business maga- recently made its way into finance research (e.g., Kaustia
zine articles boast recommendations about the “best value and Knüpfer, 2008; Malmendier and Nagel, 2011, 2013;
funds” or the “best growth funds.” Wall Street professionals Schoar and Zuo, 2013), we hypothesize that an individual's
are educated about value and growth investing already in specific life experiences affect behavior, including the indivi-
business school, with many Master of Business Administra- dual's investment style, later in life. We consider several
tion (MBA) programs today offering, e.g., value investing potentially relevant, and plausibly exogenous, life experiences
courses. Most important, from the perspective of academic of individuals. We analyze whether experiencing an adverse
research, one of the most debated issues in the past several and significant macroeconomic event, e.g., growing up during
decades is the differential returns of investments in value the Great Depression, affects an individual's value versus
versus growth stock portfolios, that is, the value premium growth orientation. We also analyze the impressionable years
debate (e.g., De Bondt and Thaler, 1985; Fama and French, during an individual's life course, e.g., the economic conditions
1992, 1993, 1996; Lakonishok, Shleifer, and Vishny, 1994; when an individual entered the labor market for the first time.
Daniel and Titman, 1997).3 Finally, we also examine the socioeconomic status (SES) of the
Despite all this attention to value and growth investing, rearing environment in which the individual grew up.
little research has attempted to explain the determinants of The experience of Benjamin Graham and T. Rowe Price, Jr.,
an individual's investment style (e.g., Kumar, 2009a). That is, constitute a colorful illustration of some of our hypotheses.
why are some investors relatively more value oriented, while Graham is commonly dubbed the “father of value investing”
others are more growth oriented? In this paper, we argue because he preferred stocks with comparatively low valuation
that differences in investment styles across individuals, in ratios and other characteristics that subsequently came to
principle, stem from two non-mutually exclusive sources: a define value investing. Price, the founder of the large money
biological predisposition that translates into a preference for management company with his name, is often referred to as
value or growth stocks and environmental factors that dete- the “father of growth investing” because of his preference for
rmine an individual's portfolio tilt with respect to value and companies characterized by strong earnings growth, research
growth. and development (R&D) intensity, and innovative technology.
In recent years, individual characteristics of importance for Their different investment styles could very well have a
portfolio choice, e.g., having a propensity to take financial risk biological basis, but this is not possible to investigate without
or exhibiting investment biases, have been shown to be partly data on their genetic differences. Graham grew up very poor,
explained by an individual's genetic composition (e.g., Cesarini, with his father passing away unexpectedly when he was
Dawes, Johannesson, Lichtenstein, and Wallace, 2009; Barnea, young and his mother losing the family's savings in the stock
Cronqvist, and Siegel, 2010; Cesarini, Johannesson, Magnusson, market crash known as the Panic of 1907. Among his brothers,
and Wallace, 2012; Cronqvist and Siegel, 2014). As a result, we Graham was often tasked with bargain hunting at different
hypothesize that an individual's investment style has a biolo- grocery stores (e.g., Carlen, 2012). In comparison, Price had a
gical basis, i.e., a preference for value versus growth stocks privileged upbringing, his father being an medical doctor who
could partially be ingrained in an investor from birth. We served as a surgeon his entire professional career for a rapidly
begin our empirical analysis by assessing whether and to what expanding railroad company, a growth company at that time.
extent variation in investment style across individual investors We hypothesize that such differences in life experiences can
correspond to genetic variation across these investors. contribute to differences in investment styles later in life.
We then examine which individual characteristics explain Our research contributes a new perspective to the long-
investment style and relate the evidence to portfolio choice standing value versus growth debate in finance. First, an inv-
and asset pricing models that account for the value premium estor's style has a biological basis. A preference for value ver-
(e.g., Fama, 1996; Larsen and Munk, 2012; Gârleanu, Kogan, sus growth stocks is partially ingrained in an investor already
Panageas, 2012), as in the empirical investigation of household from birth. We estimate that genetic differences across indi-
portfolio choices in Betermier, Calvet, and Sodini (2014). In viduals explain about 26% of the variation in value versus
rational models, differences in portfolio holdings are generally growth orientation, if using price-to-earnings (P/E) ratios as an
determined by investors' hedging demands. Behavioral mod- investment style measure, and about 27% if using Morning-
els of the value premium, meanwhile, suggest that the value star's Value-Growth Score. Second, we examine which indivi-
premium arises due to overreaction or excessive extrapolation dual characteristics explain investment style. Concurring with
of past performance or due to non-standard preferences. prior household finance evidence supporting risk-based the-
ories of the value premium (e.g., Betermier, Calvet, Sodini,
2014), we find that investors' hedging demands related to
human capital and displacement risk as well as behavioral
2 biases in form of a preference for speculative assets contribute
See https://www.fidelity.com/learning-center/mutual-funds/grow-
th-vs-value-investing. to investment style. Finally, an investor's style is explained by
3
The value premium debate has not been limited to only the US life course theory in that experiences, both earlier and later in
stock market. It extends to several international stock markets (e.g., Chan, life, are related to investment style. In particular, investors
Hamao, and Lakonishok, 1991; Fama and French, 1998; Daniel, Titman,
and Wei, 2001) and also to other asset classes (e.g., Asness, Moskowitz,
and Pedersen, 2013). We refer to Fama and French (2012) for recent
4
empirical evidence on the prevalence of a value premium in international For further details and references related to life course theory, see,
stock markets. e.g., Giele and Elder (1998) and Elder, Johnson, and Crosnoe (2003).

Please cite this article as: Cronqvist, H., et al., Value versus growth investing: Why do different investors have different
styles? Journal of Financial Economics (2015), http://dx.doi.org/10.1016/j.jfineco.2015.04.006i
 H. Cronqvist et al. / Journal of Financial Economics ] (]]]]) ]]]–]]] 3

with adverse macroeconomic experiences have stronger 2.2. Finance models

preferences for value investing later in life. For example,
those who grew up during the Great Depression have port- Numerous studies consider the return premium asso-
folios with average P/E ratios that are 1.7 (or about 11% at ciated with investments that are long value (cheap) stocks
the median) lower, controlling for individual characteristics. and short growth (expensive) stocks, both in US markets
Consistent with an impressionable years hypothesis, those (e.g., Fama and French, 1992) and outside the US (e.g.,
who entered the labor market for the first time during an Fama and French, 1998; Asness, Frazzini, and Pedersen,
economic recession are also more value oriented later on. 2012). What are the implications for the portfolio compo-
We also find that those who grew up in a lower status soc- sition of individual investors?
ioeconomic rearing environment have a stronger value orie-
ntation later in life.
The paper is organized as follows. Section 2 reviews related 2.2.1. Rational models
research. Section 3 introduces our data. Section 4 reports our Lynch (2001), Jurek and Viceira (2011), and Larsen and
results and robustness checks, and Section 5 concludes. Munk (2012) derive the optimal asset allocation for an
investor whose entire wealth consists of financial assets.
Differently from traditional portfolio choice models that
focus on the allocation between the risk-free asset and the
2. Related research equity market, these studies explicitly include value and
growth stocks with return properties estimated from US
In this section, we review models that have implications data. All three studies find that investors should hold value
for why some individual investors could prefer a more stocks due to the high Sharpe ratio of value stocks. This
value-oriented investment style while others could be more value tilt decreases with investors' risk aversion and with
growth-oriented.5 the investment horizon. Longer investment horizons lead
to a lower value tilt and generally to a relatively higher
allocation toward growth stocks because of investors'
intertemporal hedging demands.8 Intertemporal hedging
2.1. Evolutionary models
demands arise due to return predictability. For example,
mean reversion is smaller for value than for growth stocks,
Economists have for some time argued that preferences,
making value stocks relatively riskier over longer horizons,
for example with respect to risk, are shaped by evolutionary
consistent with evidence from asset pricing research (e.g.,
forces (e.g., Robson, 2001). Empirical research has provided
Campbell and Vuolteenaho, 2004).
evidence that variation of preferences across individual inves-
Fama (1996) and Cochrane (2007) consider the portfo-
tors reflects genetic differences between individuals. For
lio choice implications of the value premium from an
example, Cesarini, Dawes, Johannesson, Lichtenstein, and
equilibrium perspective. The value premium arises
Wallace (2009) and Barnea, Cronqvist, and Siegel (2010) show
because it represents exposure to a priced state variable.
that about 30% of the cross-sectional variation in financial risk
In these multifactor models, investors hold multifactor
preferences is explained by biological predispositions.6
efficient portfolios consisting of the market portfolio and a
Similarly, Rayo and Becker (2007) and Brennan and Lo
hedge portfolio, the mimicking portfolio for the state
(2011) present evolutionary models to explain how behaviors,
variable, for example, in the form of the Fama and French
which are not rational in standard economic models that
HML (high minus low) portfolio. Because the market
emphasize utility maximization by individual agents and are
portfolio needs to be held in equilibrium, an investor's
therefore considered behavioral biases, could emerge from
allocation will deviate from the market portfolio to the
human evolution.7 Recent research has again provided empiri-
extent that the investor is different from the average
cal evidence that is consistent with a relation between, on the
investor. Investors with below average risk aversion or
one hand, biological predispositions and, on the other hand,
with below average exposure to the state variable under-
behavioral biases (e.g., Cesarini, Johannesson, Magnusson, and
lying the value premium have a long position in the HML
Wallace, 2012; Cronqvist and Siegel, 2014).
portfolio and thereby a value tilt in their overall portfolio,
We thus examine to what extent investors' choices of
while more risk-averse investors with above average
value- versus growth-oriented investments reflect innate pre-
exposure to the underlying state variable tilt their portfolio
ferences or biases and show the relative importance of biol-
toward growth via a short position in the HML portfolio.
ogical and environmental factors for investors' invest-
The nature of the priced state variable underlying the
ment style.
value premium is still the object of much research in finance.
For example, Liew and Vassalou (2000) find a positive
5
For a review of empirical evidence related to value and growth correlation between HML and future gross domestic product
investing, see, e.g., Chan and Lakonishok (2004). (GDP) growth, suggesting that those particularly exposed to
6
Some research in the intersection of finance and neuroscience has
macroeconomic conditions would overweight growth stocks
even identified specific candidate genes involved in explaining differ-
ences in financial risk taking across individuals (e.g., Kuhnen and Chiao, at the expense of value stocks. Bansal and Yaron (2004) relate
2009; Dreber, Apicella, Eisenberg, Garcia, and Zamore, 2009; Zhong, the value premium to long-run consumption risk, such that
Israel, Xue, Ebstein, and Chew, 2009).
7
For further details and references, see, e.g., Cosmides and Tooby
8
(1994), Chen, Lakshminarayanan, and Santos (2006), and Santos and Larsen and Munk (2012) find only very slight horizon dependence
Chen (2009). due to small hedging positions.

Please cite this article as: Cronqvist, H., et al., Value versus growth investing: Why do different investors have different
styles? Journal of Financial Economics (2015), http://dx.doi.org/10.1016/j.jfineco.2015.04.006i
4 H. Cronqvist et al. / Journal of Financial Economics ] (]]]]) ]]]–]]]

long horizon investors should be expected to favor growth 2.2.2. Behavioral models
over value stocks. Fama and French (1996) suggest that While the above models provide a rational explanation
negative shocks to distressed value firms lower the value of of the value premium and of investors' asset allocation
human capital, implying that those with relatively more choices, several behavioral models suggest that the value
human capital should hold growth instead of value stocks. premium arises due to systematic mispricing. For example,
More recently, Gârleanu, Kogan, and Panageas (2012) show in Lakonishok, Shleifer, and Vishny (1994) argue that the
a general-equilibrium overlapping-generations model how return of growth, or glamour, stocks is not related to a
technological innovation reduces the value of existing firms source of systematic risk but is the result of investor
as well as the human capital of older workers. In their model, sentiment, and they provide evidence that value investing
though, financial capital is more exposed to this displacement results in higher returns because it exploits behavioral
risk than human capital, such that agents with more financial biases of some investors. In other behavioral models, the
wealth relative to total wealth favor growth stocks over value value premium reflects positive feedback trading (e.g., De
stocks. Long, Shleifer, Summers, and Waldmann, 1990; Hong and
Differently from the above models that introduce non- Stein, 1999; Barberis and Shleifer, 2003), conservatism, and
market risk state variables to explain the value premium, representativeness (e.g., Barberis, Shleifer, and Vishny,
others such as Zhang (2005) and Petkova and Zhang (2005) 1998) or overconfidence (e.g., Daniel, Hirshleifer, and
suggest that the value premium is consistent with the Subrahmanyam, 1998; Daniel, Titman, and Wei, 2001). As
conditional capital asset pricing model (CAPM).9 While the a result, we can expect that those who display more beh-
ability of conditional CAPMs to empirically fit asset return avioral biases have a stronger growth orientation.
data is unclear (e.g., Lewellen and Nagel, 2006; Gilbert,
Hrdlicka, Kalodimos, and Siegel, 2014), portfolio choice impli-
2.3. Life course theory
cations of conditional CAPM models differ from those above,
in particular with respect to investors' risk aversion. Specifi-
In this section, we review previous research related to
cally, differences in risk aversion across investors simply lead
life course theory.
to different allocations between the market portfolio and the
risk-free asset. That is, a particularly risk-seeking investor
uses leverage to increase her holding of the market portfolio 2.3.1. Macroeconomic experiences
instead of tilting her portfolio toward value stocks. In practice, Experiencing an adverse and significant macroeconomic
though, some individuals can face leverage constraints (e.g., event can have long-term and persistent effects on an
Constantinides, Donaldson, and Mehra, 2002) and could use individual's behavior much later in life. The Great Depression
value or growth stocks as a substitute for leverage.10 is the macro event that has so far been studied most in-
Finally, independently of the nature of the value premium depth in social science, and a variety of outcomes have been
and even if it reflects only time-varying market exposure, we examined. See Elder (1974) for one of the first and most
expect rational investors to hold value and growth stocks to extensive studies of the long-term effects of the Great
the extent that such tilts allow investors to hedge back- Depression. Several researchers have argued that the Great
ground risks such as labor income risk. For example, we Depression created a depression generation, whose behavior
expect a rational investor whose idiosyncratic labor income affected the macro economy for decades after the Depression
is negatively correlated with the return of value (growth) ended. For example, Friedman and Schwartz (1963) suggest
stocks, and who has not insured her labor income otherwise, that the Great Depression “shattered” beliefs in capitalism.
to overweight value (growth) stocks in her portfolio (e.g., In their Depression Babies study, Malmendier and Nagel
Cochrane, 2007; Davis and Willen, 2013).11 (2011) show that individuals who have experienced relatively
low stock market returns in their lives subsequently do not
participate in the stock market and they take significantly less
financial risk if they do participate. Others have analyzed the
importance of recent return experiences on the behavior of
9
Jagannathan and Wang (1996), PalaciosHuerta (2003), Santos and young investors in the 1990s (e.g., Vissing-Jørgensen, 2002;
Veronesi (2006), and Sylvain (2014) also offer (approximate) conditional Greenwood and Nagel, 2009). Graham and Narasimhan (2004)
CAPM explanations of the value premium. These models, though, require show that corporate executives who experienced the Great
the inclusion of human capital in the total wealth portfolio. Santos and Depression choose more conservative capital structures.
Veronesi (2006) show that these models imply two-factor models
consisting of the market return and the return to (aggregate) human
Recessions have also been shown to significantly affect out-
capital. comes ranging from fertility (e.g., Ben-Porath, 1973) to infants'
10
We thank the referee for this insight. Retail investors could use a and adults' health (e.g., Ruhm, 2000; Dehejia and Lleras-
value tilt if value stocks are riskier or they could use a growth tilt to Muney, 2004). Other economists have also found that macro
capture the stronger upside performance of growth stocks. If leverage
events have long-term effects on individual preferences. For
constraints are widespread, such a growth tilt could contribute to the
emergence of the value premium similar to the return difference example, Alesina and Fuchs-Schüendeln (2005) report that
between high and low beta stocks in Asness, Frazzini, and Pedersen post-reunification East Germans (particularly older cohorts)
(2012) and Frazzini and Pedersen (2014). have stronger preferences for, e.g., redistribution than other-
11
Empirical research in finance has so far produced inconclusive wise similar West Germans. More recently, Malmendier and
evidence on whether individual investors' hedging demands affect
portfolio choice (e.g., Heaton and Lucas, 2000; Massa and Simonov,
Nagel (2013) show that differences in life experiences of high
2006; Bonaparte, Korniotis, and Kumar, 2014; Addoum, Korniotis, and or low inflation predict differences in subjective inflation
Kumar, 2014). expectations.

Please cite this article as: Cronqvist, H., et al., Value versus growth investing: Why do different investors have different
styles? Journal of Financial Economics (2015), http://dx.doi.org/10.1016/j.jfineco.2015.04.006i
 H. Cronqvist et al. / Journal of Financial Economics ] (]]]]) ]]]–]]] 5

Those who have more salient experiences of difficult preschool (kindergarten) environment explains some asset
economic conditions, characterized by the absence of fin- allocation decisions later in life, such as contributing to a 401
ancial resources, can develop a more value-oriented inve- (k) retirement savings plan and owning a home.13
stment style, with a preference for stocks that could seem In this study, we focus on the rearing environment
relatively cheaper. within the family during an individual's upbringing. We
hypothesize that the socioeconomic status of the rearing
2.3.2. Impressionable years environment in which an individual grows up explains
Several studies in social psychology suggest that experi- differences in investment style later in life.14
ences in early adulthood are important for preferences later
in life (e.g., Krosnick and Alwin, 1989). An individual's core 3. Data
beliefs and preferences seem to crystallize during a period of
great neurological plasticity in early adulthood – the so- In this section, we introduce our data.
called impressionable years – and remain largely unchanged
thereafter.12 In economic research, Giuliano and Spilimbergo
3.1. Individual characteristics
(2014) has recently showed that experiencing an economic
recession during the impressionable years (18–25 years old)
To study the extent to which variation in investment
significantly affects redistribution and political preferences
styles across a large sample of individual investors
much later in life.
reflects innate differences, we employ data on identical,
We also examine whether an individual entered the labor
or monozygotic (MZ), and fraternal, or dizygotic (DZ),
market for the first time in an economic recession. This mea-
twins. We construct our data set by matching a large
sure comes with the caveat that it is somewhat less exo-
number of twins from the Swedish Twin Registry (STR),
genous compared with the impressionable years because
the world's largest twin registry, with data from indivi-
individuals can to some extent endogenously choose when
dual tax filings and other databases. In Sweden, twins are
they enter the labor market by increasing their investment in
registered at birth, and the STR collects additional data
education. We still find it informative to examine the time of
through in-depth interviews.15 STR's data provide us with
an individual's labor market entry because it has been shown
the zygosity of each twin pair.16
to be important in other studies of economic outcomes (e.g.,
Table 1 reports summary statistics for the twins in our
Malmendier, Tate, and Yan, 2011; Schoar and Zuo, 2013). This
data set and their individual characteristics. Panel A
is also a period when many individuals start to invest in the
shows that we have data on a total of 10,490 identical
stock market, so it seems possible that an individual's
twins and 24,486 fraternal twins, who participate in the
investment style could also be affected.
stock market. Opposite-sex twins are the most common
An effect on investment style of labor market condi-
(38%); identical male twins are the least common (13%).
tions at the time of an individual's first employment can
Panel B reports summary statistics for individual char-
also be less direct. For example, Oyer (2008), Kahn (2010),
acteristics, including age, education, marital status, net
and Oreopoulos, von Wachter, and Heisz (2012) report that
worth, and disposable income, which we include as
the market conditions at the time of an individual's labor
controls when we estimate models in Section 4. The
market entry have persistent effects on earnings because
average size of investors' holdings of stocks and equity
of the initial labor market conditions affecting an employ-
mutual funds in our data set, about $33,500, is compar-
ee's task-specific human capital accumulation. That is, an
able to those in other data sets of a broad set of individual
effect on an individual's value versus growth orientation
could be indirect through an effect on earnings. As a result,
it is important to control for earnings in our analysis. 13
Even the pre-birth, i.e., in utero, environment has been shown to
predict subsequent economic outcomes and behaviors. See, e.g., Almond
and Currie (2011) and Cronqvist, Previtero, Siegel, and White (2014).
2.3.3. Rearing environment 14
We refer to, e.g., Bisin and Verdier (2000, 2001) for work related to
The argument that the rearing environment, and other the social transmission of preferences and behavior from parents to their
early life experiences, can have significant long-term and children.
15
persistent effects on an individual's behaviors later in life has STR's databases are organized by birth cohort. The Screening
recently made its way into economic research. Most existing Across Lifespan Twin, or SALT, database contains data on twins born
1886–1958. The Swedish Twin Studies of Adults: Genes and Environment
studies investigate outcomes such as education and earnings. database, or STAGE, contains data on twins born 1959–1985. In addition
For example, economists have shown that birth order and to twin pairs, twin identifiers, and zygosity status, the databases contain
family size affect educational attainment and earnings later variables based on STR's telephone interviews (for SALT), completed
in life (e.g., Black, Devereux, and Salvanes, 2005). Relatively 1998–2002, and combined telephone interviews and Internet surveys
(for STAGE), completed 2005–2006. For further details about STR, see
few studies examine outcomes of primary interest to finan-
Lichtenstein, Sullivan, Cnattingius, Gatz, Johansson, Carlström, Björk,
cial economists. An exception is Chetty, Friedman, Hilger, Svartengren, Wolk, Klareskog, de Faire, Schalling, Palmgren, and
Saez, Schanzenbach, and Yagan (2011) who report that the Pedersen (2006).
16
Zygosity is based on questions about intra-pair similarities in
childhood. One of the questions was: Were you and your twin partner
12
Recent research on neurological development shows that, in the during childhood “as alike as two peas in a pod” or were you “no more
developing brain, the volume of gray matter in the cortex gradually alike than siblings in general” with regard to appearance? STR has
increases until about adolescence but then sharply decreases as the brain validated this method with DNA analysis as having 98% accuracy on a
prunes away neuronal connections that are deemed superfluous to the subsample of twins. For twin pairs for which DNA (deoxyribonucleic acid)
adult needs of the individual (e.g., Spear, 2000). has been collected, zygosity status is based on DNA analysis.

Please cite this article as: Cronqvist, H., et al., Value versus growth investing: Why do different investors have different
styles? Journal of Financial Economics (2015), http://dx.doi.org/10.1016/j.jfineco.2015.04.006i
6 H. Cronqvist et al. / Journal of Financial Economics ] (]]]]) ]]]–]]]

Table 1
Summary statistics: individual characteristics.
This table reports summary statistics for the sample of twins and their individual characteristics. The variables are defined in Appendix A.

Panel A: Twins by zygosity and gender

Identical twins Fraternal twins

Same

Same sex: sex: Opposite

Variable All twins Male Female All male female sex All

N 34,976 4,496 5,994 10,490 5,064 6,300 13,122 24,486

Percentage 100 13 17 30 14 18 38 70

Panel B: Individual characteristics All twins Identical twins Fraternal twins

Variable N Mean Median Standard deviation Mean Median Standard deviation

Age 153,743 50.88 52 16.39 55.41 56 14.15

College or More 153,743 54% 100% 50% 45% 0% 50%
High School 153,743 22% 0% 41% 25% 0% 43%
No Education Data Available 153,743 9% 0% 29% 10% 0% 30%
Married 153,743 53% 100% 50% 59% 100% 49%
Net Worth (US dollars) 153,030 140,934 73,336 336,746 153,406 80,819 728,368
Financial Wealth (US dollars) 153,743 63,500 23,000 263,750 68,250 25,375 633,750
Share in Equities 153,030 58% 61% 34% 57% 59% 34%
Disposable Income (US dollars) 153,743 36,418 27,813 43,403 39,661 29,434 66,227
Labor Income (US dollars) 153,030 32,875 29,500 24,500 33,250 29,000 27,000
Labor Income Correlation 152,425
3%
4% 40%
5%
7% 39%
Investment Bias Index 123,158 4.28 4.00 2.34 4.24 4.00 2.31
Volatility 132,042 26% 23% 14% 25% 23% 14%

investors, e.g., 24,600 euros in Grinblatt and Keloharju individual, we compute the value-weighted average ratio for
(2009).17 each year in the panel. Appendix A reports definitions for our
investment style measures. For mutual funds, we also construct
3.2. Investment-style measures two measures: Morningstar's Value-Growth Score, which
varies from
100 (value) to þ400 (growth); and name-
Prior to the abolishment of the wealth tax in Sweden in based Value/Growth Measure, which is
1 if a fund's name
2007, all Swedish banks, brokerage firms, and other financial contains “value,” þ1 if a fund's name contains “growth” or
institutions were required by law to report to the Swedish Tax “technology,” and zero otherwise. We use the same method as
Authority information about individuals' portfolios (i.e., stocks, for stocks to construct a measure for each individual and year.
bonds, mutual funds, and other securities) owned on Decem- Panel A of Table 2 reveals that while identical and fra-
ber 31. We have matched the individuals in our data set with ternal twins are relatively similar with respect to these value
portfolio data between 1999 and 2007, the entire period for versus growth orientation measures, significant variation
which data are available. For each individual, our data set exists in investment style across different investors. It is this
contains all securities owned at the end of the year (identified variation that we decompose and explain in Section 4. Panel
by each security's International Security Identification Number B shows that all measures of the value versus growth orie-
(ISIN)), the number of each security owned, and the end of the ntation of investors' portfolios are significantly positively
year value. Security level data have been provided by S&P correlated, suggesting that investors have a consistent pre-
CapitalIQ and Morningstar. In our data set, there is a total of ference for certain investment styles.
about two thousand different individual stocks and about one
thousand different mutual funds. 4. Results
We categorize each investor's value versus growth tilt on a
continuum. For individual stocks, we construct two measures We first examine to what extent an individual's invest-
of value versus growth orientation using different scaled prices: ment style is explained by biological predispositions. We
P/E Ratio (price/earnings) and P/B Ratio (price/book).18 For each then investigate the relation between individual charac-
teristics and an investor's value versus growth orientation.
Finally, we analyze environmental experiences that life
17
We use the average end-of-year exchange rate 1999–2007 of
8.0179 Swedish krona per US dollar to convert summary statistics in
the table. When we estimate models in Section 4, all values are in (footnote continued)
Swedish krona. level does not change any of the reported results. Our results are also
18
We check that our results are robust to outliers. Following Capital robust to variable transformations (e.g., log) used to reduce the skewness
IQ's practice, the ratios are censored at 0 and 300. Winsorizing at the 1% of scaled price distributions.

Please cite this article as: Cronqvist, H., et al., Value versus growth investing: Why do different investors have different
styles? Journal of Financial Economics (2015), http://dx.doi.org/10.1016/j.jfineco.2015.04.006i
 H. Cronqvist et al. / Journal of Financial Economics ] (]]]]) ]]]–]]] 7

Table 2
Summary statistics: investment style measures.
This table reports summary statistics (Panel A) and correlations (Panel B) for the investment-style measures. The variables are defined in Appendix A.
nnn
indicates significance at the 1% level.

Panel A: Summary statistics

Identical twins Fraternal twins

Variable N Mean Median Standard deviation Mean Median Standard deviation

Individual stocks
P/E Ratio 138,063 23.3 15.4 22.9 22.5 14.9 21.8
P/B Ratio 151,729 3.3 2.3 3.6 3.2 2.2 3.4
Mutual funds
Morningstar's Value-Growth Score 147,818 156.0 153.6 25.6 154.8 152.8 24.8
Name-based Value/Growth Measure 195,438 0.08 0.00 0.22 0.07 0.00 0.21

Panel B: Correlations

Morningstar

Variable P/E Ratio P/B Ratio Value-Growth Score

P/B Ratio 0.26nnn

(N¼ 77,980)
Morningstar's Value-Growth Score 0.06nnn 0.04nnn
(N¼ 59,577) (N ¼65,341)
Name-based Value/Growth Measure 0.06nnn 0.02nnn 0.51nnn
(N¼ 77,981) (N ¼85,509) (N ¼66,017)

course theory suggests can influence an individual's inve- 4.1.1. Evidence from correlations
stment style. Fig. 1 reports correlations by genetic similarity, i.e., for
identical twins and fraternal twins (separately for same-
and opposite-sex twins), for measures of value versus
growth orientation. Panel A presents results for individual
4.1. Biological predispositions and investment style stocks; Panel B, for mutual funds.
Several conclusions emerge from this exercise. First, we
In this section, we first report separate correlations for find that identical twins' investment styles are significantly
identical versus fraternal twins for each of our measures of more correlated compared with fraternal twins. For example,
investment style. We then provide formal estimation the Pearson correlation coefficient among identical twins is
results from decomposing the variation in investment 0.49 for the average P/E ratio of their stock holdings, com-
style into genetic and environmental components. To do pared with 0.35 among fraternal twins (0.34 for opposite-sex
so, we use empirical methods from quantitative behavioral fraternal twins). Using the P/B ratio, we find a correlation of
genetics research that have recently been employed also in 0.39 for identical twins and of 0.25 for fraternal twins, again
research in economics (e.g., Cesarini, Dawes, Johannesson, confirming a larger correlation for identical twins relative to
Lichtenstein, and Wallace, 2009; Barnea, Cronqvist, and fraternal twins. A similar conclusion emerges for mutual
Siegel, 2010). The approach involves maximum likelihood funds. For example, the Pearson correlation coefficient among
estimation (MLE) of a random effects model but relies on identical twins is 0.31 for the average Value-Growth Score by
an intuitive and simple insight: Identical twins share 100% Morningstar for their mutual fund holdings, compared with
of their genes, while the average proportion of shared only 0.14 for fraternal twins (0.12 for opposite-sex fraternal
genes is only 50% for fraternal twins,19 so if identical twins twins). Similarly, for our name-based value versus growth
are more similar with respect to their investment styles measure we find a correlation of 0.44 for identical twins and a
than are fraternal twins, then there is evidence that value correlation of 0.29 for fraternal twins. That is, genetically more
versus growth orientation is partly explained by genetic similar investors have more similar investment styles. This
predispositions. For further details, see Appendix B. evidence suggests that genetic differences affect value versus
growth orientation among individual investors.
Second, we find that the correlations among identical
twins are significantly below one. That is, even genetically
19
Genome sequencing has recently revealed that humans and, e.g., identical investors show significant differences with respect
the common chimpanzee (Pan troglodytes) share about 96% of their genes to their investment styles. This evidence shows the impor-
(e.g., Mikkelsen, 2005), and the genetic overlap is even greater among tance of the environment in shaping an investor's value
humans. That is, the 50% refers to only the proportion of genes that
makes different humans different from each other. For some individual
versus growth orientation, and it emphasizes the impor-
characteristics, particularly physical attributes, there is no variation tance of analyzing the effect on investment style of experi-
across individuals. ences and events during an individual's life course.

Please cite this article as: Cronqvist, H., et al., Value versus growth investing: Why do different investors have different
styles? Journal of Financial Economics (2015), http://dx.doi.org/10.1016/j.jfineco.2015.04.006i
8 H. Cronqvist et al. / Journal of Financial Economics ] (]]]]) ]]]–]]]

Idential twins Fraternal twins Fraternal twins, same sex Fraternal twins, opposite sex
0.50
0.45
Pearson Correlation Coefficient
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
P/E Ratio P/B Ratio

Idential twins Fraternal twins Fraternal twins, same sex Fraternal twins, opposite sex
0.50

0.45
Pearson Correlation Coefficient

0.40

0.35

0.30

0.25

0.20

0.15

0.10

0.05

0.00
Morningstar's Value-Growth Score Name-based Value/Growth Measure

Fig. 1. Correlations by genetic similarity (individual stocks and mutual funds). Panel A reports Pearson correlation coefficients for investment style for
different types of twin pairs. The investment style measures are calculated using individual stock holdings only. Panel B reports Pearson correlation
coefficients for investment style for different types of twin pairs. The investment style measures are calculated using mutual fund holdings only. All
variables are defined in Appendix A.

4.1.2. Evidence from variance decomposition between 26% and 40% and are statistically significant for each
Tables 3 and 4 report results from decomposing invest- investment-style measure. The C component is significantly
ment styles into genetic and environmental variation. We smaller and varies between 0% and 11%. The remaining
report the proportions of the variation in investment styles variation in investment style is explained by individual-
across individuals that are explained by genetic (A), common specific experiences and events.21
environmental (C), and individual-specific environmental (E) While the evidence reported so far involves individual
factors (for details, see Eq. (5) in Appendix B). We first regress stocks, we also decompose the variation in investment style
each investment style measure on a set of individual char- using data on mutual funds. We use two measures, the Value-
acteristics as well as year fixed effects and then we decompose Growth Score by Morningstar and a name-based value/
the residual variation. We include the following individual growth measure. These measures provide a salient way for
characteristics in each case: gender, age, education, marital an individual investor to choose exposure based on his or her
status, disposable income, and net worth. value versus growth preference. Our conclusions from Table 3
The evidence in Table 3 confirms the correlation evidence are not affected. The estimates of the A component vary
and shows that variation across investors with respect to
value versus growth orientation of their stock portfolio is
partially genetic.20 The estimates of the A component vary (footnote continued)
components for these measures are similar to those reported for standard
measures such as P/E and P/B ratios.
21
The E component is also absorbing idiosyncratic measurement
20
We focus on the most commonly used measures in practice related error. Because our data set comes from the Swedish Tax Agency, which
to value versus growth orientation. As a robustness check, we also obtains the data directly from financial institutions, reporting errors
examine price/sales (P/S) and price/cash flow (P/C) ratios. The A should be relatively rare in our specific sample.

Please cite this article as: Cronqvist, H., et al., Value versus growth investing: Why do different investors have different
styles? Journal of Financial Economics (2015), http://dx.doi.org/10.1016/j.jfineco.2015.04.006i
 H. Cronqvist et al. / Journal of Financial Economics ] (]]]]) ]]]–]]] 9

Table 3 variation when including a large set of individual char-

Variance decomposition of investment style: individual stocks. acteristics (e.g., Brunnermeier and Nagel, 2008). Overall,
This table reports results from maximum likelihood estimation. The
we conclude that an individual's investment style has a
different investment style measures are modeled as linear functions of
observable individual characteristics (gender, age, education, marriage biological basis, i.e., a preference for value versus growth
status, disposable income, and net worth), year fixed effects, and unobser- stocks is partially ingrained in an investor from birth.
vable random effects representing additive genetic effects (A), share It is important to emphasize several assumptions, and
environmental effects (C), and an individual-specific error (E). For each associated caveats, of the models we estimate. First, the equal
estimated model, we report the variance fraction of the residual explained
by each unobserved effect (A Share, for the additive genetic effect; C Share,
environments assumption (EEA) is one important assump-
for common environmental effect; E Share, for the individual-specific tion. If an individual's parents or others in the environment
environmental effect) and the bootstrapped standard errors (one thousand treat identical twins more similarly than fraternal twins, then
resamples). The variables are defined in Appendix A. nnn and nn indicate the estimated genetic component (A) can be upward-biased.
significance at the 1% and 5% level, respectively.
From research in behavioral genetics research, when the EEA
Estimate P/E Ratio P/B Ratio has been challenged most rigorously, the evidence sugg-
ests that bias from violations of the EEA is not first order
A Share 0.255nn 0.396nn (e.g., Bouchard, 1998).22 Second, the model we estimate
0.067 0.094 assumes an additive genetic component. We also estimate
C Share 0.112nn 0.000
0.046 0.049
ADE models, with an additional factor for dominance genetic
E Share 0.632nnn 0.604nnn variation. We find that the resulting D components are small
0.027 0.060 in magnitude and not statistically significant from zero, while
Individual characteristics Yes Yes the A components continue to account for a similar propor-
Year fixed effects Yes Yes
tion of the variation. This reduces concerns that our estimated
N 63,592 63,592
model is misspecified. Third, including opposite-sex twins in
our analysis may result in a bias in favor of a genetic
component, but we have checked that our results are similar
if we restrict our analysis to only same-sex twins. Fourth,
Table 4
Variance decomposition of investment style: mutual funds.
positive assortative mating between the twins' parents would
This table reports results from maximum likelihood estimation. The make fraternal twins more similar relative to identical twins
different investment style measures are modeled as linear functions of and would bias the estimate of the genetic (A) component
observable individual characteristics (gender, age, education, marriage downward (e.g., Neale and Maes, 2004). Finally, the model we
status, disposable income, and net worth), year fixed effects, and unobser-
estimate does not consider interactions between genes and
vable random effects representing additive genetic effects (A), shared
environmental effects (C), and an individual-specific error (E). For each the environment (e.g., some environmental factors can pre-
estimated model, we report the variance fraction of the residual explained vent genetic factors from expressing themselves) and genetic
by each unobserved effect (A Share, for the additive genetic effect; C Share, and environmental factors might not be uncorrelated (e.g.,
for common environmental effect; E Share, for the individual-specific
genetic factors could steer the individual to be exposed to
environmental effect) and the bootstrapped standard errors (one thousand
resamples). The variables are defined in Appendix A. nnn and nn indicate
certain environments).23
significance at the 1% and 5% level, respectively.

4.2. Individual characteristics and investment style

Estimate Morningstar's Value- Name-based Value/
Growth Growth
Score Measure We examine which individual characteristics explain an
investor's investment style, and we relate the evidence to
nnn nn
A Share 0.273 0.257
rational and behavioral portfolio choice and asset pricing
0.011 0.014
C Share 0.000 0.000 models discussed in Section 2.2.2. We run panel regres-
0.000 0.006 sions of each investor's annual P/E ratio, the value-
E Share 0.727nnn 0.743nnn weighted average of all stocks in the investors' portfolio
0.011 0.011 at the end of the year, on investor characteristics typically
Individual Yes Yes
determined in the previous year. All specifications include
characteristics
Year fixed effects Yes Yes year fixed effects. We use two-way clustering by year and
N 85,388 85,388 twin pair (e.g., Thompson, 2011), i.e., the residual correla-
tion can be different from zero if observations are in the
same year or in the same twin pair.
In Column 1 of Table 5, we regress the average P/E ratio
between 26% and 27%, and they are statistically significant on standard socioeconomic characteristics. We also
for both investment-style measures, while estimates of the include the proportion of an individual's financial assets
C component are 0% in both cases. invested in risky equities, which is a common measure of
Even though estimates of the A component based on
holdings of individual stocks (Table 3) are less precise than
22
those based on holdings of mutual funds (Table 4), each of Researchers have studied twins reared apart for which there is no
the investment style measures reveals a statistically sig- common parental environment, and these studies generally produce
estimates similar to those using twins who were reared together (e.g.,
nificant A component. It should also be emphasized that Bouchard, Lykken, McGue, Segal, and Tellegen, 1990).
recent studies related to individual investor behavior have 23
For an extensive review of research on gene-environment inter-
had difficulties explaining even 10% of the cross-sectional actions, see Rutter (2006).

Please cite this article as: Cronqvist, H., et al., Value versus growth investing: Why do different investors have different
styles? Journal of Financial Economics (2015), http://dx.doi.org/10.1016/j.jfineco.2015.04.006i
10 H. Cronqvist et al. / Journal of Financial Economics ] (]]]]) ]]]–]]]

Table 5
Individual characteristics and investment style.
This table reports coefficient estimates from panel regressions with year fixed effects. Robust standard errors are two-way clustered by twin pair and
year. The variables are defined in Appendix A. nnn and nn indicate significance at the 1% and 5% level, respectively.

Variable (1) (2) (3) (4) (5)

Age
0.15nnn
0.10nnn
0.08nnn
0.06nnn
0.07nnn

(0.05) (0.04) (0.03) (0.02) (0.03)
Male
0.19 0.07
0.20
0.49nn
0.31
(0.24) (0.22) (0.19) (0.21) (0.23)
Married 0.21 0.35n 0.23 0.16 0.34
(0.24) (0.19) (0.18) (0.18) (0.25)
Log (Net Worth) 0.56
0.38
0.52nn
0.51n
0.49n
(0.40) (0.26) (0.26) (0.26) (0.29)
Log (Disposable Income) 1.71nnn 0.54n 0.27 0.26 0.30
(0.57) (0.33) (0.28) (0.24) (0.27)
Share in Equities 2.30 2.99n 2.83n 2.65 2.82
(1.41) (1.64) (1.66) (1.65) (1.77)
Log (Financial Wealth) 1.31nnn 1.34nn 1.29nn 1.34nnn
(0.49) (0.53) (0.51) (0.51)
Log (Labor Income) 0.84nnn 0.78nnn 0.62nn 0.63nnn
(0.25) (0.29) (0.26) (0.22)
Labor Income Correlation 0.60nnn 0.69nnn 0.68nnn 0.56nnn
(0.18) (0.21) (0.22) (0.21)
College or More 1.55nnn 1.31nnn 1.12nnn 0.85nn
(0.38) (0.45) (0.39) (0.38)
High School 0.96nnn 0.89nnn 0.80nnn 0.60nn
(0.29) (0.29) (0.26) (0.25)
No Education Data Available 0.51 0.69 0.54 0.35
(0.44) (0.49) (0.51) (0.56)
Investment Bias Index 0.67nnn 0.11 0.11
(0.24) (0.32) (0.33)
Volatility 25.20nnn 22.64nnn
(8.24) (8.35)

Constant 7.08 3.86 5.68 3.18 3.00

(9.06) (8.40) (8.12) (8.37) (8.96)
Year fixed effects Yes Yes Yes Yes Yes
N 106,518 106,518 77,078 77,078 49,528
Sample All All All All Leverage o 20%
R-squared 0.19 0.19 0.20 0.22 0.22

investors' financial risk-taking propensity (e.g., Merton, include investors' educational achievement,24 their labor
1969; Samuelson, 1969). We find that portfolios of older income, and the full-sample correlation between investors'
(younger) investors are significantly more value- (growth- labor income growth and changes in GDP per capita.
) oriented. The average P/E ratio of the stock portfolio of a Consistent with the predictions of multifactor efficient
65 year old investor is 6.0 (39% at the median) lower portfolio choice models (e.g., Fama, 1996; Cochrane,
compared with a 25-year-old. If younger investors have 2007) and the view that the value premium is related to
longer investment horizons, this result is consistent with human capital, we find that investors' with more human
the portfolio choice models of Lynch (2001), Jurek and capital in the form of more education and higher levels of
Viceira (2011), and Larsen and Munk (2012). At the same labor income prefer growth stocks, as do investors whose
time, the results in Column 1 do not suggest that investors' labor income covaries more positively with GDP growth.
risk aversion is significantly related to investors' value or That is, investors whose labor income is reduced in bad
growth tilts. The share invested in risky equities as well as states of the world prefer growth over value stocks, a
investors' net worth have a positive, but insignificant effect behavior that seems consistent with models in which the
on the average P/E ratio. Men who are often more risk value premium represents compensation for distress risk
seeking than women (e.g., Sundén and Surette, 1998; (e.g., Fama and French, 1993). We also find support for the
Croson and Gneezy, 2009) have an insignificant value tilt prediction by Gârleanu, Kogan, and Panageas (2012) that
in their portfolios. Besides age, only disposable income has in a world with technological innovation and displacement
a significant effect on the portfolio's value versus growth risk agents with relatively more financial wealth tilt their
orientation. Possibly consistent with human capital-rel- portfolios toward growth stocks. Overall these results are
ated hedging motives, investors with higher disposable in line with predictions of rational models for the value
income exhibit a preference for growth stocks. premium and consistent with Betermier, Calvet, and Sodini
In Column 2, we examine such hedging motives in
more detail. First, to better capture investors' human
capital and its exposure to a possible state risk factor, we 24
“Less than a high school degree” is the omitted educational outcome.

Please cite this article as: Cronqvist, H., et al., Value versus growth investing: Why do different investors have different
styles? Journal of Financial Economics (2015), http://dx.doi.org/10.1016/j.jfineco.2015.04.006i
 H. Cronqvist et al. / Journal of Financial Economics ] (]]]]) ]]]–]]] 11

(2014) who also analyze Swedish individual investors' The result in Column 1 in Panel A of Table 6 shows that
portfolio style and report similar findings. individuals who grew up during the Great Depression
In addition to risk-taking and hedging motives, we also show significantly more value orientation in their stock
explore whether investors who exhibit more behavioral biases portfolios several decades later in life. We find that those
in the investment domain are more growth oriented. Our who grew up during the Great Depression have portfolios
evidence is also supportive of such a prediction. The results in with average P/E ratios that are about 1.7 (11% at the
Column 3 of Table 5 suggest that a 1 standard deviation median) lower compared with those of other investors. It
increase in the bias index of Cronqvist and Siegel (2014) is important to emphasize that we control for disposable
corresponds to an average P/E ratio that is about 1.6 higher income and net worth, which can also be affected by a
(10% at the median).25 Several authors have suggested that Great Depression experience, so our results are not simply
investors could prefer stocks with speculative characteristics reflecting long-term wealth differences. In Column 2 we
such as high volatility (e.g., Shefrin and Statman, 2000; Barberis analyze the subsample of those born 1910 to 1939, i.e., a
and Huang, 2008; Kumar, 2009b; Dorn and Huberman, 2010). subset of only older individuals. Some of these individuals
In Column 4, we explicitly test for this specific explanation and were treated by growing up during the Great Depression,
also include the contemporaneous annualized monthly volati- while others were arguably not treated as severely but are
lity of an investor's portfolio. The effect of the bias index slightly younger or slightly older than the Great Depres-
decreases and becomes insignificant, and we find strong sion cohort. The previous conclusion of a Depression Baby
support that investors who select high volatility portfolios also effect remains. While economically still significant, the
tilt their portfolio toward growth stocks. effect is statistically weaker (p-value ¼11.9%).
Finally, investors could select volatile growth stocks Even with age as a control variable, the short sample
because they face leverage constraints that prevent them period makes it challenging to separate a Great Depression
from increasing the riskiness of their portfolio otherwise. In cohort effect from a life-cycle effect. We therefore analyze
Column 5, we therefore consider a subset of investors whose an individual's overall GDP growth experience during his
debt to asset ratio is at most 20% and often zero and who are or her life as an alternative measure of macroeconomic
therefore unlikely leverage constraint. The results are essen- experiences. We measure the average GDP growth from an
tially unchanged relative to those of the full sample and individual's birth year until the start of our data set. In
suggest that leverage constraints do not explain investors' Column 3, we find that experiencing stronger GDP growth
portfolio choice with respect to value and growth stocks. results in a relatively more growth-oriented investment
Overall, we find support for rational as well as behavioral style. We find that for a 100 basis points per year higher
finance models. In particular, our results suggest that investors average GDP growth experience, the average P/E ratio of
hedge human capital and displacement risks with growth the stock portfolio of the investor is 1.4 higher (9%
stocks. At the same time, we find that investors' biases, in compared with the median) compared with those of other
particular a possible preference for lottery type stocks, also investors, controlling for individual characteristics.
make them favor growth over value stocks. Therefore, the Research in financial economics has so far provided
value premium could reflect both a risk-based compensation inconclusive evidence on whether more distant or more
and mispricing due to speculative retail investors (see, e.g., recent experiences are weighted more in individual investor's
Han and Kumar, 2013). financial decision making. Some recent papers, e.g.,
Malmendier and Nagel (2011), explicitly estimate a weighting
4.3. Life course theory and investment style function and report that more recent stock market returns
affect stock market participation and financial risk-taking
In this sub-section, we examine to what extent differential relatively more than distant experiences. Other papers (e.g.,
life experiences and events of individuals explain differences Cronqvist, Previtero, Siegel, and White, 2014) emphasize early
in investment styles much later in life. Based on pre-existing life experiences (even the prenatal and in utero environment)
research in social psychology and neuroscience, we consider for financial risk taking propensities later in life, following a
several types of potentially relevant, and plausibly exogenous, large literature in economics related to the fetal origins
life experiences of individuals: macroeconomic experiences, hypothesis (see, e.g., Currie, 2011 and the references therein).
the impressionable years, and the rearing environment. Some papers emphasize the early postnatal environment (e.g.,
4.3.1. Macroeconomic experiences Cunha and Heckman, 2010). For example, Chetty, Friedman,
First, we analyze whether a long-term and persistent Hilger, Saez, Schanzenbach, and Yagan (2011) show that the
effect exists on an individual's investment style of growing pre-school (kindergarten) environment explains some asset
up during the Great Depression. We investigate the effect allocation decisions later in life. Still other papers, e.g., Schoar
on value versus growth orientation of being born between and Zuo (2013), report empirical evidence consistent with the
1920 and 1929, using the same Depression Baby definition importance of the impressionable years and the first labor
as Schoar and Zuo (2013).26 market entry. We have reestimated the result for individual
GDP growth experience using a simple weighting function.
We assign proportionately increasing and decreasing weights
25
The bias index covers six prominent investment biases. It is available to more recent experiences. For an individual who is T years
only as a time-invariant measure. For details, see Cronqvist and Siegel (2014). old, the weights for year i ¼ 1…T are specified by
26
Sweden was affected by the Wall Street Crash of 1929 and was also PT PT
the origin of the so-called Kreuger Crash of 1932, with adverse interna- i ¼ 1 iw ¼ 1 and i ¼ 1 ðT þ1
iÞw ¼ 1, respectively, where
tional macroeconomic consequences deepening the Great Depression in w A ½0; 1. We find a statistically insignificant relation, i.e., for
several countries, including the US. value versus growth orientation we cannot conclude that

Please cite this article as: Cronqvist, H., et al., Value versus growth investing: Why do different investors have different
styles? Journal of Financial Economics (2015), http://dx.doi.org/10.1016/j.jfineco.2015.04.006i
12 H. Cronqvist et al. / Journal of Financial Economics ] (]]]]) ]]]–]]]

Table 6
Life course theory and investment style: macroeconomic experiences.
This table reports coefficient estimates from panel regressions with individual characteristics and year fixed effects. Robust standard errors are two-way
clustered by twin pair and year. The variables are defined in Appendix A. nnn and nn indicate significance at the 1% and 5% level, respectively.

(1) (2) (3)

Panel A: Great Depression and gross domestic product (GDP)
growth experiences
Depression Baby
1.74nnn
0.72
0.64 0.46
Individual GDP Growth Experience 1.42nn
0.57
Constant 5.10nnn
1.40 2.49
1.35 3.98 1.64
Individual characteristics Yes Yes Yes
Year fixed effects Yes Yes Yes
Sample All Born 1910 to 1939 All
N 107,658 31,932 107,658
R-squared 0.19 0.17 0.19

Panel B: Individual high minus low (HML) return experience

Individual Sweden HML Experience 0.01
0.47
0.02 0.41
Individual Sweden HML Experience 0.72
 Professional Finance Experience 1.02
Professional Finance Experience
3.34
5.49
Constant 9.42nnn 19.56nn
(3.18) (8.70)
Individual Characteristics Yes Yes
Year Fixed Effects Yes Yes
Sample First job after 1975 First job after 1975
N 27,523 8,755
R-squared 0.20 0.18

more distant, or more recent, GDP experiences are more market entry and the impressionable years (18–25 years).
important. There exists no similar classification of recessions to the
Finally, we examine whether Individual HML Experience, National Bureau of Economics Research's business cycle
based on the return on a Swedish HML factor portfolio, database for our sample, so we analyze several alternative
explains subsequent investment styles. That is, do those who measures of recessions. We define a recession as a period
have experienced higher HML returns during his or her life with a year of negative GDP growth71 year.28 Column 1
develop into a more value-oriented investor? We construct in Panel A of Table 7 shows that individuals who entered
the measure in a similar way to the individual GDP growth the labor market during a recession have portfolios with
experience measure using data provided by Kenneth an average P/E ratio that is 0.9 lower (6% at the median)
French.27 In Column 1 in Panel B of Table 6, we report no compared with those of other investors, although the
statistically significant relation between individual HML exp- effect is not statistically significant.
erience and value versus growth orientation. For a subsa- In Column 2 we report results for the most severe
mple we are also able to collect detailed data on professi- economic recessions someone in our sample experienced.
onal finance experience based on the International Standard We find that those who entered the labor market in a
Classification of Occupations (ISCO-88) by the International severe recession, i.e., during World War I, the Great
Labour Organization (ILO). We then interact individual HML Depression, or World War II, have portfolios with average
experience with an indicator variable for professional finance P/E ratios that are 3.2 lower (21% at the median) compared
experience. In Column 2, we find that the interaction effect is with those of other investors. That is, the estimated effect is
statistically insignificant, i.e., more financially sophisticated about three times larger for experiencing the most severe
investors do not seem influenced by HML experience. recessions. It is important to emphasize that our model
specification controls for disposable income and net worth,
i.e., there is a direct effect of economic recessions on inve-
4.3.2. Impressionable years
stment style later in life, in addition to any indirect effect
We also examine persistent effects of the economic
from lower income of those who entered the labor market
conditions during the years of an individual's first labor
in economic recessions (e.g., Oreopoulos, von Wachter, and
Heisz, 2012).
27
We compute HML returns from the year an individual entered the
labor market for the first time. This increases the sample size for the HML
factor portfolio because we have data starting only in 1975. Our results
are also robust to using data for the HML factor portfolio for Europe (e.g.,
28
Fama and French, 2012) for which we have data starting only in 1990 (not Our results are similar if we include only the years of negative GDP
tabulated). The value premium in Swedish stocks has been shown to be growth, but such a measure is more susceptible to criticisms of exogene-
around 8% (e.g., Fama and French, 1998; Hansson, 2004). ity compared with a measure that also includes 7 1 year.

Please cite this article as: Cronqvist, H., et al., Value versus growth investing: Why do different investors have different
styles? Journal of Financial Economics (2015), http://dx.doi.org/10.1016/j.jfineco.2015.04.006i
 H. Cronqvist et al. / Journal of Financial Economics ] (]]]]) ]]]–]]] 13

Table 7
Life course theory and investment style: impressionable years.
This table reports coefficient estimates from panel regressions with individual characteristics and year fixed effects. Robust standard errors are two-way
clustered by twin pair and year. The variables are defined in Appendix A. nnn, nn, and n indicate significance at the 1%, 5%, and 10% level, respectively.

Variable (1) (2) (3) (4)

Panel A: Years of first labor market entry
First Labor Market Entry in Recession
0.87
0.59
First Labor Market Entry in Severe Recession
3.24nn
1.37
Individual characteristics Yes Yes
Year fixed effects Yes Yes
N 95,272 95,272
R-squared 0.19 0.19
Panel B: Impressionable years
18–25 Years Old in Severe Recession
3.07nnn
2.90nnn
1.11 0.99
10–17 Years Old in Severe Recession
0.65n
0.79nn
0.34 0.38
26–33 Years Old in Severe Recession
1.26
1.24
0.98 1.07

Individual characteristics Yes Yes Yes Yes

Year fixed effects Yes Yes Yes Yes
N 105,393 105,393 105,393 105,393
R-squared 0.19 0.19 0.19 0.19

Finally, in Panel B we find an equally strong effect if Table 8

we examine whether an individual experienced a Life course theory and investment style: rearing environment.
This table reports coefficient estimates from panel regressions with
severe economic recession during the impressionable
individual characteristics, year fixed effects, and birth cohort (by decade)
years (e.g., Giuliano and Spilimbergo, 2014). We also fixed effects. Robust standard errors are two-way clustered by twin pair
compare the effects of the economic conditions during and year. The variables are defined in Appendix A. nn indicates signifi-
the 18–25 years with experiences somewhat earlier cance at the 5% level.
(10–17 years) and somewhat later (26–33) in life. We
Variable (1) (2)
find that the effect of severe recession experiences
during the 10–17 year period is of smaller economic Log (Parents' Net Worth) nn
0.21 0.22nn
magnitude (0:79=2:90 ¼ 27:2%) compared with the 0.10 0.10
effect for the impressionable years. A statistical test Individual characteristics included Yes Yes
Year fixed effects Yes Yes
reveals that the difference between the impressionable
Birth cohort (decade) fixed effects No Yes
years effect and the 10–17 years effect is statistically N 22,484 22,484
significant. We conclude that the economic conditions R-squared 0.15 0.15
during the years of an individual's first labor market
entry and the impressionable years are important for
the individual's investment style later in life.
parental wealth distribution (10th percentile) have portfolios
4.3.3. Rearing environment with average P/E ratios that are 0.92 (or about 6.2% at the
We also investigate whether the rearing environment has median) lower compared with investors at the top of the
significant and persistent effects on an individual's invest- distribution (90th percentile). We show that this effect is
ment style later in life. We examine the socioeconomic status robust within each generation by controlling for birth cohort
of an investor's parents. We are not able to measure parents' (decade) fixed effects, so this result is not specific to, e.g., the
SES precisely when an individual grew up, so we use parents' Great Depression cohort.
net worth at the start of our data set as a proxy measure.29 Overall, we find support for the hypothesis that life exp-
The results are reported in Table 8. We find that individuals eriences affect an individual's investment style. By control-
who grew up in a lower SES environment, i.e., relatively poor, ling for education, income, and net worth, we can rule out
show significantly more value orientation in their stock that these effects operate merely through an investor's
portfolios later in their lives. Investors at the bottom of the economic circumstances. Instead, our evidence is consistent
with experiences early in life affecting investors' preferences
and beliefs. If shared by a sufficiently large part of the pop-
29
To avoid that our results are affected by parents who were alive in ulation, such experiences could result in legacy effects that
1999, but passed away during our sample period, and consequently affect affect the relative demand for value versus growth stocks
the portfolio composition of their children through wealth shocks or
direct inheritances of equity assets (e.g., Andersen and Nielsen, 2011), we
and thereby possibly the value premium, similar to the imp-
exclude individuals whose parents passed away during our sample lications of Friedman and Schwartz (1963) and Cogley and
period. Sargent (2008) for the equity premium.

Please cite this article as: Cronqvist, H., et al., Value versus growth investing: Why do different investors have different
styles? Journal of Financial Economics (2015), http://dx.doi.org/10.1016/j.jfineco.2015.04.006i
14 H. Cronqvist et al. / Journal of Financial Economics ] (]]]]) ]]]–]]]

5. Conclusion

We find that several factors explain an individual investor's

A.2. Individual characteristics
style, i.e., the value versus growth orientation of the investor's
stock portfolio. First, we estimate that genetic differences acr-
oss individuals explain a significant proportion of the variation
in value versus growth orientation, whether we analyze Male Indicator that is one if an individual is male and
zero otherwise
P/E ratios as an investment-style measure, or Morningstar's College or More Indicator that is one if an individual has
Value-Growth Score. Such a biological basis of individuals' attended college and zero otherwise
investment style likely reflects partially innate preferences as High School Indicator that is one if an individual has
well as genetically influenced characteristics of investors' eco- completed high school (gymnasium) and zero
otherwise
nomic circumstances that affect portfolio choice.
No Education Data Indicator variable that equals one if no
Second, we examine in detail which individual character- Available educational data are available for an individual
istics explain investment style. We find that investors' hedging and zero otherwise
demands related to human capital and displacement risk as Married Indicator that is one if an individual is married
well as behavioral biases contribute to investment style. and zero otherwise
Net Worth Difference between end-of-year market values
Investors with more human capital and whose labor income of an individual's assets and liabilities (in
is more correlated with GDP growth hold more growth stocks. nominal Swedish krona)
We also find that investors who exhibit more behavioral Disposable An individual's disposable income (in nominal
biases in the investment domain in particular in the form of a Income Swedish krona), i.e., sum of income from labor,
business, and investment, plus received
preference for speculative assets are more growth oriented.
transfers, less taxes, and alimony payments
Finally, we find that an investor's style is explained by life Share in Equities Market value of direct and indirect equity
course theory in that experiences, both earlier and later in life, holdings divided by market value of all
are related to investment style. Investors with adverse macro- financial assets
economic experiences have stronger preferences for value Financial Wealth The market value of an individual's financial
assets as reported by Statistics Sweden at the
investing later in life, even when differences in income and end of each year, expressed in nominal Swedish
net worth are accounted for. For example, those who grew up Krona. Financial assets include checking,
during the Great Depression have portfolios with average P/E savings, and money market accounts, (direct
ratios that are significantly lower several decades later in life. and indirect) bond holdings, (direct and
indirect) equity holdings, investments in
Consistent with an impressionable years hypothesis, those
options and other financial assets such as
who enter the labor market for the very first time during a rights, convertibles, and warrants
severe economic recession are also more value-oriented later Labor Income An individual's work-related income (in
on. This evidence contributes to a growing literature in finance nominal Swedish krona) in a given year
and economics that shows the importance of life experiences Labor Income Time series correlation of an individual's log of
Correlation labor income growth and GDP growth in
and events for economic behavior later in life.
Sweden. GDP growth data are from the World
Our results have several implications for understanding Development Indicators (WDI)
the value premium. They suggest that the value premium Investment Bias For each of six investment behavior, the
could reflect both risk-based compensation and mispricing Index variable is zero (least biased), one, or two (most
biased). The behaviors are diversification,
due to investors' behavioral biases. But our findings also
home bias, turnover, disposition effect,
imply that the overall composition of investor population, performance chasing, and skewness
with respect to genetic make-up, age, and life experiences, preference. The index is the sum across the
can affect the relative demand for value versus growth investment behaviors
stocks and in the end potentially the value premium. Volatility Using twelve monthly return observations for
each asset in an individual's equity portfolio,
we calculate the average annualized return
Appendix A. Variable definitions volatility for each individual's portfolio and
year
A.1. Investment style measures

A.3. Life-cycle experience and events

P/E Ratio Value-weighted price to earnings ratio for
an individual's direct stock holdings. Data
are from CapitalIQ
P/B Ratio Value-weighted market to book value of Depression Baby Indicator that is one if an individual is
equity ratio for an individual's direct stock born 1920–1929, and zero otherwise
portfolio. Data are from Capital IQ Individual GDP Growth Average GDP growth in Sweden from an
Morningstar's Value- Value-weighted Morningstar score of value Experience individual's birth year to 2000
Growth Score versus growth from
100 (value) to þ 400 Individual Sweden HML Average Fama and French Sweden HML
(growth). Data are from Morningstar Experience return from the year an individual
Name-based Value/ Value-weighted indicator that is
1 if a entered the labor market for the first
Growth Measure fund's name contains “value,” þ1 if a fund's time (if after 1975) to 2000. Data are
name contains “growth” or “technology,” from Kenneth French
and zero otherwise. Data are from Professional Finance Using data on an individual's
Morningstar Experience occupation, based on the International

Please cite this article as: Cronqvist, H., et al., Value versus growth investing: Why do different investors have different
styles? Journal of Financial Economics (2015), http://dx.doi.org/10.1016/j.jfineco.2015.04.006i
 H. Cronqvist et al. / Journal of Financial Economics ] (]]]]) ]]]–]]] 15

Standard Classification of Occupations σ2 is the sum of the three variance components (σ 2 ¼

(ISCO-88) by the International Labour
Organization (ILO) and available for a
σ 2a þ σ 2c þ σ 2e ).
subset of our sample, we identify Identification of variation due to aij, ci, and eij is possible
individuals with work experience due to constraints on the covariance matrices for these eff-
related to finance ects. These constraints are the result of the genetic similarity
First Labor Market Entry Indicator that is one if an individual of twins and assumptions about upbringing and other aspects
in Recession entered the labor market for the first
time during a year with negative GDP
of the common environment. Consider two twin pairs i¼1,2
growth7 1 year and zero otherwise with twins j¼ 1,2 in each pair, where the first is a pair of
First Labor Market Entry Indicator that is one if an individual identical twins and the second is a pair of fraternal twins. The
in Severe Recession entered the labor market for the first additive genetic effects are a ¼ ða11 ; a12 ; a21 ; a22 Þ0 . Identical
time during World War I, the Great
and fraternal twin pairs differ in their genetic similarity, i.e.,
Depression, or World War II and zero
otherwise the off-diagonal elements related to identical twins in the
Parents' Net Worth Difference between market values of matrix in Eq. (2) are 1 as the proportion of shared additive
combined assets and liabilities (in genetic variation is 100% between identical twins. In contrast,
nominal Swedish krona) of an for fraternal twins the proportion of the shared additive
individual's parents at the end of 1999
genetic variation is on average only 50%,32 i.e., the off-dia-
gonal elements related to fraternal twins in the matrix in
Eq. (2) are 1/2.33 As a result, for these two twin pairs, the
covariance matrix with respect to aij is
2 3
Appendix B 1 1 0 0
6 7
61 1 0 0 7
In this Appendix, we describe the empirical methodol- CovðaÞ ¼ σ 2a 6
60 0
7: ð2Þ
4 1 1=2 7
5
ogy employed to decompose the cross-sectional variation
0 0 1=2 1
in individual investors' investment styles into genetic and
environmental components. We model the value versus The common environmental effects are c ¼ ðc11 ; c12 ; c21 ;
growth orientation, vgij, for twin pair i and twin j (1 or 2) c22 Þ0 . The model assumes that identical and fraternal twins
as a function of observable socioeconomic individual char- experience the same degree of similarity in their common
acteristics Xij and three unobservable random effects, an environments (the equal environments assumption). That
additive genetic effect, aij, an effect of the environment is, the off-diagonal elements related to either identical or
common to both twins (e.g., upbringing), ci, and an indi- fraternal twins in the matrix in Eq. (3) are 1. Assuming that
vidual-specific effect, eij, which also absorbs idiosyncratic identical and fraternal twins experience the same degree
measurement error: of similarity in their common environment, any excess
similarity between identical twins is due to the greater
vg ij ¼ β 0 þ β1 Xij þaij þci þ eij : ð1Þ proportion of genes shared by identical twins than by
fraternal twins. As a result, for the two twin pairs, the cov-
In quantitative behavioral genetics research, this model ariance matrix with respect to ci is
2 3
is referred to as an ACE model, where A stands for additive 1 1 0 0
genetic effects, C for common environment, and E for ind- 61 1 0 07
6 7
ividual-specific environment.30 The additive genetic com- CovðcÞ ¼ σ 2c 6 7: ð3Þ
40 0 1 15
ponent aij in Eq. (1) represents the sum of the genotypic
0 0 1 1
values of all genes that influence an individual's behavior.
Each individual has two, potentially different, versions The individual-specific environmental effects are e ¼
(alleles) of each gene (one is from each parent), and each ðe11 ; e12 ; e21 ; e22 Þ0 . These error terms represent, for exam-
version is assumed to have a specific, additive effect on the ple, life experiences, but also idiosyncratic measurement
individual's behavior. The genotypic value of a gene is the error. That is, the off-diagonal elements related to either
sum of the effects of both alleles present in a given individual. identical or fraternal twins in the matrix in Eq. (4) are 0. As
Consider, for example, two different alleles A1 and A2 for a
given gene and assume that the effect of the A1 allele on 32
Genome sequencing has recently revealed that humans and, e.g.,
investment style is of magnitude α1, and the effect of the A2 the common chimpanzee (Pan troglodytes) share about 96% of their genes
allele is α2. An individual with genotype A1A1 would (e.g., Mikkelsen, 2005), and the genetic overlap is even greater among
experience the genetic effect 2α1 , and genotype A1A2 would humans. That is, the 50% refers to only the proportion of genes that
have a genetic effect of α1 þ α2 .31 We also assume that aij, ci, makes different humans different from each other.
33
For an intuitive explanation of the proportion of the shared
and eij are uncorrelated with one another and across twin
additive genetic variation for fraternal twins as well as non-twin siblings,
pairs and normally distributed with zero means and var- consider a single gene, of which one parent has allele A1 and A2, while
iances σ2a, σ2c , and σ2e , so that the total residual variance the other parent has allele A3 and A4. Any of their off-spring will have
one of the following combinations as they get one allele from each
parent: A1A3, A1A4, A2A3, or A2A4. Suppose one fraternal twin is of type
30
See, e.g., Falconer and Mackay (1996) for a more detailed discus- A1A3. The overlap with the fraternal twin sibling will be 1 if the sibling is
sion of quantitative behavioral genetics research. of type A1A3, 1/2 if type A1A4, 1/2 if type A2A3, and 0 if the type is A2A4.
31
The extent to which the effect of two different alleles deviates This implies an average overlap of 1/2. For a formal derivation, see e.g.,
from the sum of their individual effects is called dominance deviation. Falconer and Mackay (1996).

Please cite this article as: Cronqvist, H., et al., Value versus growth investing: Why do different investors have different
styles? Journal of Financial Economics (2015), http://dx.doi.org/10.1016/j.jfineco.2015.04.006i
16 H. Cronqvist et al. / Journal of Financial Economics ] (]]]]) ]]]–]]]

a result, for the two twin pairs, the covariance matrix with Brunnermeier, M.K., Nagel, S., 2008. Do wealth fluctuations generate
respect to eij is time-varying risk aversion? Micro-evidence on individuals' asset
allocation. American Economic Review 98, 713–736.
2 3 Campbell, J.Y., Vuolteenaho, T., 2004. Bad beta, good beta. American
1 0 0 0
Economic Review 94, 1249–1275.
60 1 0 07
6 7 Carlen, J., 2012. Einstein of Money: The Life and Timeless Financial
CovðeÞ ¼ σ 2e 6 7: ð4Þ
40 0 1 05 Wisdom of Benjamin Graham. Prometheus Books, Amherst, NY.
Cesarini, D., Dawes, C.T., Johannesson, M., Lichtenstein, P., Wallace, B.,
0 0 0 1 2009. Genetic variation in preferences for giving and risk taking.
Quarterly Journal of Economics 124, 809–842.
We use maximum likelihood to estimate the model in Cesarini, D., Johannesson, M., Magnusson, P.K., Wallace, B., 2012. The
behavioral genetics of behavioral anomalies. Management Science 58,
Eq. (1). Finally, we calculate the variance components A, C,
21–34.
and E. A is the proportion of the total residual variance that Chan, L.K.C., Hamao, Y., Lakonishok, J., 1991. Fundamentals and stock
is related to an additive genetic factor: returns in Japan. Journal of Finance 46, 1739–1764.
Chan, L.K.C., Lakonishok, J., 2004. Value and growth investing: review and
σ 2a σ 2a update. Financial Analysts Journal, 71–86.
A¼ ¼ ð5Þ Chen, M.K., Lakshminarayanan, V., Santos, L.R., 2006. How basic are
σ 2 σ 2a þ σ 2c þ σ 2e behavioral biases? Evidence from Capuchin monkey trading behavior.
Journal of Political Economy 114, 517–537.
The proportions attributable to the common environment Chetty, R., Friedman, J.N., Hilger, N., Saez, E., Schanzenbach, D.W., Yagan,
(C) and individual-specific environmental effects (E) are D., 2011. How does your kindergarten classroom affect your earn-
ings? Evidence from Project STAR. Quarterly Journal of Economics
computed analogously. Standard errors reported in the
126, 1593–1660.
tables are bootstrapped with one thousand repetitions. Cochrane, J.H., 2007. Portfolio theory. Unpublished working paper. Uni-
versity of Chicago, Chicago, IL.
Cogley, T., Sargent, T.J., 2008. The market price of risk and the equity
References
premium: a legacy of the Great Depression?. Journal of Monetary
Economics 55, 454–476.
Addoum, J.M., Korniotis, G.M., Kumar, A., 2014. Income risk, dynamic Constantinides, G.M., Donaldson, J.B., Mehra, R., 2002. Junior can't
style preferences, and return predictability. Unpublished working borrow: a new perspective on the equity premium puzzle. Quarterly
paper. University of Miami, Coral Gables, FL. Journal of Economics 117, 269–296.
Alesina, A., Fuchs-Schüendeln, N., 2005. Good bye Lenin or not? the Cosmides, L., Tooby, J., 1994. Better than rational: evolutionary psychol-
effect of Communism on people's preferences American Economic ogy and the invisible hand. American Economic Review 84, 327–332.
Review 97, 1507–1528. Cronqvist, H., Previtero, A., Siegel, S., White, R.E., 2014. The fetal origins
Almond, D., Currie, J., 2011. Killing me softly: the fetal origins hypoth- hypothesis is finance: prenatal testosterone and financial risk taking.
esis. Journal of Economic Perspectives 25, 153–172. Unpublished working paper. Ivey School Business School, London,
Andersen, S., Nielsen, K.M., 2011. Participation constraints in the stock Ontario, Canada.
market: evidence from unexpected inheritance due to sudden death. Cronqvist, H., Siegel, S., 2014. The genetics of investment biases. Journal of
Review of Financial Studies 24, 1667–1697. Financial Economics 113, 215–234.
Asness, C.S., Frazzini, A., Pedersen, L.H., 2012. Leverage aversion and risk Croson, R., Gneezy, U., 2009. Gender differences in preferences. Journal of
parity. Financial Analysts Journal 68, 47–59. Economic Literature 47, 1–27.
Asness, C.S., Moskowitz, T.J., Pedersen, L.H., 2013. Value and momentum Cunha, F., Heckman, J., 2010. Investing in our young people. In: Reynolds,
everywhere. Journal of Finance 68, 929–985. A. (Ed.), Cost Effective Early Childhood Programs in the First Decade,
Bansal, R., Yaron, A., 2004. Risks for the long run: a potential resolution Cambridge University Press, New York, NY.
of asset pricing puzzles. Journal of Finance 59, 1481–1509. Currie, J., 2011. Inequality at birth: some causes and consequences.
Barberis, N., Huang, M., 2008. Stocks as lotteries: the implications of American Economic Review 101, 1–22.
probability weighting for security prices. American Economic Review Daniel, K., Hirshleifer, D., Subrahmanyam, A., 1998. Investor psychology
98, 2066–2100. and security market under-and overreactions. Journal of Finance 53,
Barberis, N., Shleifer, A., 2003. Style investing. Journal of Financial 1839–1885.
Economics 68, 161–199. Daniel, K., Titman, S., 1997. Evidence on the characteristics of cross-
Barberis, N., Shleifer, A., Vishny, R.W., 1998. A model of investor senti-
sectional variation in stock returns. Journal of Finance 52, 1–33.
ment. Journal of Financial Economics 49, 307–343.
Daniel, K., Titman, S., Wei, K., 2001. Explaining the cross-section of stock
Barnea, A., Cronqvist, H., Siegel, S., 2010. Nature or nurture: what
returns in Japan: factors or characteristics?. Journal of Finance 56,
determines investor behavior?. Journal of Financial Economics 98,
743–766.
583–604.
Davis, S.J., Willen, P., 2013. Occupation-level income shocks and asset
Ben-Porath, V., 1973. Short-term fluctuations in fertility and economic
returns: their covariance and implications for portfolio choice.
activity in Israel. Demography 10, 185–204.
Quarterly Journal of Finance 3, 1350011.
Betermier, S., Calvet, L.E., Sodini, P., 2014. Who are the value and growth
De Bondt, W.F.M., Thaler, R.H., 1985. Does the stock market overreact?
investors? Unpublished working paper. McGill University, Montreal,
Canada. Journal of Finance 40, 793–805.
De Long, J.B., Shleifer, A., Summers, L.H., Waldmann, R.J., 1990. Positive
Bisin, A., Verdier, T., 2000. Beyond the melting pot: cultural transmission,
marriage, and the evolution of ethnic and religious traits. Quarterly feedback investment strategies and destabilizing rational speculation.
Journal of Economics 115, 955–988. Journal of Finance 45, 379–395.
Bisin, A., Verdier, T., 2001. The economics of cultural transmission and the Dehejia, R., Lleras-Muney, A., 2004. Booms, busts, and babies' health.
dynamics of preferences. Journal of Economic Theory 97, 298–319. Quarterly Journal of Economics 119, 1091–1130.
Black, S.E., Devereux, P.J., Salvanes, K.G., 2005. The more the merrier? The Dorn, D., Huberman, G., 2010. Preferred risk habitat of individual
effect of family size and birth order on children's education. Quarterly investors. Journal of Financial Economics 97, 155–173.
Journal of Economics 120, 669–700. Dreber, A., Apicella, C.L., Eisenberg, D.T.A., Garcia, J.R., Zamore, R.S., 2009.
Bonaparte, Y., Korniotis, G.M., Kumar, A., 2014. Income hedging and The 7R polymorphism in the dopamine receptor D4 gene (DRD4) is
portfolio decisions. Journal of Financial Economics 113, 300–324. associated with financial risk taking in men. Evolution and Human
Bouchard, T.J., 1998. Genetic and environmental influences on adult Behavior 30, 85–92.
intelligence and special mental abilities. Human Biology 70, Elder, G.H., 1974. Children of the Great Depression: Social Change in Life
257–279. Experience. University of Chicago Press, Chicago, IL.
Bouchard, T.J., Lykken, D.T., McGue, M., Segal, N.L., Tellegen, A., 1990. Elder, G.H., Johnson, M.K., Crosnoe, R., 2003. The Emergence and Devel-
Sources of human psychological differences: the Minnesota study of opment of Life Course Theory. Kluwer Academic Publishers, Norwell,
twins reared apart. Science 250, 223–228. MA pp. 3–19.
Brennan, T.J., Lo, A.W., 2011. The origin of behavior. Quarterly Journal of Falconer, D.S., Mackay, T.F.C., 1996. Introduction to Quantitative Genetics,
Finance 1, 55–108. fourth ed. Pearson Education Limited, Essex, England.

Please cite this article as: Cronqvist, H., et al., Value versus growth investing: Why do different investors have different
styles? Journal of Financial Economics (2015), http://dx.doi.org/10.1016/j.jfineco.2015.04.006i
 H. Cronqvist et al. / Journal of Financial Economics ] (]]]]) ]]]–]]] 17

Fama, E.F., 1996. Multifactor portfolio efficiency and multifactor asset Liew, J., Vassalou, M., 2000. Can book-to-market, size, and momentum be
pricing. Journal of Financial and Quantitative Analysis 31, 441–465. risk factors that predict economic growth? Journal of Financial
Fama, E.F., French, K.R., 1992. The cross-section of expected stock returns. Economics 57, 221–245.
Journal of Finance 47, 427–465. Lynch, A.W., 2001. Portfolio choice and equity characteristics: character-
Fama, E.F., French, K.R., 1993. Common risk factors in the returns on izing the hedging demands induced by return predictability. Journal
stocks and bonds. Journal of Financial Economics 33, 3–56. of Financial Economics 62, 67–130.
Fama, E.F., French, K.R., 1996. Multifactor explanations of asset pricing Malmendier, U., Nagel, S., 2011. Depression babies: do macroeconomic
anomalies. Journal of Finance 51, 55–84. experiences affect risk-taking?. Quarterly Journal of Economics 126,
Fama, E.F., French, K.R., 1998. Value versus growth: the international 373–416.
evidence. Journal of Finance 53, 1975–1999. Malmendier, U., Nagel, S., 2013. Learning from inflation experiences.
Fama, E.F., French, K.R., 2012. Size, value, and momentum in international Unpublished working paper. National Bureau of Economic Research,
stock returns. Journal of Financial Economics 105, 457–472. Cambridge, MA.
Frazzini, A., Pedersen, L.H., 2014. Betting against beta. Journal of Financial Malmendier, U., Tate, G., Yan, J., 2011. Overconfidence and early life
Economics 111, 1–25. experiences: the effect of managerial traits on corporate financial
Friedman, M., Schwartz, A.J., 1963. A Monetary History of the United policies. Journal of Finance 66, 1687–1733.
States, 1867–1960. Princeton University Press, Princeton, NJ. Massa, M., Simonov, A., 2006. Hedging, familiarity, and portfolio choice.
Gârleanu, N., Kogan, L., Panageas, S., 2012. Displacement risk and asset Review of Financial Studies 19, 633–685.
returns. Journal of Financial Economics 105, 491–510. Merton, R.C., 1969. Lifetime portfolio selection under uncertainty: the
Giele, J.Z., Elder, G.H., 1998. Methods of Life Course Research: Qualitative continuous-time case. Review of Economics and Statistic 51, 247–257.
and Quantitative Approaches. Sage Publications, Thousand Oaks, CA. Mikkelsen, T.S., 2005. Initial sequence of the chimpanzee genome and
Gilbert, T., Hrdlicka, C., Kalodimos, J., Siegel, S., 2014. Daily data is bad for comparison with the human genome. Nature 437, 69–87.
beta: opacity and frequency-dependent betas. Review of Asset Pricing Neale, M.C., Maes, H.H.M., 2004. Methodology for genetic studies of twins
Studies 4, 78–117. and families. In: NATO ASI Series D: Behavioural and Social Science.
Giuliano, P., Spilimbergo, A., 2014. Growing up in a recession. Review of Kluwer Academic Publishers B.V., Dordrecht, The Netherlands.
Economic Studies 81, 787–817. Oreopoulos, P., von Wachter, T., Heisz, A., 2012. The short-and long-term
Graham, J.R., Narasimhan, K., 2004. Corporate survival and managerial career effects of graduating in a recession. American Economic
experiences during the Great Depression. Unpublished working Journal: Applied Economics 4, 1–29.
paper. Duke University, Fuqua School of Business, Durham, NC. Oyer, P., 2008. The making of an investment banker: stock market shocks,
Greenwood, R., Nagel, S., 2009. Inexperienced investors and bubbles.
career choice, and lifetime income. Journal of Finance 63, 2601–2628.
Journal of Financial Economics 93, 239–258.
Palacios-Huerta, I., 2003. The robustness of the conditional CAPM with
Grinblatt, M., Keloharju, M., 2009. Sensation seeking, overconfidence, and
human capital. Journal of Financial Econometrics 1, 272–289.
trading activity. Journal of Finance 64, 549–578.
Petkova, R., Zhang, L., 2005. Is value riskier than growth? Journal of
Han, B., Kumar, A., 2013. Speculative retail trading and asset prices.
Financial Economics 78, 187–202.
Journal of Financial and Quantitative Analysis 48, 377–404.
Rayo, L., Becker, G.S., 2007. Evolutionary efficiency and happiness. Journal
Hansson, B., 2004. Human capital and stock returns: is the value
of Political Economy 115, 302–337.
premium an approximation for return on human capital? Journal of
Robson, A.J., 2001. Why would nature give individuals utility functions?
Business Finance and Accounting 31, 333–358.
Journal of Political Economy 109, 900–914.
Heaton, J., Lucas, D., 2000. Portfolio choice and asset prices: the
Ruhm, C.J., 2000. Are recessions good for your health? Quarterly Journal
importance of entrepreneurial risk. Journal of Finance 55,
of Economics 115, 617–650.
1163–1198.
Rutter, M., 2006. Genes and Behavior: Nature/Nurture Interplay
Hong, H., Stein, J.C., 1999. A unified theory of underreaction, momentum
Explained. Blackwell Publishers, Oxford, UK.
trading, and overreaction in assets markets. Journal of Finance 54,
Samuelson, P.A., 1969. Lifetime portfolio selection by dynamic stochastic
2143–2184.
programming. Review of Economics and Statistics 51, 239–246.
Jagannathan, R., Wang, Z., 1996. The conditional CAPM and the cross-
Santos, L.R., Chen, M.K., 2009. The evolution of rational and irrational
section of expected returns. Journal of Finance 51, 3–53.
economic behavior: evidence and insight from a non-human primate
Jurek, J.W., Viceira, L.M., 2011. Optimal value and growth tilts in long-
horizon portfolios. Review of Finance 15, 29–74. species. In: Glimcher, P., Camerer, C., Fehr, E., Poldrack, R. (Eds.),
Kahn, L.B., 2010. The long-term labor market consequences of graduating Neuroeconomics: Decision-Making and the Brain, Elsevier, London,
from college in a bad economy. Labour Economics 17, 303–316. UK, pp. 81–94.
Kaustia, M., Knüpfer, S., 2008. Do investors overweight personal experi- Santos, T., Veronesi, P., 2006. Labor income and predictable stock returns.
ence? Evidence from IPO subscriptions. Journal of Finance 63, Review of Financial Studies 19, 1–44.
2679–2702. Schoar, A., Zuo, L., 2013. Shaped by booms and busts: how the economy
Krosnick, J.A., Alwin, D.F., 1989. Aging and susceptibility to attitude impacts CEO careers and management styles. Unpublished working
change. Journal of Personality and Social Psychology 57, 416. paper. MIT Sloan School of Management, Cambridge, MA.
Kuhnen, C.M., Chiao, J., 2009. Genetic determinants of financial risk Shefrin, H., Statman, M., 2000. Behavioral portfolio theory. Journal of
taking. PLoS ONE 4. Financial and Quantitative Analysis 35, 127–151.
Kumar, A., 2009a. Dynamic style preferences of individual investors and Spear, D.O., 2000. Neurobehavioral changes in adolescence. Current
stock returns. Journal of Financial and Quantitative Analysis 44, Directions in Psychological Science 9, 111–114.
607–640. Sundén, A.E., Surette, B.J., 1998. Gender differences in the allocation of
Kumar, A., 2009b. Who gambles in the stock market? Journal of Finance assets in retirement savings plans. American Economic Review 88,
64, 1889–1933. 207–211.
Lakonishok, J., Shleifer, A., Vishny, R.W., 1994. Contrarian investment, Sylvain, S., 2014. Does human capital risk explain the value premium
extrapolation, and risk. Journal of Finance 49, 1541–1578. puzzle? Unpublished working paper. University of Chicago, Depart-
Larsen, L.S., Munk, C., 2012. The costs of suboptimal dynamic asset ment of Economics, Chicago, IL.
allocation: general results and applications to interest rate risk, stock Thompson, S.B., 2011. Simple formulas for standard errors that cluster by
volatility risk, and growth/value tilts. Journal of Economic Dynamics both firm and time. Journal of Financial Economics 99, 1–10.
and Control 36, 266–293. Vissing-Jørgensen, A., 2002. Towards an explanation of household port-
Lewellen, J., Nagel, S., 2006. The conditional CAPM does not explain asset folio choice heterogeneity: nonfinancial income and participation
pricing anomalies. Journal of Financial Economics 82, 289–314. cost structures. Unpublished working paper. National Bureau of
Lichtenstein, P., Sullivan, P.F., Cnattingius, S., Gatz, M., Johansson, S., Economic Research, Cambridge, MA.
Carlström, E., Björk, C., Svartengren, M., Wolk, A., Klareskog, L., de Zhang, L., 2005. The value premium. Journal of Finance 60, 67–103.
Faire, U., Schalling, M., Palmgren, J., Pedersen, N.L., 2006. The Swedish Zhong, S., Israel, S., Xue, H., Ebstein, R.P., Chew, S.H., 2009. Monoamine
twin registry in the third millennium: an update. Twin Research and Oxidase A gene (MAOA) associated with attitude towards longshot
Human Genetics 9, 875–882. risks. PLoS ONE 4.

Please cite this article as: Cronqvist, H., et al., Value versus growth investing: Why do different investors have different
styles? Journal of Financial Economics (2015), http://dx.doi.org/10.1016/j.jfineco.2015.04.006i