Margin of Safety: Life History Strategies and the Effects of Socioeconomic Status and Macroeconomic Conditions on Self-Selection into Accounting

Electronic copy available at: https://ssrn.com/abstract=2985135

Margin of Safety: Life History Strategies and the Effects

of Socioeconomic Status and Macroeconomic Conditions

on Self-Selection into Accounting

Justin Leiby*

University of Georgia

Paul E. Madsen

University of Florida

June 2017

*Corresponding Author 310 Herty Drive

255 Brooks Hall Athens, GA 30606 706.542.3596 jleiby@uga.edu

This paper has benefited from helpful comments from Kris Allee, Allen Blay, Bud Fennema, Karla Johnstone, Matt Kaufman, Brian Mayhew, Michelle McAllister, Chad Stefaniak (discussant) and workshop participants at Florida State University, the University of Wisconsin – Madison, Temple University, and the 2015 AAA Auditing Midyear Meeting.

Electronic copy available at: https://ssrn.com/abstract=2985135 1

Margin of Safety: Life History Strategies and the Effects

of Socioeconomic Status and Macroeconomic Conditions

on Self-Selection into Accounting

Abstract: We use experimental and archival evidence to show that people who had low socioeconomic status (SES) as children participate in the U.S. accounting labor market in distinctive and consequential ways. Drawing on life history theory, we predict and show that low SES individuals select into accounting at disproportionately high rates relative to other fields, an effect driven by accounting’s relatively high job security. Supplemental tests are consistent with these low SES individuals being a source of high quality human capital for the accounting profession, as low SES individuals selecting into accounting possess desirable attributes at relatively high rates. From a social perspective, we provide theory and evidence consistent with accounting being an important and secure source of upward social mobility in comparison to other fields. However, recessions cause selection into accounting by low SES individuals to decrease at a higher rate than in other fields, compromising these professional and social benefits. For example, our evidence is consistent with the “low SES effect” improving gender diversity among entrants into the accounting labor market during good economic times. However, lower self-selection rates during recessions are particularly pronounced among low SES females, who may thus bear the brunt of lost professional and social benefits.

Keywords: Self-selection; Human Capital; Accounting Labor Market; Socioeconomic Status, Gender Diversity; Inequality

Electronic copy available at: https://ssrn.com/abstract=2985135 2

1. Introduction

In this study, we predict and show that people who had low childhood socioeconomic status (SES)1 participate in the U.S. accounting labor market in distinctive and consequential ways. Accounting careers are relatively secure in that they reward relatively high human capital investment with stable demand and returns (Bureau of Labor Statistics 2015; Madsen 2015). We draw on life history theory, which concerns the relationship between childhood poverty and adulthood preferences for security and deferred rewards (Chisholm 1993; Griskevicius et al. 2013), to predict that low SES individuals 1) disproportionately select into accounting relative to other fields, 2) possess attributes desired by the accounting profession, and 3) select into accounting at disproportionately lower rates than other fields during times of macroeconomic uncertainty. Using experimental and archival evidence, we find support for all three of these expectations.

Our theory and findings have important implications for society and for the accounting profession. The United States has low socioeconomic mobility relative to other OECD countries (OECD 2010; Blanden 2013; Mitnik et al. 2015). Even public universities, which graduate many entry-level accountants, are marred by inequality (Chetty et al. 2017). From a social perspective, our evidence suggests that selecting into accounting is an effective “ladder” out of poverty. Relative to other fields, accounting delivers high wages, low wage variability, and high job security, regardless of SES background, which facilitates upward mobility.

Our evidence further suggests that the accounting profession faces unique risks and opportunities in the intensifying competition for talent (ACAP 2008; AICPA 2013). Despite extensive research about accounting’s outputs (e.g., reported earnings), little theory or data exists

1 _{Low SES is poor relative access to financial, educational, and social capital resources (Coleman 1988). Our} examination focuses on the effects of childhood SES, not a person’s current SES as an adult. Throughout the paper, “SES” refers to childhood SES unless otherwise noted.

3

about accounting’s human capital inputs. Understanding entry-level labor supply is critical for accounting due to its heavy reliance on young employees and early recruitment strategies (The Economist 2014). Our theory and evidence suggest that this requires a nuanced understanding of the preferences of low SES individuals. Driven by preferences for security and long-term human capital returns, large numbers of talented low SES individuals select into accounting, which benefits the profession. However, this effect reverses during recessions, during which desirable low SES individuals disproportionately select away from accounting. Together, these findings suggest an opportunity for the accounting profession to attract desirable individuals by emphasizing the security and equal opportunity offered by accounting careers.

We test our research questions with multiple methods. For a controlled test of the underlying cognitive mechanism, we first conduct an experiment that measures upper-division accounting students’ intentions to enter the accounting profession, as opposed to common, less secure alternatives such as finance. Our choice of participant group biases against observing variation in intentions to enter accounting, as most participants have nearly completed accounting degrees and have already had one or more internships with accounting firms. Prior to measuring self-selection intentions, we actively manipulate cues of uncertain macroeconomic conditions by priming half the participants with a narrative about the effects of recessions, while the other half read a neutral prime (Griskevicius et al. 2011b). Following a neutral prime, intentions to self-select into accounting are greater among low, as opposed to high SES individuals. Moreover, the recession prime decreases intentions to self-select into accounting among low SES individuals, but increases these intentions among high SES individuals. Consistent with our theorized mechanism, the self-assessed importance of job security in a profession mediates the joint effect of SES and macroeconomic uncertainty on intentions to self-select.

4

We next test our research questions using large-sample archival data to demonstrate that our theorized effects are broadly generalizable. We use data from the Higher Education Research Institute’s (hereafter, “HERI") surveys on the demographics, degree choices, and families of millions of college freshmen in the U.S conducted over several decades. In some supplemental tests, we use a subset of HERI data describing college seniors. Consistent with our experimental results, accounting disproportionately attracts low SES individuals relative to non-accounting business fields and to all non-accounting fields.2 This effect holds among college seniors, as well. We also test this effect in recession years, as opposed to non-recession years. Consistent with our predictions, selection into accounting by low SES individuals decreases in recession years, and decreases more in accounting than in non-accounting business fields and all non-accounting fields.

Moreover, we conduct multivariate analyses that control for attributes desirable to the profession, such as diverse demographics. We also gather input from a panel of partners and managers at multiple public accounting firms about attributes most likely to distinguish successful accountants, such as academic performance, motivation, and communication skills. Our findings not only are robust to controlling for these attributes, but also show that low (as opposed to high) SES individuals selecting accounting possess some of these attributes at higher rates. Low (as opposed to high) SES individuals who select accounting are similar in academic ability and motivation but exhibit greater gender diversity, self-confidence, and writing ability. This is consistent with our theorized effect channeling quality human capital into the profession. This finding is encouraging for social mobility, as other professional fields such as finance and law systematically exclude low SES individuals (Rivera 2015).

2 _{In our} comparisons, “non_-accounting business fields” refers to _{business administration, finance, international} business, marketing, management, and other business fields, whereas “all non-accounting fields” refers to the set of all fields other than accounting. “Other fields” refers to both of the aforementioned comparison groups.

5

However, our findings also show that recessions decrease low SES self-selection into accounting more than into other fields. Thus, uncertain economic conditions undermine accounting’s function as a means of upward mobility for disadvantaged people and undermine an effect that otherwise channels high quality talent into the profession. In particular, our evidence suggests that the low SES effect may be particularly effective in improving gender diversity by attracting females from disadvantaged backgrounds to accounting. However, the recession effect also disproportionately decreases selection into accounting by low SES females. If low SES individuals benefit from selecting into accounting, then low SES females disproportionately miss these benefits during uncertain macroeconomic times. While the low SES effect facilitates upward social mobility, there may be a “gender gap” in its effects.

We next compare career outcomes in accounting to those in alternative fields. This analysis can support our theory that accounting is a secure choice, and tests whether or not accounting both delivers security and does so similarly for low and high SES individuals. We examine mean wages, wage variance, and unemployment in accounting versus other substitute fields using data from the National Survey of College Graduates (NSCG) by the National Science Foundation. The data reveal that accounting degrees deliver high mean salaries with low variance, and unemployment lower than in non-accounting business fields and all non-accounting fields. That is, if people select accounting because they expect security, then accounting on average fulfills these expectations. Moreover, the benefits of accounting are more pronounced among low SES individuals, especially in comparison to finance, which is the most common alternative. This suggests that self-selection into accounting is a secure choice that can have long-term welfare benefits for low SES individuals.

In sum, our evidence shows that accounting occupies a distinctive niche in the menu of career options available to labor market entrants, in part because potential labor market entrants

6

perceive it to be exceptionally safe. Accounting provides a relatively secure entry into a business career for talented individuals from low SES backgrounds, which is encouraging in light of evidence of persistent, implicit discrimination against low SES individuals in other professions (Rivera 2015). Our evidence also suggests that accounting’s distinctive features benefit the profession by attracting low SES people with attributes desired by the profession. Given the profession’s (and accounting firms’) investments in broadening accounting’s appeal and personalizing recruitment to best compete for talent (Jeacle 2008; Carnegie and Napier 2010), it is inherently important to better understand the vector of attributes that influence a person’s interest in accounting. Childhood SES is among these attributes.

We also contribute to broader theory on career selection and life history theory. We provide the first empirical evidence of which we are aware that people are likely to have different life history strategies that influence how they view their labor market options. Seminal economic thinkers such as Alchian (1950) and Becker (1976) concurred that insights from evolutionary biology enrich our understanding of economic choices. As Becker (1976, 818) observes, “the approach of sociobiologists is highly congenial to economists, since they rely on competition, the allocation of limited resources…efficient adaptation to the environment, and other concepts also used by economists.” Career selection is part of the adaptive landscape of contemporary life, and is thus well suited to interpretation through a life history theory lens. Our study illuminates the cost / benefit tradeoffs that people make when they choose a profession by showing how and why early life experiences shape the preferences that drive adult career choices.

2. Background Literature and Theory Development

The accounting profession’s sustainability depends on its capacity to attract human capital in an increasingly competitive labor market (ACAP 2008; AICPA 2013). It is thus important to

7

better understand why people choose to become accountants. Prior literature provides limited insight into this choice, mostly from exploratory studies of the personality profiles of accounting students. The evidence suggests that accounting degree programs are dominated by a small subset of personality types favoring concrete and analytical thinking (Oswick and Barber 1998; Wheeler 2001; Swain and Olsen 2012). In a longitudinal study, Swain and Olsen (2012) find that these personality types disproportionately begin careers in accounting and remain in accounting jobs, suggesting that accounting “fits” certain personalities. Also, relative to other fields, accounting students are more conscientious and interested in making money, yet less creative and less enthusiastic about their chosen field (Saemann and Crooker 1999; Allen 2004; Madsen 2015). Blay and Fennema (2017) find that some college students exhibit inherent aptitude for accounting, but this aptitude does not lead to self-selection into accounting. This suggests that self-selection may reflect a more complex set of considerations than “personality fit” or inherent skills.

Indeed, we argue that the choice to become an accountant reflects a complex set of cost-benefit considerations involving broad social and economic forces.3 _{In particular, identifying the} forces that drive self-selection promotes a better understanding not only of the profession’s appeal to labor market entrants, but also of how this appeal may change in different environments and potentially of the profession’s broader societal function.

We focus on the impact of a person’s childhood SES on selection into accounting. SES is important in itself due to its links to social welfare, as identifying professions that represent good matches for low SES individuals can increase equality and upward socioeconomic movement. SES

3 _{If self-selection into accounting versus other fields were to follow a traditional model, then it is possible that there} would be no meaningful variation in self-selection beyond the vector of attributes affecting the enjoyment of accounting and the marginal costs of acquiring accounting skills (Willis and Rosen 1979; Guasch and Weiss 1981; Polachek 1981). However, choosing to pursue accounting does not imply aptitude for accounting (Oswick and Barber 1998; Blay and Fennema 2017) and does not imply enjoyment of accounting (Madsen 2015). Moreover, marginal costs of entry into a field, such as earning an accounting degree, can be high independent of aptitude, as is the case when SES is low (McDonough 1997). Thus, a different lens is likely useful to understanding the selection decision.

8

also links to the diversity of human capital, which professions pursue but have achieved with mixed success (Abbott 1988; Hammond 2002; Sullivan Commission 2004; Madsen 2013). In general, low SES individuals confront social and economic barriers to professional success, as low SES is associated with lower college attendance and higher college dropout rates (DeAngelo et al. 2011; Chetty et al. 2017). Rivera (2015) reports that low SES college graduates are less likely than their high SES counterparts to get jobs in elite law, banking, and consulting firms, even when they graduate from the same institutions and have better grades. In addition to these barriers, low SES is associated with lower emotional resilience to stress, further diminishing the likelihood of success in higher education and in difficult jobs (Bowles et al. 2005). In one study of entry-level accountants, low SES individuals perform as well as their high SES counterparts but exit at higher rates (Collarelli et al. 1987). The authors offer no explanations for this finding, but it does raise the possibility that low SES individuals are an underutilized source of quality human capital.

Moreover, this study focuses on SES because a person’s SES likely influences the appeal of the accounting profession’s relative security, specifically its tradeoff between costs of entry and relative security. Accounting requires significant investment in human capital, such as Bachelor’s and sometimes Master’s degrees, certifications, and ongoing education. In turn, the accounting labor market exhibits persistently low unemployment, stable demand that is robust to economic conditions, and low wage variance (AICPA 2013; Bureau of Labor Statistics 2015).

We argue that SES affects self-selection into accounting by influencing the compatibility between these fundamental attributes of accounting and a person’s broader life history strategy. In the subsections that follow, we discuss life history strategies and the effect of SES on the composition of the pool of potential entrants into accounting. Because a necessary condition of entry is the costly choice to invest resources of time, money, and effort in education, we first

9

examine the choice to attend college. As far as we are aware, no existing studies offer theoretical explanations as to how this costly choice reflects a set of deeper fundamental life history differences, particularly in low SES individuals. Our theory allows us to better define the potential labor market entrants over whom accounting and other fields compete. Among this group of potential entrants, we then examine the choice of accounting over other options.

2.1. Life History Theory and College Attendance

The idea of life history strategies originated in evolutionary biology. This theory seeks to explain how organisms (including humans) trade off current versus future resource consumption at different points in their lifespans (Schaffer 1983; Kaplan and Gangestad 2005). While early applications of life history theory focused on tradeoffs to increase survival odds or reproductive fitness, over the past two decades this theory has been applied more broadly to understand issues in the social sciences, including in economics, marketing, psychology, and sociology. For example, Wang and Dvorak (2010) draw on the theory’s biological roots to examine temporal discounting, i.e., preferring smaller immediate payoffs over larger future payoffs. The authors find that temporal discounting decreases in response to an experimental manipulation in which half of participants consume a soft drink prior to the task—that is, temporarily increasing blood glucose levels and the subjective sense that daily energy needs are fulfilled affects a common economic tendency. This is one of many examples in which life history theory and its biological foundations are useful to understanding contemporary psychological and economic issues.

For our study, career selection is well suited for examination as part of a life history strategy, because selection is an action generally taken at a given point in the lifespan (i.e., early adulthood) and involves complex tradeoffs affecting current and future resource acquisition and allocation. For example, career selection involves tradeoffs between current versus future resource

10

usage at key points throughout life—such as pursuing post-secondary education versus working full-time immediately after high school.

Life history strategies vary along a fast/slow continuum (Ellis et al. 2009). Slower (as opposed to faster) strategies are consistent with longer over shorter horizons, prioritizing saving over consumption, avoiding over accepting risks, among other distinctions (e.g., Kaplan and Gangestad 2005; Griskevicius et al. 2013).4 The trajectory towards faster or slower life history strategies begins early in life, with harsher, more uncertain childhood environments—such as those characterized by low SES or exposure to violence or disease—associated with a trajectory towards faster strategies (Wilson and Daly 1997; Low et al. 2008; Brumbach et al. 2009; Ellis et al. 2009).5 For example, people from harsher childhood environments follow a faster life history trajectory in adolescence and adulthood with riskier and earlier sexual activity, higher rates of smoking, and lower impulse control (Seltzer and Oechsli 1985; Wilson and Daly 1997; Soteriades and DiFranza 2003; Hanson and Chen 2007; Nettle 2010; Hill et al. 2016). These effects extend to economic decision making, as there are conditions in which people who grew up in harsher environments have higher rates of credit card debt, are less willing to purchase insurance, and make riskier investment choices, indicating faster life history strategies (Griskevicius et al. 2011b; 2013; Mittal and Griskevicius 2016).

Life history theory allows us to develop nuanced predictions about the effect of SES on selection into accounting. We begin by discussing the effect of low SES on the pool of potential

4 _{Biologists such as Stearns (1989) originally characterized the fundamental tradeoff in life history strategies as} somatic effort towards physical and mental growth versus reproductive effort towards attracting and retaining mates and caring for children. Through broader application of the theory to other fields, researchers have characterized the tradeoffs more broadly to include things like risk and time horizon. Griskevicius et al. (2011a) analogize the tradeoff to putting money into a savings account versus spending the money to help perpetuate the account holder’s survival. 5 _{The term} “strategy” does not imply that _{a set of actions are good or bad, or even that the person consciously} undertakes each action constituting a life history strategy. Some early life tradeoffs relevant to our study are determined by parents or social institutions, rather than choices of the individual, such as low SES limiting access to early childhood education (McDonough 1997).

11

entrants into accounting, i.e., college attendees. Because lower SES implies a disadvantaged position and thus an initially faster trajectory, having the option to pursue higher education suggests that a low SES individual has deviated from that trajectory through a series of “slower” choices as an adolescent. People from poorer backgrounds who have the option to attend college likely made “slow” tradeoffs to invest in education, delay gratification, and avoid risks in adolescence in order to overcome their disadvantaged initial position.

For example, Chen and Miller (2012) document a “shift and persist” life history strategy among some low SES individuals, indicated by high resilience, optimism, and pride in one’s achievements. Such resilience at younger ages is increasing in intelligence and self-esteem, and helps low SES individuals manage stress and adapt to social circumstances at older ages (Masten et al. 1990; Brody et al. 2013). This is consistent with evidence that low SES is positively

associated with pursuing a college education among those who exhibit resilience, e.g., by holding jobs and demonstrating financial responsibility during adolescence (Brumbach et al. 2009).

Thus, we argue that, among college attendees, the distribution of life history strategies is likely narrower when SES is low, because the initial advantages of high SES offer more wiggle room for risky or myopic choices, yet retain opportunities such as higher education. Consequently, college attendees are likely to have followed slower life history strategies than have non-attendees, and this difference is likely to be significantly greater when SES is low. This enriches existing literature, which generally observes that higher SES and educational attainment individually are associated with positive health and life outcomes, but does not examine interactive effects.

We predict that the opportunity to choose to go to college likely signals that a person from a poorer socioeconomic background has had a slower life history strategy in the past, as indicated by choices during adolescence and adulthood that are consistent with a slower strategy. It is also

12

likely that pursuing college is a more credible signal of life history strategies among people from poorer, as opposed to wealthier backgrounds. This leads to our first hypothesis,

Hypothesis 1: College attendees’ choices are more consistent with slower life history strategies than those of non-attendees, and this difference is greater for lower, as opposed to higher SES individuals.

2.2. Life History Theory and Self-Selecting into Accounting over Other Options

Among fields requiring a college education, choosing accounting over other options represents a relatively slow strategy. Jobs in the accounting profession often require incremental human capital investments, in the form of a master’s degree, certification (e.g., CPA, CMA), and continuing professional education. In turn, accounting offers incremental short- and long-term security in the form of low unemployment, stable demand, and salaries that deliver comparable returns to other business fields (Bureau of Labor Statistics 2015; Madsen 2015).

Based on the development of hypothesis one, the set of potential entrants into the accounting labor market comprises low SES individuals with predominately slow life history strategies, in addition to high SES individuals who have a broader array of strategies. Thus, all else equal, we predict that low SES is likely to increase the likelihood of self-selection of college attendees into accounting—that is, low SES individuals will be disproportionately represented in accounting relative to the set of all other fields. This leads to our second hypothesis:

Hypothesis 2: Low SES individuals select into accounting to a greater degree than non-accounting business fields and all non-non-accounting fields.

2.3. The Moderating Effect of Recessions

The effect of SES on self-selection into accounting is likely to differ across good and bad economic conditions. Although life history strategies develop early, they are not fixed and their trajectories can change in response to cues in a person’s current environment (Griskevicius et al. 2011a; b; 2013). That is, people re-calibrate towards slower or faster life history strategies in

13

response to indicators of potential changes in resource availability or life expectancy (e.g., recessions, crime rates).

Childhood SES affects how people re-calibrate preferences based on environmental cues, and does so throughout childhood and adulthood (Ellis et al. 2009; Griskevicius et al. 2011b). Among adults, lower childhood SES is associated with lower perceived control over one’s environment (Chen and Miller 2012; Mittal and Griskevicius 2014) and lower impulse control capabilities (Kochanska et al. 2001; Hill et al. 2016). Also, low childhood SES is associated with poor health outcomes among adults, and this association is driven by childhood SES but not adulthood SES (Currie and Stabile 2003; Cohen et al. 2004; Hanson and Chen 2007). Hackman et al. (2010) discuss evidence of biological and physiological differences between people who grew up with low, as opposed to high SES. This includes fMRI evidence of different brain structures for areas involved with problem-solving and threat detection.

In other words, people are physically and psychologically sensitized to the conditions that they observe during childhood. If a person observes adverse conditions like resource scarcity while making a key decision, then the person’s choice is likely to re-calibrate towards the trajectory adopted in childhood, i.e., when facing adverse conditions, poorer (wealthier) backgrounds lead to choices consistent with faster (slower) strategies.

Thus, we expect an interactive effect of SES and macroeconomic conditions on self-selection into accounting. Although a person from a poorer background may have altered trajectories during adolescence towards a slower strategy and investment in education, the trajectory of subsequent choices such as field of study and career are likely to vary depending on

14

whether they make the choice in benign, as opposed to uncertain economic conditions.6 There is ample evidence to support this logic. In a series of experiments, Griskevicius et al. (2013) find that people who grew up in poorer environments exhibit greater risk-taking and temporal discounting when they observe cues of resource scarcity, as opposed to neutral cues. By contrast, these behaviors decrease when childhood SES is high. Similarly, observing cues of current environmental uncertainty decreases low SES individuals’ impulse control, willingness to delay consumption, and willingness to purchase insurance, but has opposite effects among high SES individuals (Griskevicius et al. 2011b; 2013; Mittal and Griskevicius 2014; 2016).

Applied to our setting, uncertain macroeconomic conditions like recessions are likely to weaken preferences for the relatively slow attributes of accounting among those low SES individuals who have already entered college and are, therefore, in the set of potential entrants into accounting. That is, while low SES is likely to increase the preference for accounting in benign economic conditions, it is likely to decrease this preference in uncertain macroeconomic conditions. Thus, we predict an interaction in which the effect of SES on self-selection into accounting depends on whether or not there is a recession, i.e., conditions of resource scarcity.

Hypothesis 3: Selection into accounting among low SES individuals decreases in uncertain, as opposed to benign macroeconomic conditions. This effect is stronger in accounting than in non-accounting business fields and all non-accounting fields.

6 _{Life history strategy comprises a broad array of social, psychological, and economic considerations such as risk} aversion, temporal discounting, impulsivity, sexual activity, parental duties, etc. We have considered purely economic explanations for our predictions, but cannot conceive of a comprehensive and parsimonious explanation. Theory and evidence on the relation between wealth and, for example, risk aversion is difficult to apply to our setting, as it primarily focuses on portfolio settings (Arrow 1971; Paravisini et al. 2010). Absolute risk aversion is decreasing in wealth (i.e., incremental wealth increases the absolute level of investment in risky assets) but relative risk aversion is increasing in wealth (i.e., incremental wealth decreases the percentage of total wealth that is invested in risky assets). This would not explain H2 or H3. More broadly, as a theoretical example, it is possible that, if risk aversion is a concave function of wealth, then all else equal, low SES individuals occupy a steeper part of the curve than do high SES individuals and are thus more risk averse. This would increase the preference for accounting. However, this would not explain H3, as recessions would simply shift both high SES and low SES individuals down the curve and make each group more risk averse, increasing the preference for accounting among both groups.

15

3. Tests of H1—Life History Strategies of College Attendees versus Non-Attendees 3.1. Data

Data for testing H1 are from the National Longitudinal Survey of Youth (NLSY).7 The NLSY is a survey of roughly 10,000 randomly-selected Americans born between 1957 and 1964. It begins in 1979 and has been updated 24 times through 2012 to track participants’ ongoing labor market activities and other significant life events. This dataset captures educational and career activities, as well as self-reported lifestyle attributes and choices. It is an ideal dataset for testing H1, because it allows linkage between indicators of early childhood SES, adolescent and adult lifestyle attributes, and educational and career choices.

3.2. Variables

College attendee is our proxy for a person’s choice to pursue a college degree. We define a college attendee as a respondent in the NLSY who had completed at least four years of college by the age of 26.8 To proxy for low SES, we identified respondents whose parents had not attended any college when the survey began in 1979. Parental education level is a widely-accepted proxy of SES, as it is positively associated with income and negatively associated with student loan debt and full-time work as a source of tuition assistance (DeAngelo et al. 2011).

To find proxies for life history strategy speed, we search for variables that are both available during the appropriate time in participants’ lives (late adolescence and early adulthood when people also make career decisions) and relevant to our theory. We find five variables from the NLSY that are available in relevant years and that theory suggests will be sensitive to the speed

7 _{See http://www.bls.gov/nls/nlsy79.htm and Appendix B for more information about the NLSY. We use the NLSY} in tests of H1 but not in other tests, because the NLSY does not measure the degrees pursued by respondents attending college and graduating from college.

8 _{We repeated our tests with college attendee defined as anyone who completed one year of college by the age of 22,} to proxy some college attendance as opposed to attendance and completion. We choose to emphasize the standard of completing four years by age 26, as it seems more relevant to our labor market self-selection theory to focus on those who complete college, but our inferences are the same using either proxy.

16

of a life history strategy. We use average age at first marriage, with older ages signifying a slower strategy (Chisholm 1993, 9); average age at birth of first child, with older ages signifying a slower strategy (Griskevicius et al. 2011b, Nettle 2010); average number of children, with fewer children signifying a slower strategy (Nettle 2010); whether the respondent has smoked daily at any time in their life, with not smoking signifying a slower strategy (Petridou et al. 1997; Hill and Chow 2002); and the age at which the respondent began smoking daily, with older ages signifying a slower strategy (Petridou et al. 1997; Hill and Chow 2002).

3.3. Results

We use a difference in difference analysis to test H1. Using the variables above, we partition the sample into four groups: 1) college non-attendees who are low SES, 2) college attendees who are low SES, 3) college non-attendees who are high SES, and 4) college attendees

who are high SES. H1 is supported by larger differences between college attendee and non-attendee differences when socioeconomic status is low than when it is high, i.e., the difference between groups (1) and (2) is greater than the difference between groups (3) and (4).

Table 1 shows that all five slower life history strategy proxies are consistent with low SES college attendees being significantly more likely than low SES college non-attendees to have made choices consistent with a slow life history strategy (all p ≤ 0.02).9 Further, there are five difference in differences in Table 1 to test whether the difference between college attendees and non-attendees is greater for low SES than for high SES. The difference between college attendees and

college non-attendees is significantly greater for low SES respondents than for high SES

respondents on four of the five measures (p ≤ 0.09). The result for smoked daily is directionally consistent with our predictions but not statistically significant (p = 0.13).

17

Our results support H1 that college attendees from low SES backgrounds have especially slow life history strategies. Further, among college attendees, our results are consistent with low SES individuals exhibiting even slower strategies than high SES individuals on average number of children (p = 0.03) and smoked daily (0.07). Thus, the pool of low SES people minimally eligible to select into accounting are predominately those with slow life history strategies. This point is critical in examining determinants of entry into accounting, as we theorize that distinguishing features of accounting likely make it appealing to people pursuing slower strategies.

4. Experimental Tests of H2 and H3 – SES and Self-Selection into Accounting

We test H2 and H3 using multiple methods due to the paucity of data on self-selection and the complex array of factors that could influence this decision. To maximize the internal validity of our inferences, we first test these hypotheses with an experiment using upper-division accounting students. In particular, we use priming procedures to actively manipulate cues of macroeconomic uncertainty (i.e., recession versus neutral) and randomly assign participants to observe recession versus neutral primes. Because we cannot actively manipulate SES, the experiment only allows us to make associational claims about H2. However, randomly assigning participants to different primes does allow us to make strong causal inferences about H3, specifically, the effects of the recession prime on career intentions within a given level of SES (e.g., how do low SES individuals respond to recession, as opposed to neutral primes?).

4.1. Participants and Procedures

Experimental participants (n = 245) are business students recruited from accounting classes at a large public university in the Southeast United States, of which 209 (85.3%) were accounting majors. The sample comprises 51 (20.8%) freshmen and sophomores, 87 (35.5%) juniors, 47 (19.1%) seniors, and 59 (24.1%) Master’s students. Participants begin the experiment by reading

18

a 600-word story that contained our priming manipulation (described below). To ensure that participants attended to the prime and did not believe that the prime was related to their career intention judgment, the instructions directed participants to assume that the prime was “like a memory task” and that there would be questions about the story at the end of the study. After reading the story, participants completed a series of questions about their career intentions, answered manipulation and attention check questions, and answered questions about their childhood and current SES. To avoid deception, we included memory questions about the story to follow through on the statement in the materials that the story was like a memory task.

4.2. Variables

4.2.1. Recession prime

In the recession prime condition, participants read a story about economic hardships confronted by recent college graduates. The story—titled “Tough Times Ahead: The New Economics of the 21stCentury”—was formatted to look like a news article and was adopted from Griskevicius et al. (2011b and 2013). In brief, the story focused on economic hardships confronted by recent college graduates, including skyrocketing student loan debt, intense labor market competition, increasing food and energy costs, and diminishing funds for government social support programs. In the control condition, the story (also adopted from Griskevicius et al. 2011b and 2013) was designed to elicit similar levels of negative arousal. 10 The story focused on a person spending several hours looking for keys around their house. We could have used a neutral prime that described good economic conditions, but that would have risked manipulating negative arousal along with macroeconomic cues.

10 _{A prime unrelated to the job selection task is desirable for internal validity, as varying some aspect of the task would} risk confounding cues of economic uncertainty with uncertainty about career objectives, task enjoyment, etc. We are grateful to Vlad Griskevicius for graciously sharing his materials.

19

As manipulation checks, we asked participants in each condition to rate the extent to which the story made them believe the world will become (1) more unsafe, (2) more unpredictable, and (3) more uncertain. Participants’ responses in the recession condition were higher than those in the control condition on all three measures (all p < 0.01), but did not differ on measures of negative arousal such as anger, frustration, or stress that could influence the career intention judgment. This indicates a successful manipulation that varied the uncertainty of the environment, but held constant levels of negative arousal that could influence our measures. See Appendix A.

4.2.2. SES

To capture SES, we asked participants to report whether or not they are a first generation college student. We classified participants as low SES if they reported having no parent(s) or guardian(s) with a college degree and as high SES if they reported at least one parent or guardian with a college degree.11

Self-Selection into Accounting

We measured our primary dependent variable on a 100-point scale that asked participants to assess the likelihood that they intend to pursue a career in accounting. The instrument also asked participants to rate the likelihood with which they intend to pursue careers in close substitute fields such as finance, investment banking, or consulting. In addition, participants assessed the importance of five characteristics of a job / career: job security, high earnings, likeable colleagues, interesting work, and maximizing future opportunities.

11 _{Results using two alternative measures of childhood SES do not change our inferences. We also measured SES} based on (1) estimated childhood household income and (2) average agreement with three statements capturing childhood SES. For (2), the statements were “My family usually had enough money for things when I was growing up,” “I grew up in a relatively wealthy neighborhood,” and “I felt relatively wealthy compared to the other kids in my school.” Our inferences are identical using these measures. We also included measures of current SES to ensure that childhood SES and not current SES is the mechanism that influences decision making. When we control for current SES, our results do not change.

20 4.3. Results

4.3.1. Tests of H2 and H3

H2 predicts that low SES individuals are more likely to select into accounting. H3 predicts that uncertain macroeconomic conditions, of which recessions are a common contemporary indicator, will decrease self-selection into accounting by people from poorer backgrounds. To test this, we conduct a 2 (Prime: recession, control) X 2 (Childhood SES: low, high) Analysis of Variance (ANOVA) with the assessed likelihood of pursuing a career in accounting as the dependent measure. See Table 2 for ANOVA results and descriptive statistics.

Consistent with H2, in the neutral prime condition, intentions to select into accounting are higher when SES is low, as opposed to high (78.00 versus 63.62, F1,241 = 8.02, p < 0.01). Further, the interaction predicted by H3 is significant (F1,241 = 8.42, p < 0.01). Specifically, observing the recession prime decreased the likelihood of pursuing a career in accounting when SES is low (63.95 versus 78.00, F1,241 = 5.75, p = 0.02). Interestingly, though we had no hypothesis for how the prime would affect high SES individuals, Table 2, Panel C shows that observing the recession prime marginally increased the likelihood of pursuing a career in accounting in this group (70.57 versus 63.63, F1,241 = 7.29, p = 0.10).12 Thus, H2 and H3 are supported, and the data are consistent with our hypothesized interaction being a causal effect.

4.3.2. Supplemental Test of the Mediating Effect of Job Security

To provide further corroboration of the cognitive process that underlies our effects, we examine participants’ ratings of the importance of job security to their career decision. Our theory

12 _{Our theory focuses on low SES individuals and thus only predicts that the recession prime will more negatively} affect selection intentions among low, as opposed to high SES individuals. However, the observed disordinal interaction is consistent with the logic of our theory. If recession cues cause individuals to re-calibrate choices in the direction consistent with their upbringing, then it is possible for high SES individuals, as well as low SES individuals, to reverse their preferences.

21

argues that H2 and H3 occur because low SES college students follow slow strategies in benign conditions but adopt faster strategies in uncertain conditions. If our theory is valid, then SES and recession jointly affect a person’s likelihood of self-selection into accounting because they jointly affect the importance of job security to that person. That is, our interaction is likely to have an indirect effect on self-selection, via differences in the importance of job security.

Following the recent statistics literature, we test the significance of the indirect effect using a bootstrapping technique (Preacher and Hayes 2008; Hayes and Preacher 2013). Because our independent variable is an interaction term, we follow the guidance of Hayes and Preacher (2013) to create a multi-categorical independent variable using the linear weights for the interaction term.13 We use 5000 bootstrap re-samples of the data to calculate bias-corrected confidence intervals for the total indirect effect. Significance is indicated by confidence intervals that do not include zero. In our analysis, there is a significant indirect effect of life history strategy on self-selection into accounting, which is mediated by the assessed importance of job security (lower CI = 0.75, higher CI = 9.86). See Figure 1. Thus, our data support the conclusion that the SES by recession prime interaction affects self-selection into accounting because it affects the importance of job security.

5. Archival Tests of H2 and H3 – SES and Self-Selection into Accounting

We now turn to tests of H2 and H3 using large-sample archival evidence to provide comfort about the generalizability of our findings.

13 _{We used the following weights: -0.25 for control prime, high SES; +0.25 for recession prime, high SES; +0.25 for} control prime, low SES; -0.25 for recession prime, high SES. Following Hayes and Preacher’s (2013) guidance, we include two orthogonal control contrasts because fully representing the effects of a categorical variable with k categories requires k – 1 parameter estimates.

22 5.1. Sample

We obtain data from the Higher Education Research Institute’s (HERI) Freshman Surveys, which contain data collected from millions of incoming college freshmen annually since 1971 (UCLA 2013). The surveys describe how students choose which college to attend as well as their demographics, high school academic and extracurricular activities, opinions on a wide array of topics, values, and the field in which they intend to earn a degree. HERI data from 1971 to 1999 are available to all registered users of HERI’s website in an archive file. More recent data are available for purchase contingent on approval from HERI. In this study, we use subsets of HERI data from the archive as well as data purchased for 2000 and 2002.14 _{We conduct our main tests} using subsets of years from 1971-2002 for which our variables are available.15 The sample size in our main tests is 125,125 accounting observations, 407,235 non-accounting business observations, and 2,993,954 non-accounting observations.16

In addition, to test the robustness of H2 and to perform certain supplemental analyses (described later), we also obtain a subset of the HERI database that collects responses both when individuals were freshman and when they were seniors. This dataset is only available from 1994 – 1999, thus tests using the senior data have lower statistical power than tests using the freshman dataset. Also, no recessions occurred during 1994 – 1999, thus we cannot use it to test H3.

14 _{The 2000 and 2002 samples differ somewhat from the earlier samples because HERI was unwilling to release the} entire database to us. Rather, for 2000 and 2002 HERI provided us with all observations of students selecting any business field and a random sample of observations representing students selecting non-business fields.

15 _{See http://www.heri.ucla.edu/abtcirp.php and Appendix B for more information about HERI variables and data} availability.

16 _{Like many surveys, HERI is constructed using stratified sampling rather than strictly random sampling in order to} ensure adequate coverage of the populations of interest to the study’s designers. Because their sampling is not strictly random, traditional statistical methods applied to HERI data will produce biased estimates if they are not adjusted to correct for HERI’s sampling methodology. Statistical corrections for nonrandom sampling are common in survey data analysis and involve weighting observations using their probability sampling weight (e.g., Rosenbaum 1987; King et al. 2010). All of the means and statistical tests we report are corrected to account for survey weights.

23 5.2. Variables

5.2.1. Self-selection into accounting

To test self-selection into accounting, we partition our sample into students planning to major in accounting and those not planning to major in accounting. This is a reliable predictor of selection into accounting jobs (Madsen 2015). We use two comparison groups to test our hypothesis. We test self-selection into accounting against “non-accounting business” fields (business administration, finance, international business, marketing, management, and other business) and against “all non-accounting” fields (the set of all non-accounting fields).

5.2.2. SES

We manually search the HERI codebook and identify three measures of childhood SES that are both available throughout the whole sample period and relevant indicators of low SES. They are 1) whether or not the respondent indicated that cheap tuition was “very important” to

them when selecting a college, as prior research has shown that low SES students are particularly sensitive to tuition costs (Heller 1997, 638-642); 2) whether or not the respondent is a first generation college student, as first generation college students come from poorer families on average (Terenzini et al. 1996, 8-9), and whether or not the respondent’s estimate of their parental income is in the bottom third of the sample for that year, which is a direct measure of respondents’ perceptions about relative household income.17 For each of these measures, we create a dummy variable equal to 1 for values representing low SES and 0 otherwise.18 _{We then sum these three}

17 _{Our results are robust to alternative specifications, such as the bottom quartile or the bottom half of the distributions.} 18 _{Most ofthe HERI variables are ordinal measures with verbal anchors on each scale point, and the measures vary in} the number of scale points and in the verbal anchors used. For example, parental income has 14 points, each associated with an income range (e.g., 1 = “Less than $6,000,” 14 = “$200,000 or more”), while chea p tuition has three (1 = “Not important,” 2 = “Somewhat important,” 3 = “Very important”). There is no consensus in the literature about how best to treat this type of ordinal variable statistically. Baum (2006, 161) and Campbell (2008) argue that dichotomization is justified in large samples such as ours, but that treating variables as continuous is also allowable. In our main tests, we transform ordinal variables into dichotomous variables around their medians unless otherwise noted in the text. We run robustness tests treating all ordinal predictors as continuous and our inferences do not change.

24

dummy variables to create a low SES index. Higher values of the low SES index indicate lower SES. Our main tests involve comparing values of the low SES index for students who have selected accounting degrees against values for students selecting other degrees.

Table 3, Panel A sorts all fields in the HERI freshman database based on our low SES index in recession and non-recession years, with higher values indicating higher representation of low SES individuals. Accounting stands out relative to other fields and to business fields in particular, ranking 11th overall out of 76 fields in low SES. These results suggest that accounting is by far the most appealing business field to low SES individuals, versus management (35th), business administration (39th_{), marketing (69}th_{), and finance (70}th_{). Table 3, Panel B shows that accounting} has the 12th largest decline in the low SES index in recession years. All other business fields experience smaller declines in the low SES index during recession years, as no other business field ranks higher than 38th in the magnitude of the decline (Finance). Indeed, accounting appears to occupy a unique niche among business fields and among most fields in general.19

5.2.3. Uncertain Economic Conditions

Our proxy for uncertain economic conditions is a year in which a recession occurred. We partition the sample period into recession years and non-recession years. Recession is an indicator for any year that includes at least one month classified by the National Bureau of Economic Research as a recession month. Recession years during our sample period are 1973-1975, 1980-1982, and 1990-1991.

19 _{Of the fields with comparable low SES representation to accounting, only agriculture and nursing have} similarly-sized declines in low SES representation in recession years. Thus, these fields may fit similar life history strategies to those fit by accounting.

25 5.3. Results

5.3.1. Hypothesis 2 – Univariate Tests of the Effect of SES on self-selection into accounting

Table 4, Panel A presents univariate analyses. It provides mean values for each of our low SES indicators, both separately and when aggregated, in accounting, other business fields, and all non-accounting fields. H2predicts that a low SES background is positively associated with self-selection into accounting. Consistent with H2, accounting students have higher values than non-accounting students for each of our low SES indicators. These differences are even larger when comparing accounting to other business fields. Table 4, Panel A also presents the univariate results among college seniors. Because people from disadvantaged backgrounds are more likely to drop out of college before they become seniors, all three low SES indicators are lower in the senior dataset than in the freshman dataset. The senior dataset also has substantially smaller sample sizes and lower statistical power. Notwithstanding these limitations that bias against our findings, H2 is robust among college seniors. Thus, our univariate analyses support H2.

5.3.2. Hypothesis 2 – Multivariate Tests of the Effect of SES on self-selection into accounting

Table 4, Panel B presents results of a logit analysis that predicts the likelihood of selecting into accounting based on SES, which we estimate separately during the 1970s, 1980s, 1990s, and 2000s (only the years 2000 and 2002 due to data limitations) to illustrate the effect’s robustness over time. We estimate the following model:

Select Accounting (0 or 1) = α + β1*lowSES + β2*Female + β3*Minority + β4*GPA + β5*Drive + β6*Self-confidence intelligence + β7*Self-confidence social + β8*Writing + ε (1)

Our primary variable of interest is lowSES and we predict a positive effect for this variable, i.e., an odds ratio greater than one. We report coefficients in the form of odds ratios, which are exponentiated logit coefficients, to provide a clearer interpretation of the practical implications of

26

the effects of a given variable. In brief, odds ratios represent the number of people in our sample with a given attribute choosing accounting for every one person choosing a business degree (under the heading “freshmen: accounting versus non-acctg business students”) or any other degree (under the heading “freshmen: accounting versus all non-acctg students”). Table 4, Panel B also shows similarly calculated results for college seniors (under the “seniors” heading).

To select control variables, we collected input from seven senior managers and partners in public accounting firms about desired attributes that lead to success in an accounting career. We presented each person with a list of 16 items from the HERI survey that we expected could be related to success and asked each person to select the five that are most likely to lead to success in accounting. We control for variables that were selected by a majority of respondents. Specifically, we include GPA on a four-point scaleas a proxy for academic ability (high school GPA for college freshmen and college GPA for college seniors), an indicator variable for above average self-assessed drive as a proxy for motivation, indicator variables for above average self-assessed intellectual and social self-confidence, and an indicator variable for above average self-assessed

writing ability as a proxy for writing skills. We also include indicator variables for Female and for membership in a disadvantaged minority race or ethnicity, which we label Minority and includes Black, Hispanic, and American Indian people. We include Female and Minority to provide insight into the relation between our effect and the diversity of the accounting labor pool. See Appendix B for details about the variables.20

The results in Table 4, Panel B support the idea that low childhood SESsignificantly and positively predicts the likelihood of selecting into accounting in each of the sample periods. Thus,

20 _{The 16 variables in our questionnaire were academic ability (i.e. GPA), amount of time spent studying, amount of} participation in clubs or student organizations, computer skills, competitiveness, creativity, drive to achieve, initiative, leadership ability, mathematical ability, public speaking ability, self-confidence, self-understanding, understanding of others, volunteer work, and writing ability.

27

multivariate analyses support H2. These analyses show that our effect is also robust to controlling for race, gender, academic ability, self-confidence, motivation, and writing ability. The results also indicate a few potentially problematic patterns for accounting. While GPA is positively associated with accounting in every sample period and relative to both business and non-business fields, individuals selecting accounting appear to possess lower social self-confidence and writing ability. We explore these issues in Section 6.1 to examine whether or not the low SES effect may help overcome some deficits in the entry-level accounting labor pool. Life history theory suggests this is possible, as those adopting slower life history strategies tend to be highly capable. Because low SES individuals have different life experiences than their high SES counterparts, they may also possess high levels of different capabilities and thus improve the quality of the entry-level accounting labor pool.21

5.3.3. Hypothesis 3 – SES in recessions and non-recessions

The HERI data also permit us to provide large-sample analysis of H3, which predicts that recessions decrease preferences for accounting when SES is low, and do so to a greater degree in accounting than in other fields. Our univariate analysis of H3 is presented in Table 5, Panel A.

21 _{We conduct untabulated tests to ensure that our results are robust to inclusion of all 16 HERI items that we identified} as potentially important (i.e., the 11 attributes from HERI that were not selected by our practitioner panel in addition to the five that were selected). However, some of the 11 variables are not available in all the years for which we conducted our primary tests, including four that are not available in any of the recession years. Thus, we run a series of robustness tests as follows: (1) we run models that include all 16 variables only in those years that all variables are available (applicable only to H2), (2) we run models that include our five primary controls plus other controls that are available for all the years for which the five primary controls are available (7 of the 16 possible controls), and (3) we run models that include our five primary controls plus variables that are available in at least two recession years (9 of the 16 possible controls). In the reported univariate tests of H2 and H3, we use samples restricted to include the same sample years as those used in our multivariate tests of H2 and H3. We conduct robustness tests in which we use all sample years in our univariate difference in difference tests of H2 and H3. Univariate and multivariate tests of H2 (effect of low SES) and univariate tests of H3 (effect of low SES by Recession) are robust to all these alternate specifications for both the “non-accounting business” and “all non-accounting” comparison groups. Multivariate tests of H3 are robust to specifications (2) and (3) for the “all non-accounting” specification and to specification (2) for the “non-accounting business” comparison group. H3 is directionally consistent with our prediction, but insignificant for specification (3) for the “non-accounting business” comparison group. It is unclear whether H3’s lack of significance for this one specification for this one comparison group is due to the additional controls or to the restricted sample.

28

We conduct a difference in difference test that compares the decline in the low SES index in recession, as opposed to non-recession years in accounting to the decline in other fields. The results show that low SES representation on average drops in accounting, in other business fields, and in all other fields during recession years. This is consistent with recessions more negatively affecting rates of college attendance among people with fewer economic resources. However, the decline in low SES representation is larger in accounting than in other business fields (p < 0.01) and in the set of all other fields (p < 0.01). Thus, the results support H3.22

We report multivariate analysis of H3 in Table 5, Panel B, which is based on the model:

Select Accounting (0 or 1) = α + β1*lowSES + β2*lowSES*Recession + β3*Recession+ β4*Female + β5*Minority + β6*GPA + β7*Drive + β8*Self-confidence intelligence + β9*Self-confidence social + β10*Writing + time period dummies + ε (2) Our primary variable of interest is the lowSES*Recession interaction. We expect this interaction effect to be negative, i.e., an odds ratio lower than 1. The control variable definitions are the same as in equation (1). Because there are time trends that affect our variables of interest, including fewer first generation college students and increasing popularity of business fields in general, we also include dummy variables for each five-year time period in our sample to ensure that our analyses are not confounded by time effects. The reference time period to interpret each time period effect is 1971 – 1975, i.e., each time period effect tests the interest in accounting in that time period relative to interest in 1971 – 1975.23

22 _{Ideally, we would also be able to explicitly identify the fields chosen by people who might otherwise select into} accounting. However, we are unable to capture this data, even with our dataset of seniors that allows us to observe switches between fields. First, the senior dataset only includes years in which there were no recessions, thus we could not observe differences in substitutes between recession and non-recession years. Second, this dataset only captures alternatives to accounting for those who actually remained in college, and earning a degree is an inherently slower strategy than dropping out. Thus, the data will understate the life history speed of alternatives, and this understatement is likely stronger among low SES individuals, who are more likely to drop out of college.

23 _{We use time period dummies instead of year fixed effects because Recession is defined at the year level. In each} model, the odds ratios for selecting accounting decrease over time. For example, for every person selecting

29

As predicted, the results show a negative interaction effect in which low SES interest in accounting is lower in recession years, and this decrease is greater in accounting than it is in business fields and in the set of all fields. This interaction is striking because, in both columns of Table 5 Panel B, Recession is associated with higher interest in accounting. That is, interest in accounting among low SES individuals is lower during times when overall interest in accounting is higher. We argue that this is due to people from poorer backgrounds adopting faster trajectories in their life history strategies when they observe cues of economic uncertainty. Importantly, if the recession interaction effect simply reflected fewer low SES individuals in college during recessions due to lower application rates or higher dropout rates, then lower SES representation in accounting would not differ from that of other fields. Thus, our multivariate analyses support H3. In sum, these archival analyses provide evidence both that our theorized relations generalize in a large-sample dataset of millions of real choices and that the effects are robust over time. Low SES individuals disproportionately prefer accounting because it is consistent with a slow life history strategy. However, recessions affect patterns of selection into accounting differently than other fields, as the preference for accounting by low SES individuals disproportionately weakens.

6. Supplemental Analyses

In this section, we explore the potential implications of our findings for the profession and for low SES individuals. Specifically, our supplemental analyses first examine implications for the quality of human capital entering accounting, i.e., under what conditions do our theorized effects have benefits for accounting? We then examine job outcomes of those who select into accounting and of low SES individuals who select accounting, as opposed to other fields. This provides some

– 2000. This likely reflects other, often new business fields (e.g., information systems) growing in popularity at a faster rate than accounting enrollments over the past several decades (Madsen 2015).

30

evidence as to whether or not accounting is a slow life history choice relative to common alternatives, and whether the benefits of a slow life history strategy materialize similarly for low SES individuals and high SES individuals. If our theory suggests that accounting attracts entrants because of its security, then this analysis helps identify if accounting delivers this security. 6.1. Implications for the Quality and Diversity of Human Capital Entering Accounting

6.1.1. Low SES, Self-Selection, and Human Capital

We first examine implications of our effects for the quality of human capital in the entry-level accounting labor pool. To do so, we analyze equation (1) at each possible entry-level of lowSES to examine if any of the desirable human capital attributes vary with low SES.24 _{Recall that the}

lowSES values indicate how many of three low SES indicators a person possesses, thus SES is lower as the number of lowSES indicators increases. The first noteworthy finding is that the data are consistent with the low SES effect increasing gender diversity in the accounting labor pool. As shown in Table 6 at the lowest SES level (lowSES = 3), 1.455 women select accounting for every woman who selects another business field and 1.149 women select accounting for every woman who selects another non-business field. By contrast, women from wealthier backgrounds tend to select accounting at disproportionately low rates.

In addition, recall that the results in Table 4 Panel B identified potential deficits in the accounting labor pool. Our expert panel identified writing skills and self-confidence as critical to distinguishing successful performers in accounting, but the results show that accounting entrants lag in these categories. However, Table 6 suggests individuals selecting accounting possess higher levels of these attributes as lowSES increases, relative to business fields and to other fields. In

24 _{We also run a model with interaction terms of lowSES with each of the other variables, and the results are} comparable. We report the decomposed models at each level of lowSES because doing so provides the reader more information about trends in the data and because the odds ratios are more interpretable in this format.

31

brief, the results indicate that the quality of the entry-level accounting labor pool benefits from the low SES effect.25

6.1.2. Do Recessions Undermine Human Capital Benefits for Accounting?

If the low SES effect channels quality talent into accounting, then H3 could naturally decrease representation of this talent and have negative implications for accounting. To assess implications of H3, we analyze a version of equation (2) which has been modified to include interactions of the Recession indicator variable with our control variables at each possible level of

lowSES. As shown in Table 6 Panel B, two noteworthy effects emerge from this analysis. First, the odds ratios for the Female*Recession interaction are less than one across many levels of

lowSES, particularly when the comparison group is all non-accounting degrees.

Thus, the recession effect disproportionately reduces the representation of females from poorer backgrounds. That is, the low SES effect appears to increase gender diversity in accounting but uncertain economic conditions reverse this positive effect. This may suggest that bad economic times undermine accounting’s potential to act as a secure path to business careers for women from poor backgrounds. That is, a gender gap may exist in the potential for accounting to elevate those from poorer backgrounds. Future research into this possibility is warranted.

Second, there is a significant GPA*Recession interaction with an odds ratio greater than one for most levels of SES. This appears to benefit accounting on the surface, as academic ability increases relative to other business fields and to all other fields. We rerun equation (2) at each level of reported GPA in the HERI data in order to more deeply examine whether this effect is likely to

25 _{Inferences for other variables appear to differ depending on the comparison group, and are thus inconclusive. An} exception appears to be that the odds of disadvantaged minorities selecting accounting appears to decrease as lowSES increases. This suggests that the disproportionately high interest in accounting among low SES individuals may be driven by white, as opposed to minority individuals from poorer backgrounds. This is consistent with Hammond’s (1997) analysis of the accounting profession’s struggles to integrate minorities. Our results do show that interest in accounting is disproportionately high among minorities from wealthier backgrounds, but the low SES effect appears unlikely to increase minority representation.

32

benefit accounting. In Table 6 Panel C, the results are not definitive, as the lowSES*Recession

interaction effect is negative for students averaging a B or B+, but positive for students averaging an A-. It is not significant at any other GPA levels. That is, recessions appear to negatively affect selection into accounting among low SES individuals with above average, though not exceptional ability.

This finding may reflect subtle differences in how capability can affect a person’s life history strategy. Those with above average ability may fear being squeezed out by high ability individuals, and thus lack confidence that they will be able to experience deferred rewards. As a result, low SES individuals in this group may revert towards a faster life history strategy when they observe resource scarcity cues, in order to “get what they can when they can.”

6.2. Career Outcomes - Evidence that Accounting is Part of a Slow Life History Strategy The remainder of our analyses focus on providing evidence that accounting is, in fact, a slow life history strategy relative to potential substitutes. Further, we examine whether low SES individuals selecting accounting realize the benefits of a slow life history strategy—that is, does accounting deliver on expectations? To address this question, we compare job outcomes of accounting degree holders to three comparison groups: (1) finance degree holders, (2) other business degree holders, and (3) non-business degree holders. For brevity, we provide details on our sample and methodology in Appendix C, and summarize our findings here.

As indicators of a slow life history strategy, we examine whether accounting delivers relatively high mean wages with low wage variance, low unemployment, and high job security. Table 7 reports career outcome results for those whose highest degree is a Bachelor’s degree in the indicated field and for those whose highest degree is a Master’s degree, with the left columns

33

depicting the full sample and the right columns depicting the low SES sample only. Our low SES

proxy is first generation college students.

The data support our central theoretical assumption that accounting reflects a relatively slow life history strategy. Accounting has a relatively high mean salary with a relatively low standard deviation, which suggests that accounting degrees lead to relatively high “wage floors.” Moreover, accounting degree holders have unemployment that is lower than in finance, other business fields, and non-business fields. In general, our results indicate that, accounting careers on average deliver the attributes of a slow life history strategy. The data are roughly consistent in the full sample of accounting degree holders and the sample of only low SES individuals, suggesting that low SES accountants do not experience worse labor market outcomes.

This contrasts with a striking pattern in finance, which is a common alternative to accounting. The data suggest that Master’s degrees in finance deliver substantial rewards, but not for low SES individuals. Low SES Master’s degree holders compare unfavorably to other finance Master’s degree holders in mean earnings ($66,989 versus $87,826) and unemployment (9.88% versus 4.52%). By contrast, low SES Master’s degree holders in accounting are comparable to other accounting Master’s degree holders in terms of wages ($65,866 versus $68,552) and unemployment (0.77% versus 0.93%).

A plausible explanation is that maximizing the benefits of a Master’s degree in finance requires the type of social capital (i.e., relationships with well-connected people) that low SES

individuals possess at lower rates. This is consistent with research by Rivera (2015) that hiring practices of financial institutions systematically favor those from wealthier backgrounds. By contrast, social capital may not be as important in realizing the benefits of an accounting degree. This is consistent with our broader theory and suggests costs for low SES individuals choosing

34

accounting’s faster, most common alternative. If a low SES individual wants to compete on an equal playing field in a business career, then accounting is a relatively safe and effective choice. 7. Conclusions and Future Research Directions

The accounting profession is dependent on the supply of entry-level human capital into the labor market, but little is known about the determinants of self-selection into accounting. This study uses multiple methods and datasets to examine why people self-select into accounting, focusing on how self-selection depends on SES and the economic conditions at the time of selection. We explain our findings through the lens of life history strategies, drawn from evolutionary biology and psychology.

We find that low SES college attendees, i.e., the pool of minimally-qualified individuals to enter accounting, follow predominately slow life history strategies. That is, they make choices consistent with long-term security. Consequently, our experimental and archival evidence shows that low SES individuals disproportionately self-select into accounting relative to non-accounting business fields and all non-accounting fields. However, low SES individuals are less likely to select accounting in recession, as opposed to non-recession years, and this effect is stronger in accounting relative to non-accounting business fields and all non-accounting fields. Our evidence suggests that this joint effect on self-selection is driven by the importance of job security. Further, supplemental analyses show that accounting careers deliver the benefits of slow life history strategies to low SES individuals, in the form of high mean wages, low wage variation, and low unemployment. In sum, our study sheds light on part of accounting’s unique position in the job market, specifically, as a secure path for low SES individuals to enter professional fields.

Accounting’s potential role as a socioeconomic “ladder” is important for labor market theory in general and for the accounting profession in particular. There is evidence of systematic,

62

income. Rows labeled N show the number of observations in the HERI database for each subsample. Rows labeled N-weighted show observation counts after inflating each observation by its survey probability weight giving the number of individuals in the population of American college freshman represented by HERI observations. ***, **, and * represent statistical significance at the p < 0.01, p < 0.05, and p < 0.10 levels respectively. All p values are two-tailed. See Appendix B for dataset and variable descriptions.

TABLE 5

Archival Tests of H3

–

The Joint Effect of SES and Recession Years

Panel A: Archival Tests of H3-LowSES in Accounting and Non-Accounting across Recession Years Accounting Non-Accounting Business All Non-Accounting

Non-recession years 0.97 0.73 0.79

Recession years 0.93 0.74 0.80

Difference 0.04*** -0.01** -0.01***

Difference in Differences 0.05*** 0.05***

Panel B: Diversity in Accounting versus Non-Accounting across Recession Years Business Degrees Whole Sample

Low SES Index * Recession Year 0.964*** 0.938***

Low SES Index 1.283*** 1.191***

Recession Year 1.096*** 1.211***

Female 1.356*** 0.906***

Minority 1.162*** 1.292***

GPA H.S. 1.678*** 1.156***

Drive 0.971*** 1.066***

Self-Confidence Intelligence 1.099*** 0.928*** Self-Confidence Social 0.702*** 0.923***

Writing 0.658*** 0.513***

Years '76 to '80 0.882*** 1.618*** Years '81 to '85 0.656*** 1.933*** Years '86 to '90 0.589*** 1.719*** Years '91 to '95 0.655*** 1.221*** Years '96 to '00 0.372*** 0.822***

Constant 0.102*** 0.027***

N 532,356 3,119,057

N-Weighted 2,864,196 15,696,147

The analyses in these tables use data from the HERI Freshmen surveys. In Panel B, the columns show odds ratios. The reference time period to interpret each time period effect is 1971 – 1975, i.e., each time period effect tests the interest in accounting in that time period relative to interest in 1971 – 1975. Low SES Index is the sum of cheap tuition, first generation, and low parental income. Rows labeled N show the number of observations in the HERI database for each subsample. Rows labeled N-weighted show observation counts after inflating each observation by its survey probability weight giving the number of individuals in the population of American college freshman represented by HERI observations. ***, **, and * represent statistical significance at the p < 0.01, p < 0.05, and p < 0.10 levels respectively. All p-values are two-tailed. See Appendix B for dataset and variable descriptions.

Do Low SES Students Selecting Into Accounting Have Attributes Desired by the Accounting Profession?

Panel A: During Non-Recession Years, Do Quality Low SES Individuals Select Into Accounting?

Business Degrees All degrees

LowSES = 0 LowSES = 1 LowSES = 2 LowSES = 3 LowSES = 0 LowSES = 1 LowSES = 2 LowSES = 3 Female 1.377*** 1.371*** 1.358*** 1.455*** 0.857*** 0.930*** 1.017 1.149*** Minority 1.472*** 1.206*** 1.018 0.931 1.366*** 1.284*** 1.275*** 1.214*** GPA 1.679*** 1.692*** 1.582*** 1.486*** 1.091*** 1.154*** 1.234*** 1.191***

Drive 0.953*** 0.962* 0.99 0.975 1.096*** 1.050*** 1.029 1.041

Self-Conf. Intelligence 1.069*** 1.086*** 1.107*** 1.258*** 0.892*** 0.929*** 0.967 1.035 Self-Conf. Social 0.692*** 0.705*** 0.711*** 0.710*** 0.934*** 0.921*** 0.906*** 0.870*** Writing 0.647*** 0.653*** 0.685*** 0.712*** 0.490*** 0.512*** 0.551*** 0.576*** Years '76 to '80 1.159*** 1.293*** 1.097 1.157 2.445*** 2.002*** 1.630*** 1.622*** Years '81 to '85 0.654*** 0.749*** 0.794*** 0.790*** 2.532*** 2.229*** 2.374*** 2.634*** Years '86 to '90 0.578*** 0.667*** 0.717*** 0.663*** 2.412*** 2.081*** 2.147*** 2.341*** Years '91 to '95 0.658*** 0.687*** 0.809*** 0.675*** 1.704*** 1.276*** 1.401*** 1.410*** Years '96 to '00 0.372*** 0.409*** 0.477*** 0.467*** 1.176*** 0.898*** 0.947 1.104 Constant 0.099*** 0.120*** 0.165*** 0.232*** 0.024*** 0.029*** 0.024*** 0.026*** N 234,436 129,461 61,059 11,507 1,344,736 787,757 356,639 64,899 N-Weighted 1,060,107 733,129 400,805 91,663 5,667,387 4,154,365 2,201,530 487,914

65

TABLE 6 (continued)

Do Low SES Students Selecting Into Accounting Have Attributes Desired by the Accounting Profession?

Panel B: What are the Effects of Recessions on the Quality of Low SES Individuals Selecting Accounting?

Business Degrees All degrees

LowSES = 0 LowSES = 1 LowSES = 2 LowSES = 3 LowSES = 0 LowSES = 1 LowSES = 2 LowSES = 3 Female * Recession 0.956 0.939* 0.963 1.115 0.923*** 0.870*** 0.861*** 0.975

Minority * Rec 1.058 1.024 1.012 0.823 0.997 1.022 1.079 1.04

GPA * Recession 1.071** 1.04 1.089* 1.348** 0.991 1.015 1.044 1.242**

Drive * Recession 1.03 1.054 1.012 0.851 1.005 1.017 1.061 0.995

Self-Conf Intel * Rec 1.002 1.045 1.025 1.197 0.99 0.996 0.998 1.221* Self-Conf Social * Rec 1.005 1.03 1.039 1.16 0.971 1.002 0.977 1.042 Writing * Recession 1.023 1.04 0.904* 0.768** 0.988 1.038 0.936 0.847 Recession 0.822* 0.901 0.843 0.423** 1.272*** 1.154 1.004 0.503** Female 1.375*** 1.362*** 1.355*** 1.433*** 0.858*** 0.931*** 1.02 1.148*** Minority 1.466*** 1.202*** 1.022 0.937 1.372*** 1.286*** 1.273*** 1.211*** GPA 1.682*** 1.694*** 1.582*** 1.483*** 1.093*** 1.155*** 1.234*** 1.187***

Drive 0.955*** 0.963* 0.989 0.971 1.099*** 1.052*** 1.028 1.04

Self-Conf Intelligence 1.068*** 1.084*** 1.107*** 1.259*** 0.893*** 0.930*** 0.969 1.037 Self-Conf Social 0.691*** 0.703*** 0.710*** 0.704*** 0.942*** 0.924*** 0.908*** 0.868*** Writing 0.647*** 0.652*** 0.685*** 0.711*** 0.492*** 0.513*** 0.552*** 0.577*** Years '76 to '80 0.842*** 0.914** 0.878** 0.82 1.729*** 1.559*** 1.494*** 1.549*** Years '81 to '85 0.575*** 0.674*** 0.755*** 0.894 1.846*** 1.851*** 2.124*** 2.589*** Years '86 to '90 0.515*** 0.604*** 0.682*** 0.768** 1.676*** 1.648*** 1.827*** 2.216*** Years '91 to '95 0.593*** 0.643*** 0.787*** 0.782** 1.265*** 1.111*** 1.303*** 1.455*** Years '96 to '00 0.327*** 0.368*** 0.454*** 0.530*** 0.857*** 0.745*** 0.847*** 1.086 Constant 0.112*** 0.134*** 0.174*** 0.209*** 0.032*** 0.034*** 0.027*** 0.027***

284,962 158,401 74,940 14,053 1,643,208 959,666 437,113 79,070 1,322,833 923,945 503,267 114,152 7,090,755 5,218,399 2,777,732 609,262

66

Do Low SES Students Selecting Into Accounting Have Attributes Desired by the Accounting Profession?

Panel C: Selection into Accounting Relative to Other Business Disciplines by High School GPA

GPA = 4.00 3.67 3.33 3.00 2.67 2.33 2.00 1.00

Low SES Index * Recession 0.987 1.057* 0.943** 0.939*** 0.974 0.946 0.955 1.243 Low SES Index 1.295*** 1.290*** 1.297*** 1.277*** 1.267*** 1.257*** 1.259*** 1.145

Recession 1.059 1.003 1.116*** 1.140*** 1.096** 1.055 1.116 0.754

Female 1.457*** 1.332*** 1.331*** 1.343*** 1.332*** 1.327*** 1.376*** 1.059 Minority 1.006 1.154*** 1.080** 1.173*** 1.192*** 1.309*** 1.275*** 1.313

GPA 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000

Drive 0.984 0.908*** 0.986 0.963** 0.994 0.966 1.062 0.613

Self-Confidence Intelligence 1.131*** 1.094*** 1.125*** 1.067*** 1.053 1.103** 1.178** 2.313** Self-Confidence Social 0.726*** 0.698*** 0.694*** 0.692*** 0.729*** 0.695*** 0.670*** 1.268 Writing 0.657*** 0.617*** 0.629*** 0.681*** 0.675*** 0.712*** 0.779*** 0.955 Years '76 to '80 0.896 1.063 0.901** 0.849*** 0.739*** 0.819*** 0.739*** 1.375 Years '81 to '85 0.599*** 0.646*** 0.660*** 0.679*** 0.588*** 0.636*** 0.723** 0.602 Years '86 to '90 0.493*** 0.594*** 0.586*** 0.616*** 0.548*** 0.593*** 0.631*** 0.319** Years '91 to '95 0.536*** 0.606*** 0.654*** 0.703*** 0.613*** 0.728*** 0.784*** 0.709 Years '96 to '00 0.320*** 0.341*** 0.357*** 0.392*** 0.375*** 0.435*** 0.539*** 0.502 Constant 0.857* 0.740*** 0.590*** 0.480*** 0.420*** 0.332*** 0.232*** 0.171***

N_sub 75,491 94,472 117,781 128,122 61,506 36,877 17,581 526

N_subpop 319,933 419,697 593,821 732,532 387,039 267,772 138,950 4,453

The analyses in these tables use data from the HERI Freshmen surveys. In each panel, the columns show survey weighted logit coefficients. Low SES Index is the sum of cheap tuition, first generation, and low parental income. Rows labeled N show the number of observations in the HERI database for each subsample. Rows labeled N-weighted show observation counts after inflating each observation by its survey probability weight giving the number of individuals in the population of American college freshman represented by HERI observations. ***, **, and * represent statistical significance at the p < 0.01, p < 0.05, and p < 0.10 levels respectively. All p values are two-tailed. See Appendix B for dataset and variable descriptions.

TABLE 7

Indicators of Slow versus Fast Life History Strategies in Jobs Held By People with Accounting versus Other Degrees

All SES

Low SES Only

Mean Security Mean Security

Wages SDWages Unemp%. Satisfaction Important N Wages SDWages Unemp%. Satisfaction Important N Bachelor's degree is Highest Degree, 35 and younger

Accounting $60,124 $31,803 1.73% 89.54% 48.60% 1,545 $62,413 $30,921 0.79% 84.51% 43.46% 568 Finance $62,524 $37,668 3.70% 86.80% 43.49% 767 $67,861 $37,126 0.95% 76.07% 41.77% 202 Other business $51,384 $32,079 3.30% 86.35% 55.18% 3,671 $51,843 $30,499 3.49% 86.06% 53.13% 1,091 Non-business $49,200 $26,119 3.47% 87.25% 52.05% 35,183 $49,829 $23,849 2.90% 88.95% 51.81% 8,658

Master's Degree is Highest Degree, 35 and younger

Accounting $68,552 $23,951 0.93% 64.95% 66.35% 180 $65,866 $21,719 0.77% 98.20% 52.34% 54 Finance $87,826 $26,890 4.52% 83.87% 38.61% 433 $66,989 $29,615 9.88% 94.35% 43.27% 65 Other business $77,013 $37,894 3.08% 83.34% 49.94% 2,193 $70,952 $27,477 5.89% 85.76% 56.00% 427 Non-business $56,018 $33,413 3.01% 91.89% 55.52% 12,517 $54,543 $22,562 2.93% 93.95% 57.50% 2,448 The analyses in these tables use data from the NSCG. All wages are adjusted to 2010 dollars. See Appendix B for dataset and variable descriptions.