Big Five personality traits and sex

Sexual well-being plays an important role in the quality of life. Against this background, we provide an economics-based approach to the relationship between the Big Five personality traits and various dimensions of sexuality. From a theoretical viewpoint, personality influences sexual well-being not only by how a person feels about sex, but also by how the person behaves in a sexual relationship. Personality shapes information sharing about sexual preferences, the way dissonant sexual preferences of the partners are handled, and the extent to which a person is committed to promises made to a partner. Using a large representative dataset from Germany, we find that personality traits play a role in a person’s own sexual satisfaction, in (the self-assessment of) fulfilling their partner’s sexual needs and desires, in sexual communication, in actual and desired frequency of sex, and in extradyadic affairs. Conscientiousness contributes to a mutually beneficial sex life and increases a person’s commitment to their partner. The opposite holds true for neuroticism. While extraversion and openness to experience help realize a mutually beneficial sex life, we find no evidence that they have a commitment value. On the contrary, extraversion is associated with lower commitment to the partner. Agreeableness contributes to higher commitment. However, agreeableness appears to make people more reluctant to express their sexual needs and desires.


Introduction
Sex is an essential component of romantic relationships and quality of life. As suggested by a time diary study conducted by Kahneman et al. (2004), sex is the highest ranked activity in terms of net positive emotional affect even though it occupies a relatively small fraction of total time. Sexual well-being is associated with relational satisfaction, relationship stability, and happiness in general (Blanchflower and Oswald 2004;Cheng and Smyth 2015;Laumann et al. 2006;Rainer and Smith 2012;Schmiedeberg et al. 2017;Sprecher 2002). Sexual wellbeing is also of high political interest. The World Health Organization (2006) emphasizes that improving sexual health (i.e., a state of physical, emotional, and social well-being in relation to sexuality) remains a public health priority across the globe.
The importance of sex for quality of life gives rise to the question of which factors influence sexual well-being. Our analysis addresses this question by examining the influence of the Big Five personality traits (extraversion, openness to experience, conscientiousness, agreeableness, and neuroticism) on various key dimensions of sexuality. We provide both a theoretical discussion and an empirical analysis.
The theoretical discussion develops an economics-based approach to personality and sex. Considering the dual role of personality, we argue that personality is not only a parameter shaping a person's sexual preferences. Personality is also a parameter shaping the person's behavior in a sexual relationship. Therefore, personality influences sexual well-being through how the person behaves in a sexual relationship. First, it shapes communication and information sharing about sexual preferences. Communication helps reduce partners' incomplete information about each other's sexual preferences so they can coordinate on their preferences and realize a win-win situation. Second, personality influences how partners handle dissonant sexual preferences. There are different ways of handling dissonant preferences. On the one hand, altruism may reduce the degree of disagreement. On the other hand, partners may bargain over their sexual activities. Personality shapes a person's altruistic behavior, bargaining tactics and bargaining power. Third, personality influences how commitment problems are solved. A person may promise to be faithful to the partner or to practice or relinquish specific sexual practices but later may be tempted to break the promise. The person's commitment to the promise depends on his or her self-control, fair-mindedness, inclination to comply with norms, and willingness to reciprocate the partner's cooperative behavior.
Our empirical analysis uses the pairfam (Panel Analysis of Intimate Relationships and Family Dynamics), a large representative dataset from Germany, to examine the link between the Big Five personality traits and sexuality. The results provide evidence that the Big Five personality traits play an important role in a person's sex life. Our estimations show that personality traits have an influence on a person's own sexual satisfaction. Moreover, conforming to the notion that personality shapes the behavior in a sexual relationship, our results suggest that a person's personality also plays a role in their partner's sexual fulfillment. Thus, a person's personality influences whether partners can realize a sexual win-win situation (a mutually beneficial sex life). Examining possible transmission channels through which personality has an impact on people's sexual well-being, we show that personality traits are associated with the quality of sexual communication and the actual and desired frequency of sex. Finally, analyzing the determinants of extradyadic affairs, we demonstrate that personality has an influence on a person's commitment to their partner.
Our study contributes in several ways to the economic literature. While economists have been increasingly interested in the determinants of people's wellbeing (see Frey 2008 andFrey andStutzer 2002 for surveys), sex has remained an under-researched topic in economics. A few economic studies have examined the link between sex and happiness (Blanchflower and Oswald 2004;Cheng and Smyth 2015;Loewenstein et al. 2015;Rainer and Smith 2012) or the determinants of infidelity (Adamopoulou 2013;Bishai and Grossbard 2010;Fair 1978;Potter 2011;Smith 2012). Those studies have not considered the role of personality traits. Our study emphasizes the importance of personality for sexual wellbeing and provides broader insights into various dimensions of sexuality.
On a broader scale, we also address a general research gap in family economics. Apart from some notable exceptions (Averett et al. 2020;Boxho et al. 2020;Dupuy and Galichon 2014;Flinn et al. 2018;Lundberg 2011Lundberg , 2012, studies in family economics have paid little attention to the personality of partners as a determinant of the surplus of romantic relationships. This contrasts with studies in labor economics where a rapidly increasing body of research has examined the influence of personality traits on human capital formation and labor market outcomes (Almlund et al. 2011;Borghans et al. 2008). This development in labor market research has been identified as "one of the most exciting developments in labor economics over the past decade" (Cobb-Clark 2015: p. 1). Our study suggests that considering personality traits may lead to a similarly exciting development in family economics.
The link between personality and sex is a topic typically examined by psychologists (Allen 2019;Allen and Walter 2018). Against this background, our study follows a trend in economics to expand its scope of inquiry to topics traditionally addressed by other scientific disciplines. Of course, expanding the boundaries of economics into other social sciences by using larger and representative datasets or examining explanatory variables with greater statistical sophistication would be of limited value if it does not yield additional theoretical insights. A successful expansion of the scope of inquiry into nontraditional topics requires incorporating the new topics into economic thinking (Lazear 2000). Thus, to guide our empirical analysis, we provide an informal theoretical background discussion that integrates the psychological topic into the analytical framework of economics and, particularly, relates the topic to family economics.
The rest of the paper is organized as follows. Section 2 provides the theoretical background discussion. Section 3 describes the data and variables. Section 4 presents the empirical results. Section 5 discusses the results in light of our theoretical considerations. Section 6 concludes.

3 2 Background discussion
In what follows, we set the stage by providing a brief introduction to the Big Five model. We proceed with a theoretical discussion on the transmission channels through which the Big Five personality traits can influence sexuality. Developing an economics-based approach, we first describe general transmission channels and then relate the Big Five personality traits to these channels.

The Big Five personality traits
Psychologists view personality as enduring patterns of feelings, thoughts, and behaviors (Roberts 2009). Personality reflects the tendency of a person to respond in certain ways under certain circumstances. The most widely shared taxonomy of personality traits in psychology is the Big Five model (John et al. 2008;McCrae and Costa 2008). The Big Five model originated in Allport and Odbert's (1936) lexical approach positing that individual differences in personality are encoded in language. Analyzing personality-describing words, psychologists concluded that personality traits can be organized into five dimensions. Since Goldberg (1981), the five dimensions of personality have been known as the Big Five. The Big Five personality traits are extraversion, openness to experience, conscientiousness, agreeableness, and neuroticism. These personality traits have high predictive power for a series of important life outcomes such as educational achievement, job performance, health, mortality, criminality, and divorce (Almlund et al. 2011;Borghans et al. 2008;Roberts et al. 2007).
Extraversion reflects the degree to which a person orients his or her interests and energies toward the outer world of people, i.e., toward social attention and social interaction. Persons with a higher degree of extraversion tend to be characterized by warmth, sociability, assertiveness, activity, excitement seeking, and positive affect.
Openness to experience reflects the degree to which a person is open to change, variety, intellectual stimulation, and new cultural experiences. Persons with a higher degree of openness tend to be characterized by fantasy, esthetics, and ideas.
Conscientiousness reflects the degree to which a person is willing to comply with conventional rules, standards, and norms. Persons with a higher degree of conscientiousness tend to be characterized by order, dutifulness, achievement striving, self-discipline, and deliberation.
Agreeableness reflects the degree to which a person needs pleasant and harmonious relations with others. Persons with a higher degree of agreeableness tend to be characterized by trust, straightforwardness, cooperativeness, altruism, compliance, modesty, and tender-mindedness.
Neuroticism (emotional instability) reflects the degree to which a person experiences the world as threatening and beyond his or her control. Persons with a higher degree of neuroticism tend to be characterized by anxiety, angry hostility, psychological distress, self-consciousness, impulsiveness, and vulnerability.
It is important to note that the Big Five not only influence how a person, in general, feels about the world. The Big Five are also dispositional traits influencing the overall style of the person's adjustment to and engagement of the social world (Buss 1996;Goldberg 1981;McAdams and Pals 2006;Nettle 2006). These traits describe the degree to which the person is able to solve social adaptive problems through communication, cooperation, trust, stability, and dominance.
Similarly, economists emphasize that personality traits play dual roles (Borghans et al. 2008). On the one hand, they can be a source of pleasure, i.e., they influence the utility a person derives from social relationships. On the other hand, personality traits can be viewed as capacities and constraints in the choices the person makes. These choices, in turn, have consequences for the quality of social relationships.
At issue is how personality traits influence sexuality. In what follows, we first provide a general discussion on the transmission mechanisms through which personality may have an effect on sexuality. The discussion is developed within an economics-based approach to integrate this psychological topic into the analytical framework of economics and, in particular, to relate it to family economics. Building from that discussion, we derive specific implications for the link between the Big Five personality traits and sexuality.

An economics-based approach to personality and sex
The dual role of personality also applies to sexuality. On the one hand, personality can influence how much a person enjoys sex. Thus, from the viewpoint of economics, personality has an influence on the utility a person derives from sex. Personality can be seen as a parameter that shapes the utility functions of people. 1 To the extent people differ in their personality, they will have heterogeneous preferences for sex. Depending on personality, sex will be of higher utility for some people than for others.
On the other hand, personality is a parameter shaping a person's behavior in a sexual relationship. This behavior has an influence on the quantity and quality of sex and, hence, on the extent of sexual fulfillment in the relationship. Thus, personality influences the utility from sex through how the person behaves in a sexual relationship. The person's behavior in a sexual relationship is important for at least three reasons. First, it influences communication and information sharing about sexual preferences. Second, it influences how dissonant preferences of the partners are handled. Third, it influences how commitment problems are solved.
Sexual relationships can suffer from information asymmetries (Rainer and Smith 2012). If partners have incomplete information about each other's needs and desires, they may fail to coordinate on their preferences resulting in a suboptimal sex life. Communication helps partners reduce incomplete information about their sexual preferences. By talking and listening, partners can share information about each other's preferences for sexual practices, frequency of intercourse, or timing of orgasm.
This may allow them to find sexual activities working best for them and, hence, resulting in higher mutual sexual satisfaction. However, successful sexual communication requires communication skills, and these skills depend on the partners' personality traits. Personality influences a person's willingness to talk about own preferences and to listen to the partner. Moreover, it has an influence on whether the style of communication is characterized by warmth and empathy or by hostility and impulsiveness.
At best, partners have compatible preferences for sex. so couples with strong communicative skills can realize a sex life that is more satisfying to both of them. But when partners have partially dissonant preferences (e.g., for oral sex, anal sex, timing of orgasm, frequency of intercourse, or pornography consumption), the question arises as to how they handle these dissonant preferences. One possibility is that altruism reduces the degree of disagreement. Altruism can be modeled as an interdependency of utility functions (Bergstrom 1997;Weiss 1997). If a person not only cares about his or her own sexual enjoyment, but also about the partner's enjoyment, he or she will take the partner's preferences for sexual activities into account. This can be seen as a gift given to the partner. On the one hand, an altruistic person to some extent sacrifices his or her utility by forgoing some preferred sexual activities or engaging in activities only preferred by the partner. On the other hand, the altruistic person derives utility from the partner's sexual enjoyment. Moreover, if altruism leads to reciprocal gift exchange, a couple may realize higher mutual sexual satisfaction. Of course, people can differ in the degree of altruistic behavior, and personality has an influence on that degree. Thus, it depends on the partners' personalities whether altruism can solve the problem of dissonant sexual preferences.
It appears to be reasonable to assume that, in general, people are not completely altruistic, but to a greater or lesser extent also exhibit self-interest and care about their own sexual enjoyment. Family economics assumes that dissonant preferences of self-interested household members lead to an intra-household bargaining situation (Agarwal 1997;Komura 2013;Lundberg and Pollak 1996;Yakita 2018). This approach can also be applied to sexual preferences. Partners may solve the problem of dissonant sexual preferences by bargaining over their sexual activities. Personality has an influence on a person's bargaining strength and, hence, on whether he or she can achieve a favorable bargaining outcome. Assertiveness is likely to be associated with a more favorable bargaining outcome for the person while a high willingness to compromise may rather result in a less favorable outcome. In particular, personality may have an influence on the person's threat point. A more sociable person with better communication skills may have better opportunities to find a new partner in case of a disagreement. Better outside options increase the person's bargaining position in the current relationship and help get through his or her preferred sexual activities.
If partners reach a common understanding and agreement on their sexual relationship, commitment problems arise. The common understanding and agreement on the sexual relationship are based on promises partners make to each other. A person may promise to be faithful to the partner or to practice or relinquish specific sexual activities, but later may be tempted to break the promise. The promises made to the partner cannot be made binding through legal enforcement. Thus, the question is if there are other mechanisms ensuring that promises made to the partner are kept. An agreement may be self-enforcing in a repeated game setting (Lundberg and Pollak 1994). In a repeated game setting, partners reward each other's cooperative behavior and threaten to punish opportunistic behavior by refusing cooperation in the future. However, repeated games do not always work. A person is only deterred from opportunistic behavior if he or she does not discount the future loss of cooperation too much. Moreover, repeated games involve multiple equilibria, i.e., not only mutual cooperative behavior is an equilibrium, but also mutual opportunism. We suggest that personality is a parameter playing an important role in whether or not an agreement on sexual activities is self-enforcing. On the one hand, a person's fairmindedness, inclination to comply with norms, and willingness to reciprocate the partner's cooperative behavior increase his or her commitment to the agreement. On the other hand, sociability and communication skills may lead to alternative outside options increasing the temptation to engage in opportunistic behavior.

Implications for the link between Big Five personality traits and sex
In our empirical analysis, we examine the link between the Big Five personality traits and sexual satisfaction. Considering the dual role of personality, the Big Five may not only have an impact on sexual satisfaction by shaping a person's utility function, i.e., by influencing the pleasure the person enjoys from sexual activities. The Big Five may also have an impact through the person's behavior in a sexual relationship and, hence, through the sexual activities the couple does together. In order to gain insights into the transmission channels through which the Big Five have an influence on sexual satisfaction, we also examine a series of behavioral outcome variables. We analyze the determinants of sexual communication. This gives us insights into the role of information sharing in sexual relationships. Furthermore, we examine the degree in which a person meets the partner's needs. This provides indications of whether a person can realize a win-win situation within the sexual relationship. Moreover, we consider the actual and the desired frequency of intercourse. Finally, we examine if a person has sex with someone other than the partner. This provides insights into how the Big Five influence commitment problems and the outside options of the person.
Extraversion and openness to experience should be positively associated with sexual satisfaction. One transmission channel through which these personality traits influence sexual satisfaction is better communication with the partner. Better communication enables a more extroverted and open person to express his or her sexual preferences so that the partner can take into account these preferences. Better communication also implies that the person obtains more information about the partner's preferences so that a win-win situation may be realized. However, greater communication skills make it also easier to get in contact with other potential partners meaning that a more extroverted and open person has better outside options. 2 Better outside options increase the person's bargaining position. While this provides the second transmission channel through which extraversion and openness should result in higher sexual satisfaction, it makes the implications for the partner's sexual fulfillment ambiguous. A more extroverted and open person may get through his or her preferred sexual activities at the expense of the partner's needs and desires. Better outside options may also result in more severe commitment problems. A more extroverted and open person with better outside options may be more tempted to breach an agreement with the partner and to have sex with someone other.
By contrast, conscientiousness can be expected to have a commitment value in a sexual relationship. A higher degree of conscientiousness is associated with greater self-control, sense of responsibility, and dependability. A more conscientious person is rule-abiding and has a stronger focus on long-term relationship. Thus, a more conscientious person should be more likely to resist short-term temptations and to keep promises made to the partner (Peetz and Kammrath 2011). To the extent this induces a more cooperative behavior of the partner, a higher degree of conscientiousness can also result in increased own sexual satisfaction. Moreover, conscientiousness may contribute to a more balanced style of sexual communication leading to more successful information sharing and mutually beneficial outcomes for the partners.
Agreeableness may also contribute to more successful information sharing. An agreeable person is more empathic and tender-minded and, hence, has a more harmonious communication style allowing partners to realize a mutually beneficial sex life. Furthermore, agreeableness is positively associated with altruism (Becker et al. 2012).
Thus, an agreeable person should take the partner's needs and desires to a higher degree into account. To the extent this induces positive reciprocal behavior of the partner and, hence, mutual gift exchange in the sexual relationship, a higher degree of agreeableness can result in higher own sexual satisfaction. However, when bargaining over sexual activities, an agreeable person may compromise too much. The person may be reluctant to express her needs and desires for the sake of harmony. In particular, a higher degree of altruism can lead an agreeable person to scarify his or her sexual needs in favor of the partner's needs. This suggests that the implications of agreeableness for own sexual satisfaction are ambiguous. Nonetheless, even if an agreeable person to some extent sacrifices his or her sexual needs, this does not necessarily imply a negative influence on sexual satisfaction. Altruism means that a person also derives utility from the partner's sexual fulfillment. A higher sexual fulfillment of the partner may compensate an agreeable person for giving up some of their own sexual wishes. Moreover, agreeableness is associated with humility (McCrae and Costa 2008; Van Kampen 2012). 3 Humility implies that a person is satisfied even with less. Thus, an agreeable person may gain sexual satisfaction even if he or she sacrifices some of his or her needs.
Neuroticism can have a negative impact on sexuality for several reasons. As emphasized by Eysenck (1971Eysenck ( , 1976, a person with a higher score on neuroticism tends to have more fears about sexuality and may be more disgusted about some aspects of sexuality. This suggests that a person with a higher score on neuroticism derives less utility from sex than someone with a lower score. Moreover, neuroticism can have a negative impact on sexual satisfaction through the person's behavior in a sexual relationship. Emotional instability tends to entail inadequate and hostile sexual communication (Velten and Margraf 2017). For example, a neurotic person may overreact to criticism triggering negative responses from the partner. This results in poor information sharing and makes it difficult to realize a mutually beneficial sex life. Furthermore, negative emotions are associated with low self-control and a high discounting of the future (Loewenstein 2000). This aggravates commitment problems. A neurotic person is more likely to feel neglect or rejection from the partner, i.e., the person believes that he or she is no longer loved. This can increase the person's propensity to seek an extradyadic sexual relationship (Josephs and Shimberg 2010).

The dataset
Our empirical analysis is based on data from the pairfam (Brüderl et al. 2018;Huinink et al. 2011). A handful of studies used the data to examine some aspects of sexuality (Hajek 2019;Kislev 2020;Morgan et al. 2018;Rainer and Smith 2012;Schmiedeberg et al. 2017;Schmiedeberg and Schröder 2016;Schröder and Schmiedeberg 2015;Smith 2012). However, these studies did not consider the influence of personality traits on sexuality.
pairfam is a nationally representative panel study for Germany funded by the German Research Foundation (DFG). The focus of the pairfam study is on intimate relationships and family relations. A nucleus of themes is addressed annually. Different additional topics are sampled in consecutive waves. The survey includes both a personal interview by a professional interviewer and a self-administered questionnaire for intimate questions. The self-administered questionnaire is completed during the interview using the official survey laptop. In order to avoid interviewer effects, reporting bias, and refusal to answer questions, all questions on sexuality are asked in the self-administered questionnaire. 4 The first wave of interviews was conducted in the year 2008 for three birth cohorts: adolescents born between 1991 and 1993, young adults between 1981 and 1983, and middle-aged adults born between 1971 and 1973. Addresses were randomly drawn from the local population registers of 343 randomly chosen municipalities. More than 12,000 persons participated in the first wave. The sample of the first wave has been used as the basis for the following waves. Nonresponse patterns are similar to other panel studies based on voluntary participation. Bias due to panel attrition does not appear to be a large issue (Müller and Castiglioni 2015).
Information on the Big Five is available for the years 2009, 2013, and 2017. For the empirical analysis, we always use data from those waves for which information on both the Big Five and the respective dependent variable is available. We limit our sample to heterosexual persons who are at least 18 years old and have a partner in the respective year of observation. The average age in our estimation sample is 33 years. Table 1 shows the definitions of the dependent variables and their descriptive statistics. 5 The table also reports the years for which information on the Big Five and the respective dependent variable is available. An ordered variable for sexual satisfaction captures the utility a person derives from sex. The 11-point Likert scale of the variable ranges from 0 "very dissatisfied" to 10 "very satisfied." The regressions with this dependent variable are based on data from the years 2009, 2013, and 2017.

Dependent variables
Our theoretical background discussion suggests personality influences a person's behavior in a sexual relationship and, hence, also the partner's sexual satisfaction. Thus, in order to examine the influence of personality traits on the partner's sexual fulfillment, we consider two ordered variables for the self-assessment of whether the interviewee perceives his-or herself as being a good sex partner and being able to fulfill the partner's needs and desires. The 5-point Likert scale of these variables ranges from 1 "not at all" to 5 "absolutely." The estimations are based on data from the year 2009.
As discussed in our theoretical background section, sexual communication is one transmission channel through which a person's personality traits can influence sexual satisfaction. In order to get insights into this transmission channel, we use 5-point Likert scale variables for expressing preferences during sex and expressing sexual needs and desires in general. The estimations with these dependent variables are based on the year 2009.
As the frequency of intercourse plays an important role in sexual well-being, we also examine the link between personality and this dimension of sexuality. We use information from waves 2009, 2013, and 2017. Frequency of intercourse during the Footnote 4 (continued) on aging processes, parenthood, and relationship status. In particular, persons without a partner have a higher likelihood of refusal. Note that we focus on persons with a partner and control, among others, for age, children, and relationship status in our regressions. 5 Tables 8, 9, and 10 in the Appendix show the distribution of the dependent variables. past 3 months is measured on a 7-point interval scale: never/not in the past 3 months, once per month or less, 2-3 days per month, once per week, 2-3 days per week, more than 3 days per week, and daily. Building from Schröder and Schmiedeberg (2015), we recode the variable to indicate frequency per month: 0, 0.67, 2.5, 4, 10, 20, and 30 days per month. This allows to apply a linear regression analysis. Nonetheless, as a robustness check, we will additionally provide an interval regression.
The survey also provides information on the desired frequency of intercourse. Interviewees answer the question on whether they prefer to have less or more sex than they had during the past 3 months. The categories for the answers are as follows: much less often, somewhat less often, just as often, somewhat more often, and much more often. The analysis of the desired frequency of intercourse is based on the 2017 wave.
Finally, our theoretical discussion suggests that personality has an influence on the person's commitment to the partner. In order to examine this aspect empirically, we consider extradyadic sexual relationships as an inverse indicator of commitment. The survey provides information on extradyadic affairs during the past 2 years. Interviewees choose between four answers: (1) no extradyadic affairs of both partners, (2) extradyadic affair of the partner only, (3) extradyadic affair of the interviewee only, and (4) extradyadic affairs of both partners. Of course, while a person knows whether he or she had extradyadic affairs, the person may not be aware of the extradyadic affairs of the partner. Thus, we focus our analysis on the person's own extradyadic affairs. We define a dummy equal to 1 if the interviewee reports that he or she only or both partners had an extradyadic affair during the past 2 years. For this variable, it is irrelevant whether or not the person is aware of the possible affairs of the partner. For example, consider a situation in which both the person and the partner had extradyadic affairs. If the person is aware of the partner's affairs, he or she will respond with category 4, and the dummy for own extradyadic affairs equals 1. If the person is not aware of the partner's affairs, he or she will respond with category 3, and the dummy for own affairs still correctly equals 1. We use the information provided in waves 2010 and 2014. For the explanatory variables, we use a 1-year lag in the regressions. 6 Table 2 shows the definitions and descriptive statistics of the variables for the Big Five personality traits. As usual in large surveys John 2005, 2007;Soto and John 2017), personality is measured in the pairfam using a short version of the Big Five Inventory. Extraversion, conscientiousness, agreeableness, and neuroticism are assessed with four items. Openness to experience is assessed with five items. The items are measured on a 5-point Likert scale ranging from 1 "strongly disagree" to 5 "strongly agree." For each of the five personality traits, we add up the respective items and divide the sum by the number of items.

Table 2
Definitions and descriptive statistics of the variables for the Big Five Extraversion Score of extraversion constructed from adding up four survey items measured on a 5-point Likert scale ranging from 1 "strongly disagree" to 5 "strongly agree." The sum of items is divided by 4. The items are as follows: (1) I am usually modest and reserved, (2) I get enthusiastic easily and can motivate others easily, (3) I tend to be the strong and silent type, and (4) I am expansive and gregarious. Items 1 and 3 were recoded in inverse order before adding up 3.554

0.802
Openness Score of openness to experience constructed from adding up five survey items measured on a 5-point Likert scale ranging from 1 "strongly disagree" to 5 "strongly agree." The sum of items is divided by 5. The items are as follows: (1) I am interested in many kinds of things, (2) I am intellectual and like to contemplate things, (3) I am very imaginative, (4) I appreciate artistic and esthetic impressions, and (5) I am hardly interested in art.
Item 5 was recoded in reverse order before adding up 3.620 0.684

Conscientiousness
Score of conscientiousness constructed from adding up four survey items measured on a 5-point Likert scale ranging from 1 "strongly disagree" to 5 "strongly agree." The sum of items is divided by 4. The items are as follows: (1) I complete my tasks thoroughly, (2) I make things comfortable for myself and tend to be lazy, (3) I am proficient and work fast, and (4) I make plans and carry them out. Item 2 was recoded in inverse order before adding up 3.882 0.598 Agreeableness Score of agreeableness constructed from adding up four survey items measured on a 5-point Likert scale ranging from 1 "strongly disagree" to 5 "strongly agree." The sum of items is divided by 4. The items are as follows: (1) I tend to criticize others, (2) I trust others easily and believe that people are inherently good, (3) I can be cold and distanced in my behavior, and (4) I can be gruff and dismissive with others. Items 1, 3, and 4 were recoded in inverse order before adding up 3.292 0.700 Neuroticism Score of neuroticism constructed from adding up four survey items measured on a 5-point Likert scale ranging from 1 "strongly disagree" to 5 "strongly agree." The sum of items is divided by 4. The items are as follows: (1) I easily become depressed or discouraged, (2) I am relaxed and do not let myself be worried by stress, (3) I worry a lot, and (4) I easily become nervous and insecure. Item 2 was recoded in reverse order before adding up 2.661 0.786 Psychologists often assume that the Big Five change only modestly after they have developed in childhood and adolescence (Costa et al. 2000). Some even emphasize an important genetic component contributing to a relatively high stability of personality (Bouchard and Loehlin 2001;Kandler et al. 2010). The question of the stability of the Big Five has also been of interest to economists. Cobb-Clark and Schurer (2012) demonstrate for working-age adults in Australia that mean level changes of the Big Five personality traits are very small over a 4-year period. Considering an 8-year time frame, Elkins et al. (2017) find somewhat higher, but generally small changes even for adolescents and young adults. Anger et al. (2017) confirm for working individuals in Germany that the mean level changes of the Big Five are small over an 8-year period.
As the pairfam provides information on the Big Five for the years 2009, 2013, and 2017, we can examine the stability of the Big Five also with our data. Appendix Table 11 shows that the mean level changes of the Big Five personality traits for the periods 2009/2013, 2013/2017, and 2009/2017 are small. The table also reports Cohen's d. This measure defines the mean level change in terms of a standard deviation change of the respective trait. All values for Cohen's d amount to less than 0.2 implying that the changes can be considered very small.
Altogether, a series of studies suggest that while not literally fixed, the Big Five personality traits exhibit relatively high stability. However, there is an ongoing debate as to the role of age and major life events in intra-individual changes of personality. Specht et al. (2011) find that age has an influence throughout the life span and provide some evidence of a curvilinear influence. The evidence on the role of major life events appears to be mixed. While some studies indicate some moderate and rather specific influences of single life events (Angeli et al. 2018;Anger et al. 2017), others conclude that intra-individual changes are generally only weakly or even not related to major life events and that changes are not economically meaningful (Costa et al. 2000;Cobb-Clark and Schurer 2012). Moreover, Specht et al. (2011) suggest that there can also be reverse causation with personality influencing life events.
Whatever the exact role of life events may be, our dataset allows us to include a rich set of control variables capturing demographic, economic, and family-related factors. This should mitigate endogeneity concerns. We follow most of the studies on personality (e.g., Caliendo et al. 2014;Cobb-Clark and Tan 2011;Mueller and Plug 2006;Risse et al. 2018) and consider personality traits with suitable caution as exogenous. While the regressions may not allow definite causal inferences to be drawn, they provide a crucial first step to bring important new insights to family economics which can be interpreted in light of our theoretical considerations.

Control variables
Appendix Table 12 provides the definitions and descriptive statistics of the control variables. We control for the economic situation by including variables for the years of schooling and the person's labor market status. A variable for health satisfaction accounts for overall health status. Demographic characteristics are captured by variables for the number of children, the presence of a baby in the household and for the person's gender, religious affiliation, migration background, and age. In order to account for a nonlinear influence of age on sexuality, we also include a quadratic and a cubic age variable. The type of relationship is controlled for by variables for relationship duration and being married to the partner. For persons not married to the partner, we take into account whether or not the couple lives together in the same dwelling. We also include a variable for the number of previous marriages (and, hence, the number of previous divorces). Moreover, as East Germans appear to have more equal gender roles than West Germans, we control for residing in East Germany. 7 Finally, cohort dummies are included. In regressions with more than one wave of the data, we also control for the year of observation.

Own sexual satisfaction and satisfying the partner's sexual wishes
In the regressions, we standardize the Big Five variables to have a mean of 0 and a standard deviation of 1. This allows a straightforward interpretation of the results. The influence of a personality trait can be interpreted in terms of a one standard deviation change in the score for this trait.
Column 1 of Table 3 shows the key results on own sexual satisfaction. 8 The determinants of sexual satisfaction are estimated by using the random effects ordered logit model. 9 The random effects model accounts for the cross-period correlation of individual-specific error terms. Furthermore, we cluster the standard errors at the individual level using the Huber-White sandwich estimator. All of the five personality variables take significant coefficients. Extraversion, openness to experience, 7 See, for example, Jirjahn and Chadi (2020) for a discussion of gender roles in East and West Germany. 8 Control variables are included in the regressions, but are suppressed to save space. Full results are available from the authors upon request. 9 We prefer the random effects (RE) over a fixed effects (FE) model. The FE model only accounts for within variation of variables but throws away all the information contained in the cross-sectional variation in the data. Thus, time-invariant variables cannot be included. While variables with small variation across time may be included, this can result in highly inefficient estimates. Attenuation bias due to measurement errors is also likely to be more severe in FE regressions (Swaffield 2001). Finally, while the FE model solves the problem of unobserved time-invariant influences, it can aggravate a possible bias due to the omitted time-varying variables because dropping the between variation increases the influence of time-varying misspecification on parameter estimates (Plümper and Troeger 2019). A potential shortcoming of the RE model is the requirement that the random effects are uncorrelated with the explanatory variables. However, Clark and Linzer (2015) provide Monte Carlo simulations showing that the RE model may be preferred even if this requirement is violated. As long as the correlation between random effects and explanatory variables is not too high and, hence, the resulting bias is small enough, the lower variance of the RE estimator produces root mean square errors below those of the FE estimator. This advantage is particularly strong in regressions including variables with small within variation. Moreover, note that comparisons of RE and FE models are often made under the assumption that both include the same explanatory variables. Yet, the RE model allows to control for time-invariant factors reducing the risk that the random effects are correlated with the explanatory variables. conscientiousness, and agreeableness are significantly positive determinants of sexual satisfaction while neuroticism is a significantly negative determinant. The results are not only statistically significant, but also quantitatively meaningful. A one standard deviation increase in the neuroticism score is associated with a 7 percentage point lower likelihood of reporting one of the three highest categories of the 11-point Likert scale for sexual satisfaction. Given that we have 43% of observations in these categories, this implies a decrease of 16%. A one standard deviation increase on the scale for conscientiousness increases the likelihood of reporting one of the three highest satisfaction categories by about 5 percentage points. Taking again into account that there are 43% of observations in these categories, this implies an increase of 12%. The marginal effects of agreeableness, openness, and extraversion are 3, 2, and 1 percentage points, respectively.
As suggested by our theoretical considerations, a person's personality traits may not only influence his or her own sexual satisfaction, but also the partner's sexual satisfaction. In order to examine the influence of a person's personality on the partner's sexual satisfaction, we use interviewees' self-assessments and estimate the determinants of being a good sex partner and fulfilling the partner's sexual needs and desires. Columns 2 and 3 provide the key results of the ordered logit regressions. The estimations show no significant association between agreeableness and the two indicators of the partner's sexual satisfaction. Extraversion, openness to experience, and conscientiousness are significantly positive determinants of being a good sex partner and fulfilling the partner's sexual needs and desires. Neuroticism is a significantly negative determinant. These influences are also quantitatively meaningful. For example, a one standard deviation increase on the scale for conscientiousness is associated with a 6 percentage point higher likelihood of reporting one of the two highest categories of the 5-point scale for being a good sex partner. Given that we have 64% of observations in the two highest categories, this implies an increase in the likelihood by 9%.

Sexual communication
In order to examine possible transmission channels, we now turn to the link between personality traits and sexual communication. Table 4 presents the key results of ordered logit regressions on the determinants of expressing preferences during sex and expressing sexual needs and desires in general. Extraversion, openness to experience, and conscientiousness are positive determinants of the two indicators of sexual communication whereas agreeableness and neuroticism are negative determinants. Again, the influences are not only statistically significant, but also quantitatively meaningful. For example, a one standard deviation increase in the extraversion score is associated with a 5 percentage point higher likelihood of reporting one of the two highest categories of the 5-point scale for expressing preferences during sex. Given that we have 73% of observations in the two highest categories, this implies an increase in the likelihood by 7%.

Frequency of sex
Previous research has shown that the frequency of sex plays a role in happiness (Blanchflower and Oswald 2004;Cheng and Smyth 2015;Schmiedeberg et al. 2017). This gives rise to the question of how personality traits influence the frequency of sex. Table 5 shows the key results of a random effects GLS regression on the determinants of frequency of intercourse per month. As a check of robustness, we also present a random effects interval regression using the initial categories for the frequency of sex. For the interval regression, we translated the interval boundaries shown in Appendix Table 10 into the frequency of intercourse per month. The interval regression and the linear regression on the midpoints of the intervals yield very similar results. While the variable for agreeableness does not take a significant coefficient, the other four personality traits emerge as significant determinants. Extraversion, openness to experience, and conscientiousness are positively associated with the frequency of intercourse. Neuroticism is negatively linked to the frequency of intercourse. To provide a quantitative illustration, a one standard increase in the neuroticism score implies a decrease in the monthly frequency of intercourse by almost half a day. Taking into account that the average frequency is about 6 days per month, this is an 8 percentage decrease in the frequency of intercourse.
So far, we have considered the actual frequency of sex. We recognize that the desired frequency of sex may differ from the actual one (Loewenstein et al. 2015;Smith et al. 2011). Thus, using a multinomial logit approach, we estimate the determinants of desiring much less frequent, somewhat less frequent, somewhat more frequent, or much more frequent sex. The reference group consists of persons who are satisfied with the actual frequency of intercourse. Table 6 provides the key results. Only a few of the results are significant. Extraversion is significantly associated with a higher likelihood of desiring much more frequent sex. Conscientiousness is significantly linked with a lower probability of desiring much less frequent sex. Neuroticism is significantly associated with both a higher likelihood of desiring much less frequent sex and a higher likelihood of desiring somewhat more and much more frequent sex. The estimations show no significant influence of agreeableness and openness.

Extradyadic affairs
Finally, we examine the role of personality traits in having extradyadic affairs during the last 2 years. This gives insights into how personality influences a person's commitment to the partner. Table 7 shows the key results of a random effects logit estimation. Since the share of observations with extradyadic affairs is at 3% relatively low, the table also presents Firth's logit estimation as a robustness check. 10 The random effects regression and Firth's model yield very similar results. While openness to experience does not emerge as a significant determinant, the variables for the other four personality traits take significant coefficients. While extraversion and neuroticism are significantly positive determinants of extradyadic affairs, conscientiousness and agreeableness are significantly negative determinants. For a quantitative illustration, let us consider a one standard deviation increase on the scale for conscientiousness. This increase is associated with a 0.7 percentage point lower likelihood of having extradyadic affairs. Given that there are 3% of observations with extradyadic affairs, this implies a change in the likelihood by about 23%.

Further estimations
We performed a series of further estimations to check the robustness of the results. 11 One robustness check focused on the minimum age of persons in our estimation sample. The findings presented in this study are based on an estimation sample of persons who are at least 18 years old. One may ask if the results are influenced by younger persons whose personalities might still be forming. Thus, as a robustness check, we excluded younger persons and ran the regressions only for those who are at least 25 years old. This exercise yielded a very similar pattern of key results. We also experimented with alternative specifications of the regressions to check the robustness of the results. We ran the regressions without controls for the presence of a baby, number of previous marriages, and duration of the relationship. Exclusion of these controls did not change the key results. Table 6 Determinants of desire for less frequent or more frequent sex Control variables are included in all regressions but are suppressed to save space. The Big Five variables are standardized. Method: multinomial logit. The reference group consists of persons who prefer to have sex just as often as they had during the past 3 months. The table shows the estimated coefficients. Z-statistics are in parentheses. The average marginal effects are in square brackets * p < 0.1; **p < 0.05; ***p < 0.01 Furthermore, in order to mitigate the remaining endogeneity concerns, we used 4-year lags of the Big Five personality traits in our main regression on sexual satisfaction. This implied that we lost one wave of observations so the estimation was based on panel data from the years 2013 and 2017; i.e., sexuality in 2013 was explained by personality traits in 2009 and sexuality in 2017 by personality traits in 2013. The controls were still taken from the respective actual year. This exercise confirmed a significantly negative influence of neuroticism and significantly positive influences of openness, conscientiousness, and agreeableness on sexual satisfaction. However, the coefficient on extraversion was less precisely estimated. This was likely due to the substantially lower number of observations.
We also used our main regression on sexual satisfaction to examine the issue of item nonresponse in more detail. Item nonresponse might lead to a non-randomly selected estimation sample. A non-randomly selected sample, in turn, can result in biased estimates if there are unobserved factors influencing both a person's sexuality and his or her propensity to answer the corresponding question. To account for such unobserved factors, we used Heckman's two-stage sample selection correction. In the first stage, we used a probit regression to estimate the determinants of answering the question on sexual satisfaction. The probit results were used to calculate the inverse Mills ratio. In the second stage, the inverse Mills ratio was included as an additional explanatory variable in the satisfaction regression. To avoid that identification relies solely on the inherent nonlinearity of Heckman's approach, we imposed an additional exclusion restriction. The first-stage probit, but not the second-stage regression, included a variable for the interviewee's expressed willingness to participate in future waves of the survey. Moreover, we also excluded some controls from the second-stage regression which did not emerge as significant determinants of sexual satisfaction. The inverse Mills ratio did not take a significant coefficient in the satisfaction regression. Most importantly, including the inverse Mills ratio did not change the key results in the satisfaction regression. Thus, the key results persisted even when correcting for possible sample selection due to item nonresponse. Furthermore, we examined whether the link between personality and sexuality is heterogeneous and depends on circumstances. In particular, we examined if the influence of personality on sexuality depends on the duration of the relationship. A longer duration of the relationship means that the partners have more experience with each other and may indicate that the relationship becomes more serious. In order to examine the moderating role of relationship duration, we added interaction terms of relationship duration with the Big Five to the regressions. While this exercise also confirmed our key pattern of results, we additionally found some significant interactions with conscientiousness and neuroticism. The positive influence of conscientiousness on sexual communication, fulfillment of the partner's sexual wishes, and frequency of intercourse is increasing with the duration of the relationship. The negative influence of neuroticism on own sexual satisfaction, sexual communication, and frequency of intercourse is decreasing with relationship duration.
Finally, note that our study focuses on persons who have a partner. This reflects our interest in the role personality plays in romantic relationships. Of course, personality can also have a role in the sex life of singles since singles may engage in casual sex. Thus, as a check of robustness, we added singles to our estimation sample. We expanded the specification of the regressions by including a dummy for singles and variables for the interaction of this dummy with the Big Five personality traits. This exercise also confirmed our key results. Moreover, it showed that extraversion plays an even stronger role for singles in sexual satisfaction, fulfilling the partner's sexual wishes, sexual communication, and frequency of sex.

Discussion of results
Our results show that the Big Five personality traits play an important role in the sex life of people. The Big Five personality traits not only play a role in a person's sexual satisfaction. They also play a role in how the person behaves in a sexual relationship. This conforms to our theoretical considerations suggesting that personality traits influence sexual communication and information sharing, the way dissonant sexual preferences of the partners are handled, and the extent to which the person is committed to promises made to the partner.
Neuroticism is associated with lower sexual satisfaction. On the one hand, a higher degree of emotional instability may entail that a person derives less utility from sex as he or she fears sex or is disgusted about some aspects of sexuality. On the other hand, our results suggest that a higher degree of emotional instability negatively affects sexual satisfaction through the person's behavior in the sexual relationship. Neuroticism is negatively associated with (the self-assessment of) being a good sex partner and fulfilling the partner's sexual needs and desires. This indicates that a person's emotional instability also negatively affects the partner's sexual satisfaction and makes a mutually beneficial sex life less likely. Our findings on sexual communication corroborate this view. Neuroticism is negatively associated with expressing preferences during sex and the ability of expressing sexual needs and desires in general. This conforms to the notion that emotional instability entails inadequate and hostile sexual communication and information sharing. Furthermore, neuroticism is associated with a lower frequency of sex and a lower likelihood of being satisfied with the actual frequency of sex. Interestingly, a higher degree of neuroticism increases both the likelihood of desiring less frequent and the likelihood of desiring more frequent sex. This indicates that a neurotic person has rather volatile sexual preferences and is driven by impulsivity. Hence, it is more difficult for the person and the partner to coordinate their preferences and to handle dissonant preferences. Finally, our estimations show that neuroticism is associated with an increased likelihood of having extradyadic affairs. This conforms to the notion that lower self-control and a higher discounting of the future entail more severe commitment problems.
Conscientiousness is associated with higher sexual satisfaction. Moreover, it is also positively associated with being a good sex partner and fulfilling the partner's sexual needs and desires. Thus, our findings fit the notion that a higher degree of conscientiousness helps realize a win-win situation within the sexual relationship. As suggested by our theoretical considerations, conscientiousness may contribute to a more balanced style of sexual communication, a more fair-minded and cooperative handling of dissonant sexual preferences, and a higher commitment to promises made to the partner. Indeed, our empirical results confirm a positive role of conscientiousness in sexual communication. Conscientiousness is positively associated with expressing preferences during sex and expressing sexual needs and desires in general. Moreover, our results provide evidence that conscientiousness has a commitment value in a sexual relationship. Conscientiousness is associated with a lower likelihood of having extradyadic affairs.
Agreeableness is also associated with higher sexual satisfaction. However, our estimations provide no evidence that improved sexual communication is a transmission channel. Quite the contrary, we find evidence of a negative role of agreeableness in sexual communication. Agreeableness is negatively associated with expressing preferences during sex and expressing sexual needs and desires in general. As suggested by our theoretical considerations, there can be two opposing influences. On the one hand, agreeableness may contribute to a more harmonious and empathic style of communication. On the other hand, a stronger desire for harmony and a higher degree of altruism may imply that an agreeable person to some extent sacrifices his or her personal needs. Our empirical results on sexual communication suggest that the latter influence dominates. The interesting point is that an agreeable person nonetheless experiences higher sexual satisfaction. This may be explained by a higher degree of humility implying that an agreeable person gains sexual satisfaction even if he or she sacrifices some of his or her needs. Furthermore, our estimations show that agreeableness is associated with a lower likelihood of having extradyadic affairs. This suggests that agreeableness has a commitment value.
Extraversion is associated with both higher sexual satisfaction and higher frequency of sex. Moreover, it is a positive determinant of being a good sex partner and fulfilling the partner's sexual needs and desires. Thus, extraversion appears to contribute to a mutually beneficial sex life of the person and the partner. Our results suggest that improved sexual communication is one transmission channel. Extraversion is positively associated with expressing preferences during sex and the ability of expressing sexual needs and desires in general. However, extraversion is also associated with a higher likelihood of having extradyadic affairs. Thus, while a person's extraversion contributes to a mutually beneficial sex life of the couple, at the same time, it appears to entail more severe commitment problems. The positive link between extraversion and extradyadic affairs also indicates that a more extroverted person has better outside options increasing his or her bargaining power within the sexual relationship. A stronger bargaining power enables a more extroverted person to get through his or her preferred sexual practices. Altogether, our results conform to the notion that a person's extraversion increases both the joint surplus generated by a couple's sexual relationship and the person's share in this surplus.
We also find evidence that openness to experience is positively associated with sexual satisfaction and frequency of intercourse. Furthermore, our estimations provide evidence that openness to experience is a positive determinant of being a good sex partner and fulfilling the partner's sexual needs and desires. This suggests that an open person is more able to take into account the partner's sexual preferences. Our results also show that openness to experience contributes to improved sexual communication. It is positively associated with expressing preferences during sex and the ability of expressing sexual needs and desires in general. This indicates that a person who is more open to the partner's sexual preferences has better opportunities to communicate his or her preferences to the partner. In contrast to extraversion, we find no evidence that openness to experience is associated with extradyadic affairs. Thus, our analysis provides no evidence that openness to experience has an influence on commitment problems or the person's outside options.

Conclusions
In this study, we have developed an economics-based approach to the role personality plays in people's sex life. Personality not only shapes a person's preferences for sex. It also shapes a person's behavior in a sexual relationship. This behavior has an influence on the quality and quantity of sex and, hence, on the sexual well-being of the person and his or her partner. Personality shapes sexual communication and information sharing, the way dissonant sexual preferences of the partners are handled, and the extent to which the person is committed to promises made to their partner.
Using a large representative dataset from Germany, our empirical results confirm that personality plays an important role in various dimensions of sexuality. The Big Five personality traits have an influence on own sexual satisfaction, sexual fulfillment of the partner, sexual communication, actual and desired frequency of sex, and extradyadic affairs. In summary, conscientiousness contributes to a mutually beneficial sex life and increases the person's commitment to the partner. The opposite holds true for neuroticism. Neuroticism not only decreases the person's and the partner's sexual well-being. It is also associated with a lower commitment to the partner. While extraversion and openness to experience help realize a mutually beneficial sex life, we find no evidence that they have a commitment value. Quite the contrary, the estimations suggest that extraversion is associated with lower commitment to the partner. Agreeableness is associated with a higher commitment. However, it appears to make people more reluctant to express their sexual needs and desires.
We recognize the need for future research within this theme. Economists have been increasingly interested in the determinants of life satisfaction. Against this background, a small number of econometric studies have examined the link between sex and happiness. Future research could fruitfully examine if this link depends on personality traits. Personality may not only influence the quality and quantity of sex but may also moderate how sex translates into life satisfaction. Furthermore, comparative analyses could be a fruitful avenue for future research. Future studies could systematically examine the relationship between personality and sex for different cultures, countries, and societies. Finally, on a broader scale, our study suggests that giving sex and personality a more prominent role in family economics could lead to interesting theoretical and empirical advancements in this field. For example, it would be interesting to examine the influence of sexuality on family formation and dissolution. Appendix Table 8  Table 9  Table 10  Table 11 Table 12