Technical ReportPDF Available

Sex and Happiness

Authors:

Abstract and Figures

We examine the relationship between sexual activities and happiness using a sample of 3800 adults from China. We establish some firm results about the contribution of sexual activities and relationships to happiness for an important country other than the United States. Our main finding is that those who have more sex and better quality sex, proxied by frequency of orgasm and emotional and physical satisfaction with one's primary sex partner, are happier. Another major finding is that the happiness maximizing number of sexual partners is one. We also identify important gender differences between men and women. For men, there is a stronger relationship than for women, between the frequency, and physical aspects, of sexual intercourse and happiness. For women, there is a stronger relationship than for men between giving, and receiving, affection to/from their primary sexual partner and happiness.
Content may be subject to copyright.
1
DEPARTMENT OF ECONOMICS
ISSN 1441-5429
DISCUSSION PAPER 39/14
Sex and Happiness
Zhiming Cheng and Russell Smyth
Abstract
We examine the relationship between sexual activities and happiness using a sample of 3800 adults
from China. We establish some firm results about the contribution of sexual activities and
relationships to happiness for an important country other than the United States. Our main finding is
that those who have more sex and better quality sex, proxied by frequency of orgasm and emotional
and physical satisfaction with one’s primary sex partner, are happier. Another major finding is that
the happiness maximizing number of sexual partners is one. We also identify important gender
differences between men and women. For men, there is a stronger relationship than for women,
between the frequency, and physical aspects, of sexual intercourse and happiness. For women, there
is a stronger relationship than for men between giving, and receiving, affection to/from their
primary sexual partner and happiness.
Keywords: Happiness, sexual activities, sexual intercourse, subjective wellbeing.
JEL classification: H00, J00, D60.
We thank Vinod Mishra and Andrew Oswald for helpful suggestions on an earlier version of this paper.
© 2014 Zhiming Cheng and Russell Smyth
All rights reserved. No part of this paper may be reproduced in any form, or stored in a retrieval system, without the prior written
permission of the author.
2
Introduction
A large, and emerging, economics literature exists that examines the determinants of
happiness (see eg. Clark et al, 2008; Dolan et al., 2008; Frey & Stutzer, 2002). One
activity that has received little examination in this literature is the relationship
between sexual activity and happiness. There seems to be only one published article
in the economics literature on the relationship between sex and happiness
(Blanchflower & Oswald, 2004).
1
This is surprising given the centrality of sex in
people’s lives. As Woloski-Wruble et al. (2010, p. 2402) put it: “individuals are
sexual beings throughout their lives”. People clearly consider sex to be an important
contributor to overall happiness. For example, Kahneman et al. (2004) specifically
asked women to rate the happiness value of several daily activities. Of 19 activities,
sex received the highest happiness value. While one might object to such research on
the basis that economists have no place peering into people’s bedrooms, the reality is
that economists have explored several other aspects of sex. Some examples include
economic returns to beauty and unprotected sex among sex workers (Islam & Smyth,
2012); the effect of sexual activity on wages (Drydakis, 2013) and the role of
education in contributing to sexual satisfaction (Rainer & Smith, 2012).
Blanchflower and Oswald (2004) examine the relationship between various measures
of sexual activity and happiness using a dataset on approximately 16,000 Americans.
These authors find a strong positive correlation between sexual activity and
happiness. A limitation of their findings, however, is that they were unable to
1
Other studies in the sexual psychology and sociology literatures have looked at the relationship
between specific aspects of sexual activity and happiness. Examples are Doran & Price (2014)
(viewing pornography); Dogan et al. (2013); Laumann et al. (2006); Woloski-Wruble et al. (2010)
(sexual satisfaction or sexual wellbeing); Abdolsalehi-Najafi & Beckmanm (2013) (sexual guilt);
Wadsworth (2014) (frequency of sex and frequency of sex of likely reference groups); and Monk-
Turner and Turner (2010) (subjective wellbeing among those who exchange sex and money). A related
study examines the effect of smoking on sexual behaviour and subjective wellbeing (Yamamura,
2014). None of these studies address the issue of causality between sexual activity and happiness.
3
establish causation between sexual activity and happiness. As Blanchflower and
Oswald (2004 p. 394) put it: “Although, for the sake of persuasive identification, it
would be desirable to have instrumental variables for sexual activity [we make] no
adjustment for endogeneity. Our instinct is that solving the endogeneity problem
working out whether sex causes happiness or causality runs in the opposite direction
will be particularly difficult. Future work will have to return to this issue”.
As Blanchflower and Oswald (2004) allude to, intuitively it is difficult to think of
convincing instrumental variables (IV) that are correlated with sexual activity, but not
happiness. To address this, we estimate the relationship between various measures of
sexual activity and happiness using a novel identification strategy, proposed by
Lewbel (2012), which utilizes a heteroskedastic covariance restriction to construct an
internal IV. It has the advantage that it can be used to derive causal estimates where
other sources of identification, such as conventional IVs, are not available. To do so,
we use a dataset with a randomly selected group of 3800 adults in China.
Our results confirm, and extend, several of Blanchflower and Oswald’s (2004)
substantive findings for an important country other than the United States. Key among
these is that having more frequent sex with a primary partner in the past 12 months
has a positive effect on happiness and that the optimal number of sexual partners to
maximize happiness in the previous 12 months is one. Consistent with Blanchflower
and Oswald (2004) we also find that engaging in extra-marital sex and paying for sex
lowers happiness. We also consider the effects of other sexual activities on happiness,
not considered by Blanchflower and Oswald (2004). These include measures of the
frequency of orgasm, sexual satisfaction, partner affection and frequency of
masturbation. This allows us to identify important gender differences between men
4
and women that go beyond the findings in Blanchflower and Oswald (2004).
Specifically, we find for men, there is a stronger relationship than for women,
between the frequency, and physical aspects, of sexual intercourse and happiness. For
women, there is a stronger relationship than for men between giving, and receiving,
affection to/from their primary sexual partner and happiness.
Method
We estimate the following empirical function:
happinessi = f(Xi, S, εi) (1)
where happiness is the happiness of the ith respondent; X is a vector of personal
characteristics; S is a vector of sexual activity variables; and ε is the error term.
Equation (1) can be estimated using ordered logit or ordinary least squares (OLS).
The former is often more popular among economists, while psychologists tend to
favour using OLS. Ferrer-i-Carbonell and Frijters (2004) suggest that results are not
sensitive to the choice of OLS, that treats the dependent variable as cardinal, or
ordered probit/logit methods that treats it as ordinal. While economists typically treat
the happiness variable as ordinal, on theoretical grounds Ng (1997, 2008) advocates
treating subjective wellbeing as cardinal. We estimate Equation (1) using OLS. The
main reason for doing so is that OLS provides benchmark estimates for comparison
with the Lewbel (2012) estimates, derived using generalized method of moments
(GMM). We know of no equivalent method to the Lewbel (2012) approach using an
ordered probit/logit model. However, in robustness checks, which are not reported
given the number of estimates involved, we also treated happiness as ordinal and the
sign and significance of the sexual activity variables are similar to when we use OLS.
5
The problem with OLS estimates is that they only show the association between
sexual activity and happiness. In other words, they leave open the question: Are
people who have more sex happier or do happier people have more sex? The latter is
perfectly plausible if happier people represent more attractive partners, which seems
likely. In the absence of a conventional IV for sexual activity in the dataset, and not
even certain that an intuitive argument can be made for the existence of such a
variable, we employ the Lewbel (2012) method which relies on the existence of
heteroskedasticity in the data to establish causality. The relevant model is:
happiness= X’β1 + Sϒ11 ξ1=α1U + V1 (2)
S= X’β2 2 ξ2=α2U + V2 (3)
Let happiness be happiness and S be sexual activity. U denotes the individual’s
unobserved characteristics, which affects both his or her sexual activity and
happiness. V1 and V2 are idiosyncratic errors. Some of the structural parameters in the
above equations are not identifiable without additional information. Generally one
obtains identification by either imposing equality constraints on the coefficients of X
(i.e. using OLS regression), or assuming that one or more elements of β1 are equal to
zero. This permits the estimation of Equation (2) using instrumental variables.
The Lewbel (2012) approach uses the heteroskedasticity in the data to estimate the IV
regression. The Lewbel (2012) approach involves taking a vector Z of observed
exogenous variables and utilizing [Z-E(Z)]ξ2 as an instrument, provided that:
E(X ξ1)=0, E(X ξ2), cov(Z, ξ1, ξ2) = 0 (4)
and there is some heteroskedasticity in ξj. The vector Z could be a subset of X or equal
to X. Using the above chosen set of instruments, one can use GMM to estimate the IV
regression, as one would do with conventional IVs. As ξ2 is a population parameter,
6
and it cannot be directly observed, we use its sample estimate
, obtained from the
first stage regression and consequently use the vector [Z-E(Z)]
as IVs.
A potential limitation of the Lewbel (2012) method is that it relies on a set of
assumptions. Thus, the utility of the method depends on the extent to which the
assumptions seem reasonable (see Mishra et al. 2014 for a discussion of these points).
Given that the reader may be sceptical of an identification strategy that does not rely
on conventional IVs, it is important to fully canvass, and be aware of, the issues.
First, the assumptions specified in Equation (4), are all based on population
parameters and are generally non testable. This, however, does not represent a major
obstacle because most of these assumptions are standard. As Lewbel (2012, pp. 69)
puts it: “These are all standard assumptions, except that one usually either imposes
homoscedasticity or allows for heteroskedasticity, rather than requiring
heteroskedasticity”. Thus, the only nonstandard assumption is heteroskedasticity.
Consider the assumption of heteroskedasticity in ξj. The exact form of
heteroskedasticity requirement as derived in Lewbel (2012) is cov(Z, ξ22)0.
However Lewbel (2012) suggests using the estimate of the sample covariance
between Z and squared residuals from the first stage regression linear regression on X
to test for this requirement, using the Breusch and Pagan (1979) test for
heteroskedasticity. As noted by Lewbel (2012, p.71), “… if cov(Z, ξ22) is close to or
equal to zero, then [Z-E(Z)]ξ2 will be a weak or useless instrument, and this problem
will be evident in the form of imprecise estimates with large standard errors”. In the
data used in this paper, the Breusch and Pagan test for heteroskedasticity is highly
significant throughout, indicating that the sample estimate of cov(Z, ξ22) is different
from zero and the heteroskedasticity assumption for Lewbel (2012) is fulfilled.
7
Another potential problem with the Lewbel (2012) estimates is that they may be
sensitive to the choice of Z. This issue is compounded because there is currently no
basis to justify the particular choice of Z. This leaves open the possibility that one
could use alternative choices of Z to obtain results consistent with one’s priors or to
skew the conclusions in one direction consistent with one’s beliefs. In the results
below we follow existing studies and report results based on Z = all of X. However, to
address this point, in robustness checks, we also estimated the model for various sets
of Z. We do not report the results given the sheer number of estimates, but we find
that the results are generally not sensitive to the specific choice of Z and that the
heteroskedasticity assumption is met, no matter what subset of X is chosen as Z.
A final point of which to be aware upfront is that the Lewbel (2012) estimates (based
on higher order moments) may not be as reliable as good conventional instruments
that meet all standard exclusion restrictions. As to the reliability of his method,
Lewbel (2012, p.67) states: The resulting identification is based on higher moments
and so is likely to produce less reliable estimates than identification based on standard
exclusion restrictions, but can be useful in applications where traditional instruments
are not available. When estimating relationships such as that between sexual activity
and happiness, this is important because the choice becomes one between speaking in
terms of association or trying to get at causality using estimates which albeit may not
be as reliable as those obtained if good conventional IVs were available.
Ideally, to give the reader added assurance one would like to point to examples in the
literature in which there is endogeneity and the Lewbel (2012) method has definitely
improved the answer. To do this, one would need to point to instances in which the
Lewbel (2012) method has been shown to work as well as a strong external IV. Given
8
that the Lewbel (2012) method is designed for, and is typically employed in, those
cases in which there are (potentially) weak instruments (see eg. Kelly et al., 2014;
Sabia, 2007a, 2007b) or no instruments at all (see eg Huang et al., 2009; Mishra &
Smyth, 2013, 2014; Mishra et al., 2014), such instances are hard to come by. The only
example, of which we are aware, is the illustration in Lewbel (2012). Lewbel (2012)
shows that the estimates with his method are very close to those using income as an
IV for expenditure when estimating an Engel Curve for food.
Data
We use data from the China Health and Family Life Survey (CHFLS), which was a
national probability survey of sexual behaviour, collected from China’s adult
population (aged 20-60) in 1999-2000 across all provinces, municipalities and
autonomous regions (excluding Hong Kong, Macau and Tibet). Respondents were
selected from 14 strata, 48 primary sampling units and 60 neighbourhoods with
probabilities proportional to their respective populations. Large cities and coastal
areas were oversampled. Altogether, 5000 people were identified to be in the sample.
Of these, 3,821 people (76.4 per cent) completed the survey. This response rate
compares favourably with the response rate in large-scale surveys of sexual behaviour
in other countries, such as France, the UK and the US (see Parish et al., 2007a). The
CHFLS is described in more detail in Parish et al. (2007a, 2007b).
To measure happiness, we use responses to the question: ‘Taken all together, how
would you say things are these days would you say that you are very happy, happy,
unhappy or very unhappy?’ This is similar to the single item indicator used in
Blanchflower and Oswald (2004) and in other studies in the economics of happiness
literature. It is generally accepted that in questions with the same, or similar wording
9
to this, the proportion of people providing wellbeing scores are fairly stable with most
people responding that they are ‘happy’. This was the case in this survey, with the
mean value for the dependent variable being just over 3 (on a 1-4 scale).
To measure sexual activity we explore a range of indicators including frequency of
sexual intercourse and orgasm, number of concurrent partners, sexual practices,
sexual satisfaction and the commodification of sex (eg. paying for sex, being paid for
sex and viewing pornography). These are obviously sensitive issues to ask about in
surveys. For this reason one might be concerned about the reliability of the data.
Theoretically, there could be upward or downward biases in the data. On the one
hand, a respondent might want to appear boastful about sexual conquests and report
more sexual partners than one has really had or report higher sexual satisfaction than
one really enjoyed, leading to over reporting. On the other hand, a respondent might
feel awkward revealing that they engage in certain sexual practices or view
pornography or might want to conceal extramarital affairs leading to under reporting.
To ensure the privacy of respondents, respondents in the CHFLS were not
interviewed at home, but were asked to attend secure neighbourhood facilities. The
neighbourhood facilities were typically either a private room in a local hotel in larger
locales or local meeting place in smaller locales, such as villages. The first half of the
interview involved a series of questions administered by the interviewer. In the
second half of the interview, the respondent entered responses to the questions about
sexual activities in a laptop containing the questionnaire. In most cases this was done
without any involvement from the interviewer. There were only 13 per cent of
respondents, who were primarily older women in rural locales, who needed assistance
entering their responses. The use of computer-assisted self-administered interviews
10
for the sensitive questions dealing with sexual activities should minimize the biases
associated with answering these questions. In any case, like Blanchflower and Oswald
(2004), we take the responses at face value in the analysis below.
Table 1 provides descriptive statistics for the sample as a whole, as well as broken
down by gender, for the dependent variable, the sexual activities we study and the
control variables typically employed in econometric happiness equations. A criticism
of Blanchflower and Oswald (2004) is that they did not control for the respondent’s
health, which is potentially an important correlate of both happiness and the extent to
which one participates in, and enjoys, a range of sexual activities (Wadsworth, 2014).
We control for health in each of the regression models we present below.
--------------------
Insert Table 1
----------------------
Results
In each of the tables we present results from the happiness equations above. For the
sake of brevity, we just report the results for the sexual activity variable(s). However,
all happiness equations included a full set of controls. In particular, it is important to
note that all happiness equations controlled for both age and age squared. The results
for the control variables were as expected and consistent with prior research for urban
China (see eg. Appleton & Song, 2008) and other parts of the world (see eg. Dolan et
al., 2008). In all cases we report both the OLS and Lewbel (2012) GMM estimates. In
the discussion we focus on the GMM estimates given they show causality.
2
--------------------
Insert Table 2
----------------------
2
In all GMM specifications Hansen’s J test statistic did not reject the null hypothesis that the Lewbel
(2012) instruments are valid.
11
Table 2 presents estimates for frequency of sexual intercourse. Panel A presents the
results for frequency of sexual intercourse with one’s primary partner in the past 12
months. Compared to the reference case (sexual intercourse at least once a day),
having infrequent, or no, sexual intercourse (once or less a month) lowers happiness
by 0.17 points for the full sample and for females and 0.16 points for males. These are
relatively large effects. To illustrate, these effects amount to roughly one quarter of
the size of being in very poor health, relative to being in very good health, on
happiness. Having sex two to three times a month also lowers happiness for the full
sample by 0.08 points and for males by 0.14 points, relative to the reference case.
Following existing studies (see e.g. Clark & Oswald, 2002; Levinson, 2012; Moore &
Shepherd, 2006), we estimate the shadow price of sexual intercourse at the mean
monthly income level. The shadow price for frequency of sexual intercourse is high.
For the full sample, an average person needs 1664 RMB (205 per cent of mean
monthly income) to offset the negative effect of having sex only two to three times a
month or 3670 RMB (452 per cent of mean monthly income) to offset the negative
effect of only having sex once or less a month, compared to once a day or more. For
males, having sex just two or three times a month or having sex just once or less a
month are equivalent to a loss of 226 per cent and 250 per cent of monthly mean
income respectively compared to the reference case. For males, having sex once or
less a month is equivalent to an income loss of 524 per cent of mean monthly income.
Panel B of Table 2 presents estimates for having sexual intercourse at least once a
week with one’s primary partner in the past 12 months. Having sex at least once a
week with one’s primary partner raises the happiness score by 0.07 points for the full
sample and 0.08 points for males, but has no effect for females. The shadow price of
12
having sex at least once a week with one’s primary partner is equivalent to 180 per
cent higher monthly income for the full sample and 116 per cent higher monthly
income for males. Panel C of Table 2 gives the results if the respondent stated that
the frequency of sex with one’s primary partner was just right. Having the
(subjective) frequency just right, increased the happiness score by 0.13 points for the
full sample and 0.15 points for males, but had no statistically significant effect on the
happiness of females in the sample. The shadow prices of having the frequency of
sexual intercourse just right are equivalent to 345 per cent higher monthly income for
the full sample and 236 per cent higher monthly income for males.
The results in Table 2 suggest several conclusions. First, those who have frequent
sexual intercourse with their primary partner (once or more a day) are happier than
those who have no, or infrequent, sexual intercourse with their primary partner.
Second, having regular sexual intercourse with one’s primary partner (at least once a
week) and being satisfied with the frequency of sexual intercourse raises happiness.
Third, the frequency of sex is more important for the happiness of men than women.
The latter result is consistent with the findings of several studies in the sexual
psychology literature that the frequency of sexual intercourse is more important for
men than women and that there is a higher association between frequency of sexual
intercourse and sexual satisfaction for men than women (see eg. McNulty & Fisher,
2008; Richters et al., 2003; Schwartz & Young, 2009). The sexual psychology
literature has speculated, without offering any conclusions, that this difference might
reflect biological differences between men and women (Schwartz & Young, 2009), be
related to differences in how men and women view relationships or reflect female
socialization (Gagnon & Simon, 1973; Giddens, 1992; Tiefer, 2004).
13
--------------------
Insert Table 3
----------------------
We interacted frequency of sex with income. Individuals with higher income have a
higher opportunity cost of leisure, which potentially reduces their satisfaction from
additional sex. The results are reported in Panel A of Table 3. Blanchflower and
Oswald (2004) found that income interacted with sexual frequency was insignificant.
However, we find some support for the idea that individuals with higher income get
less satisfaction from additional sex. Having sex at least once a week and having a
medium income, relative to the reference of low income, increases happiness, but the
interaction between having sex at least once a week and medium income has a
negative sign for the full sample as well as males and females separately.
We next interacted frequency of sexual intercourse with education. The results are
reported in Panel B of Table 3. Consistent with Blanchflower and Oswald (2004), we
find some evidence that sex brings more happiness to the better educated. For both the
full sample, and for males, those with a university degree get more happiness from
additional sex. As discussed by Rainer and Smith (2012), a potential explanation for
this result is that individuals with higher human capital can better articulate their
sexual preferences under conditions of incomplete information during sex. Happiness
derived from shared intimacy also depends on the ability to listen to, and learn about,
the sexual preferences of one’s partner. Individuals with higher human capital may be
better listeners and better able to learn by doing. Another possibility is that better
educated individuals have more liberal sexual values, which may translate into higher
satisfaction from additional sex for example through willingness to experiment.
These effects could be reinforced by assortative mating. Better-educated individuals
14
partner with other better educated individuals (Pencavel, 1998), potentially improving
communication in the bedroom and contributing to higher happiness.
------------------
Insert Table 4
----------------------
In Table 4 we build on the results in Panel A of Table 3 and examine the relationship
between frequency of masturbation and income. Given that people with high income
have a high opportunity cost of leisure, this may make them time sensitive and less
willing to invest time in developing a sexual relationship with a partner. This could be
manifest in either finding a partner or finding time to have sex with an existing
partner. Consequently, we expect people with high income to get more happiness
from the immediate, time sensitive, sexual gratification from masturbation. We find
some evidence consistent with this expectation. For the full sample, we find that
individuals with medium income get more happiness from masturbation than those
with low income and for females, those with both medium income and high income
derive more happiness from masturbation than those with low income.
------------------
Insert Table 5
----------------------
Table 5 reports results for the effect of the number of concurrent partners on
happiness. Panel A reports the results for having two or three concurrent partners on
happiness. Having two concurrent partners lowers the happiness score for the full
sample (0.08 points) and for females in the sample (0.19 points), while having three
concurrent partners lowers the happiness score of males by 0.19 points. In contrast,
having one partner increases the happiness score by 0.13 points for the full sample,
0.12 points for males and 0.15 points for females. To illustrate, the negative effects of
15
having more than one partner is roughly the same as the size of being in temporary
work (-0.12 to -0.14 points), relative to having a regular job, on happiness. For the
full sample, we estimated a compensating income equivalent of 1940 RMB and 2997
RMB for having two and three concurrent partners respectively; in contrast, having
just one partner is valued at 2959 RMB in terms of its effect on happiness.
Panel B reports the results if the respondent suspects that his/her primary partner is
having sexual intercourse with someone else. The three possible responses are
‘definitely yes’, ‘probably’ and ‘definitely no’. Treating the reference as ‘definitely
yes’, those who think that it is only probable that their partner is cheating and those
who are certain that their partner is not cheating, report higher happiness. Panel C
reports the results for extramarital sex. Engaging in extramarital sex lowers happiness
by 0.11 points for the full sample, 0.09 points for males and 0.25 points for females.
The compensating income equivalent is higher among females (6300 RMB) than
among the full sample (2232 RMB) or among males (1213 RMB).
The main conclusions from Table 5 are that having two or three concurrent partners
generally lowers happiness, suspecting that your primary partner is cheating lowers
happiness and engaging in extramarital sex lowers happiness. The latter finding is
consistent with the results reported in Blanchflower and Oswald (2004). Overall, the
results are consistent with the tenor of findings in the sexual psychology and
counselling literatures that monogamy is positively correlated with relationship
satisfaction and general wellbeing, at least among heterosexual couples (Atkins et al.
2001; Banfield & McCabe, 2001; Schwartz & Young, 2009).
--------------------
Insert Table 6
----------------------
16
Table 6 considers one outcome of sexual activity; namely, frequency of orgasm, on
happiness. Panel A reports results for the frequency of the respondent’s orgasm.
Higher frequency of orgasm increases happiness by 0.05 points for the full sample,
but has no effect for males and females separately. Panel B reports results for the
primary partner’s frequency of orgasm. If one’s partner has a higher frequency of
orgasm, this increases the happiness score for the full sample (0.04 points) and the
happiness score of men (0.04 points), but has no effect on the happiness score of
women. For the full sample, the income equivalent of experiencing a higher
frequency of orgasm on a four point scale is 1060 RMB (amounts to 130 per cent of
monthly income) and the income equivalent of having one’s partner experience a
higher frequency of orgasm is 1149 RMB (amounts to 142 per cent of monthly
income). Overall, for the full sample, the frequency with which one has an orgasm in
sexual intercourse and the frequency with which one’s primary partner experiences
orgasm has roughly the same effect on happiness. This result is consistent with the
findings in the sexual psychology literature that the frequency of one’s orgasms and
those of one’s primary partner are associated with sexual satisfaction among
heterosexual long-term couples (Haning et al., 2007; Young et al., 1998).
--------------------
Insert Table 7
----------------------
Table 7 examines the relevance of a variety of sexual practices on happiness. Several
sexual practices had no significant effect on happiness (genitals caressed by partner,
women on top, from behind, giving and receiving oral sex and anal sex). Kissing
raises happiness scores by 0.07 points for the full sample and 0.05 points for males,
with no significant effect for females. Caressing the female partner’s breast increases
17
the happiness scores of both male and female partners. Caressing one’s partner’s
genitals raises the happiness score of women (0.45 points), but lowers the happiness
score of men (0.04 points), although the latter is only weakly significant. Receiving
an unpleasant demand from one’s partner lowers happiness for the sample as a whole
(0.14 points) and for males (0.08 points) and for females (0.24 points).
--------------------
Insert Table 8
----------------------
Table 8 considers the effect of several aspects of sexual satisfaction with one’s
primary partner on happiness. Both emotional and physical satisfaction increases the
happiness score for the sample as a whole as well as each gender. The coefficients are
slightly larger for females than males in both cases. For the sample as a whole, having
enough foreplay, if one’s partner knows how to please and satisfy one during sex and
being thrilled by sex increase happiness by similar amounts (0.07-0.10 points). On the
other hand, again for the sample as a whole, feeling ashamed during sex, feeling dirty
about sex, feeling dirty about genital secretions during sex and fantasizing about
having sex with others during sex lowers happiness by a similar amount (0.07-0.10
points). These are consistent with previous findings in the sexual psychology
literature that various aspects of sexual satisfaction are associated with relationship
satisfaction and overall wellbeing (see eg. Abdolsalehi-Najafi & Beckman, 2013).
We do see some gender differences. Having enough foreplay, that one’s partner
knows how to please and satisfy one during sex, feeling ashamed during sex, being
thrilled by sex, feeling dirty about genital secretion and fantasizing about others
during sex has a significant effect on the happiness of men, but not women.
18
--------------------
Insert Table 9
----------------------
Table 9 shows the effect of partner affection on happiness. Panel A reports the results
for expressing affection for one’s primary partner, while Panel B reports the results
for receiving affection from one’s primary partner. The results for the sample as a
whole are insignificant, but we observe gender differences. Expressing and receiving
affection has no statistical effect on the happiness of men, but it increases the
happiness of women by 0.23 points. The positive effects of expressing, and receiving,
affection are bigger than the size of ln income (0.16 points) on happiness
amongwomen and amount to approximately one quarter of the size of being in very
good health (0.91 points), relative to being in very poor health, on happiness. In terms
of income equivalence, expressing affection and receiving affection increases an
average woman’s stated happiness by an amount equal to 8647RMB and 7256RMB
The findings in Tables 2, 8 and 9 point to some interesting gender differences. The
results in Tables 2 and 8 suggest that the frequency of sexual intercourse and
satisfaction from sexual intercourse are more important for the happiness of men than
women. The results in Table 9 suggest expressing, and receiving affection with one’s
partner are more important for overall happiness of women than men. Together, these
results are consistent with previous findings that frequency of sexual activities and
getting satisfaction from physical sexual acts is more important to, and is more
closely associated with, overall happiness or wellbeing for men than women. On the
other hand, the relationship between emotional intimacy and happiness for women is
more context-based and tied more broadly to feelings of affection for one’s partner
(McNulty & Fisher, 2008; Schwartz & Young, 2009; Traeen, 2010). Again, the
19
reasons for these gender differences in the sexual psychology literature are not
entirely clear, but may be partly biological and partly reflect the different manner in
which men and women conceive relationships and sexual intimacy.
--------------------
Insert Table 10
----------------------
Table 10 examines the effect of the commodification of sex on happiness. Panel A
examines the effect of paying for sex. Those who reported that they had paid for sex
had lower happiness scores (0.14 points for the full sample and 0.22 points for men).
Being paid for sex considerably lowered the happiness scores of women by 0.44
points, supporting the findings in Blanchflower and Oswald (2004). Having an erotic
massage in the previous 12 months had no effect on happiness for the overall sample,
but separately lowered the happiness of men (0.09 points) and women (0.30 points).
Viewing pornography had no statistical effect on happiness.
--------------------
Insert Table 11
----------------------
Table 11 reports the results for the effect of having premarital sex on happiness. For
the sample as a whole, the only significant variable is having premarital sex with a
commercial sex partner, which lowers happiness by 0.18 points. This finding is
consistent with the result for paying for sex in Table 10. Turning to the separate
regression results for males and females, for males having premarital sex with
someone whom you do not later marry increases happiness by 0.15 points, while for
females having premarital sex with a short-term partner or commercial sex partner
lowers happiness by 0.28 and 0.98 points respectively. These results are entirely
consistent with the gender differences identified in earlier tables. Men derive
satisfaction from premarital sex with partners they do not later marry, while for
20
women, premarital sex with short-term partners has a negative effect on happiness,
presumably because such relationships lack sustained emotional intimacy.
--------------------
Insert Table 12
----------------------
Table 12 reports the results for having unwilling sex on happiness. Panel A reports the
results for having unwanted sex. The precise question was: ‘Up till now, has it
occurred that you were unwilling, but still had to concede to having sex with
someone? For the full sample, those who answered yes to this question had lower
happiness scores (0.14 points). Having unwilling sex lowered the happiness of women
(0.15 points), but had no effect on the happiness of men. Panel B reports the results
for having had sex only to satisfy one’s partner in the previous 12 months. Having
had sex only to satisfy one’s partner lowers happiness for the full sample (0.07 points)
and for men (0.08 points), but has no significant effect on the happiness of women.
Conclusion
There is a sizeable economics of happiness literature, but the effect of sexual
activities and sexual behaviour on happiness has been relatively underexplored in the
economics literature. Yet, sexual behaviour and activities are an important component
of people’s lives and can be addressed with standard econometric tools. Blanchflower
and Oswald (2004) is the only study, of which we are aware, in the economics
literature that directly tackles the relationship between sexual activity and happiness.
A limitation of that study is that the authors do not address the endogeneity of sexual
activities. Thus, they are unable to identify if sexual activities cause happiness or
vice-versa. The fundamental problem is that it is very difficult, if not impossible, to
21
think of an appropriate IV. We address this issue using the Lewbel (2012) method to
construct an internal IV, based on the heteroskedasticity in the data.
Several conclusions emerge from the results. The first is that having sexual
intercourse has a positive effect on happiness. Those who had regular sex (at least
once a week) with a primary partner in the past 12 months are happier than those who
do not and those who had frequent sex (once a day or more) with a primary partner in
the past 12 months are happier than those who have infrequent, or no, sex. Those
whose subjective assessment of the frequency with which they had sex was just right
were happier than those who thought they had too much, or too little, sex.
A second finding is that there is some evidence that human capital and income has an
effect on happiness from sex. Individuals with higher income derive more happiness
from masturbation, and less happiness from more frequent sex, presumably reflecting
their higher opportunity cost of time allocated to leisure. Highly educated individuals
derive more happiness from more frequent sex, which may be a reflection of being
better communicators in the bedroom and better able to learn by doing.
A third finding is that in addition to the frequency of sex, the quality of sex also
matters. Using frequency of orgasm as a proxy for quality, individuals who
experience more frequent orgasms are happier. Similarly, those who are more
emotionally and physically satisfied with their primary sexual partner are happier.
A fourth finding is that the optimal number of sexual partners is one. Being in a
committed, long-lasting relationship with one sexual partner has a positive effect on
22
happiness, relative to having two or three concurrent partners. Similarly, engaging in
extra-marital sex or being paid for sex with multiple partners lowers happiness.
A fifth finding is that there are important gender differences between men and
women. For men, there is a stronger relationship than for women, between the
frequency and physical aspects of sexual intercourse and happiness. In particular,
there is a stronger relationship between the frequency of sex and happiness for men
than women. There is also a stronger relationship between several aspects of sexual
satisfaction relating to the sexual intercourse and happiness for men than women. This
extends to one of the main outcomes of sexual intercourse; namely, partner’s
frequency of orgasm. For women, the relationship between emotional and sexual
intimacy and happiness appears to be more contextual. For women, there is a stronger
relationship than for men between giving, and receiving, affection and happiness.
Importantly, our results confirm several of the main substantive findings in
Blanchflower and Oswald (2004), but do so for a country other than the United States.
While more research needs to be done to be more definitive, our results, together with
those of Blanchflower and Oswald (2004), are at least suggestive that the relationship
between many aspects of sexual activity and happiness are universal.
23
References
Abdolsalehi-Najafi, E. and Beckman, L.J. (2013). Sex guilt and life satisfaction in
Iranian-American women. Archives of Sexual Behavior, 42, 1063-1071.
Appleton, S. and Song, L. (2008). Life satisfaction in urban China: Components and
determinants. World Development, 36, 2325-2340.
Atkins, D.C., Baucom, D.H. and Jacobson, N.S. (2001). Understanding infidelity:
Correlates in a national random sample. Journal of Family Psychology, 15,
735-749.
Banfield, S. and McCabe, M. (2001). Extra relationship involvement among women:
Are they different from men? Archives of Sexual Behavior, 30, 119-140.
Blanchflower, D. and Oswald, A. (2004). Money, sex and happiness: An empirical
study. Scandinavian Journal of Economics, 106, 393-415.
Breusch, T.S. and Pagan, A.R. (1979). Simple test for heteroscedasticity and random
coefficient variation". Econometrica ,47 (5), 12871294.
Clark, A. E., Frijters, P. and Shields, M.A. (2008). Relative income, happiness and
utility: An explanation for the Easterlin paradox and other puzzles. Journal of
Economic Literature, 46, 95-144.
Clark, A. E. and Oswald, A. J. (2002). A simple statistical method for measuring how
life events affect happiness. International Journal of Epidemiology, 31, 1139-
1144.
Dogan, T., Tugut, N. and Golbasi, Z. (2013). The relationship between sexual quality
of life, happiness and satisfaction with life in married Turkish women. Sex
Disability, 31, 239-247.
Dolan, P., Peasgood, T. and White, M. (2008). Do we really know what makes us
happy? A review of the economic literature on the factors associated with
subjective wellbeing. Journal of Economic Psychology, 29, 94-122.
Doran, K. and Price, J. (2014). Pornography and marriage. Journal of Family and
Economic Issues (in press).
Drydakis, N. (2013). The effect of sexual activity on wages. IZA Discussion Paper
No. 7529.
Ferrer-i-Carbonell, A. and Frijters, P. (2004). How important is methodology for the
estimates of the determinants of happiness? Economic Journal, 114, 641-659.
Frey, B. and Stutzer, A. (2002). Happiness and Economics. Princeton NJ: Princeton
University Press.
Gagnon, J. and Simon, W. (1973). Sexual Conduct: The Social Sources of Human
Sexuality. Chicago: Aldine.
Giddens, A. (1992). The Transformation of Intimacy: Sexuality, Love and Eroticism
in Modern Societies. Palo Alto, CA: Stanford University Press.
Haning, R.V., O’Keefe, S.L., Randall, E.J., Kommor, M.J., Baker, E. and Wilson, R.
(2007). Intimacy, orgasm likelihood and conflict predict sexual satisfaction in
heterosexual male and female respondents. Journal of Sex and Marital
Therapy, 33, 93-113.
Huang, H-C., Lin, Y-C. and Yeh, C-C. (2009). Joint determination of inequality and
growth. Economics Letters, 103, 163-166.
Islam, A. and Smyth, R. (2012) The economic returns to good looks and risky sex in
the Bangladesh commercial sex market. The B.E. Journal of Economic
Analysis & Policy (Contributions) (2012) 12(1) (Article 22).
Kahneman, D., Krueger, A., Schkade, D.A., Schwarz, N., Stone, A.A. (2004). A
survey method for characterizing daily life experience: the day reconstruction
method. Science, 306, 1776-1780.
24
Kelly, I.R., Dave, D.M., Sindelar, J.L., Gallo, W.T. (2014). The impact of early
occupational choice on health behaviors. Review of the Economics of the
Household (in press).
Laumann, E.O., Paik, A., Glaser, D.B, Kang, J-H., Wang, T., Levinson, B., Moreira,
E.D., Nicolosi, A. and Gingell, C. (2006). A cross-national study of subjective
sexual well-being among older women and men: Findings from the global
study of sexual attitudes and behaviors. Archives of Sexual Behavior, 35, 145-
161.
Levinson, A. (2012). Valuing public goods using happiness data: The case of air
quality. Journal of Public Economics, 96, 869-880.
Lewbel, A. (2012). Using heteroscedasticity to identify and estimate mismeasured and
endogenous regressor models. Journal of Business & Economic Statistics,
30(1), 67-80.
McNulty, J.K. and Fisher, T.D. (2008). Gender differences in response to sexual
expectancies and changes in sexual frequencies: A short-term longitudinal
study of sexual satisfaction in newly married couples. Archives of Sexual
Behavior, 37, 229-240.
Mishra, V. Nielsen, I., Smyth, R. and Newman, A. (2014). The job satisfaction-life
satisfaction relationship revisited: Using the Lewbel estimation technique to
estimate causal effects using cross-sectional data. Department of Economics
Monash University Discussion Paper No. 26/14.
Mishra, V. and Smyth, R. (2013). Are more senior academics really more research
productive than junior academics? Evidence from Australian law schools.
Scientometrics, 96(2), 411-425.
Mishra, V. and Smyth, R. (2014). Workplace policies and training in China: Evidence
from matched employee-employer data. International Journal of Manpower (in
press).
Monk-Turner, E. and Turner, C. (2010). Subjective well-being among those who
exchange sex and money, Yunnan, China and Thailand. Social Indicators
Research, 99 (1), 13-23.
Moore S. and Shepherd, J.P. (2006). The cost of fear: shadow pricing the intangible
costs of crime. Applied Economics, 38(3), 293-300.
Ng, Y-K. (1997). A case for happiness, cardinalism and interpersonal comparability.
Economic Journal, 107, 1848-1858.
Ng, Y-K. (2008). Environmentally responsible happy nation index. Toward an
internationally acceptable national success indicator. Social Indicators
Research, 85, 425-446.
Parish, W.L., Luo, Y., Stolzenberg, R., Laumann, E.O., Farrer, G. and Pan, S.
(2007a). Sexual practices and sexual satisfaction: A population based study of
Chinese urban adults. Archives of Sexual Behavior, 36, 5-20.
Parish, W.L., Laumann, E.O. and Mojola, S.A. (2007b). Sexual behavior in China:
Trends and comparisons. Population and Development Review, 33(4), 729-
756.
Pencavel, J. (1998). Assortative mating by schooling and the work behavior of wives
and husbands. American Economic Review, 88(2), 326-329.
Rainer, H. and Smith, I. (2012), Education, communication, and wellbeing: an
application to sexual satisfaction. Kyklos, 65(4), 581-598.
Richters, J., Grulich, A.E., Visser, R.O, Smith, A.M.A. and Rissel, C.E. (2003). Sex
in Australia: Sexual and emotional satisfaction in regular relationships and
25
preferred frequency of sex among a representative sample of adults. Australian
and New Zealand Journal of Public Health, 27, 171-179.
Sabia, J. (2007a). Early adolescent sex and diminished school attachment: Selection
or spillovers? Southern Economic Journal, 74(1), 239-268.
Sabia, J. (2007b). The effect of body weight on adolescent academic achievement.
Southern Economic Journal, 73(4), 871-900.
Schwartz, P. and Young, L. (2009). Sexual satisfaction in committed relationships.
Sexuality Research and Social Policy, 6(1), 1-17.
Tiefer, L. (2004). Sex is Not a Natural Act and Other Essays. Boulder, CO.:
Westview Press.
Traeen, B. (2010). Sexual dissatisfaction among heterosexual Norwegians in couple
relationships. Sexual and Relationship Therapy, 25(2), 132-147.
Wadsworth, T. (2014). Sex and the pursuit of happiness: How other people’s sex lives
are related to our sense of well-being. Social Indicators Research (in press).
Woloski-Wruble, A.C., Oliel, Y., Leefsma, M. and Hochner-Celnikier, D. (2010).
Sexual activities, sexual and life satisfaction and successful aging in women.
Journal of Sexual Medicine, 2401-2410.
Yamamura, E. (2014). Smokers’ sexual behaviour and their satisfaction with family
life. Social Indicators Research (in press).
Young, M., Denny, G., Luquis, R. and Young, T. (1998). Correlates of sexual
satisfaction in marriage. Canadian Journal of Human Sexuality, 7, 115-127.
26
Table 1 Descriptive statistics
All
Female
Gender (%)
49.81
Age
38.91
37.78
Age squared
1513.99
1427.33
Marital status (%)
Never married/de facto
9.21
7.17
First married
84.92
87.59
Remarried
2.52
2.08
De facto
0.68
0.44
Divorced
0.79
0.66
Widowed
1.88
2.06
Schooling (years)
7.60
6.95
ln monthly income
5.38
4.98
Health (%)
Very poor
0.67
0.38
Poor
11.44
11.90
Neutral
21.34
22.14
Good
38.66
37.72
Very good
27.89
27.86
Employment (%)
Never worked
1.26
2.11
Unemployed
8.37
12.02
Retired
3.54
3.69
Temporary work
12.12
11.87
Regular work
74.7
70.31
Urban (%)
28.97
27.80
Region (%)
South coast
4.91
4.95
East coast
6.46
6.33
Inland
30.21
29.53
North
11.21
10.66
Northeast
12.83
12.74
Central west
34.38
35.79
Happiness (1: very unhappy; 2: unhappy; 3: happy; 4: very happy)
3.03
3.01
Frequency of sex with primary partner in the past 12 months (%)
Once a day or more
1.86
1.93
Three to six times a week
7.05
5.34
One to two times a week
41.99
42.07
Two to three times a month
30.73
32.05
Once or less a month
18.38
18.60
Frequency of sex is just right with primary partner in the past 12 months (%)
60.53
67.76
Have sex at least once a week with primary sexual partner in the past 12 months (%)
50.89
49.34
Masturbate once a week or more (%)
4.24
1.80
Has one primary sexual partner (%)
89.38
90.73
Have two concurrent sexual partners (%)
7.82
4.09
Have three concurrent sexual partners (%)
1.78
0.14
Primary partner has another concurrent partner (%)
Definitely yes
2.01
2.07
Probably
9.63
10.57
Definitely no
88.36
87.36
Have extramarital sex (%)
11.3
5.26
Have premarital sex (%)
No
97.28
98.78
Yes, and then married
0.14
0.12
Yes, and now engaged
1.14
0.67
Yes, but did not marry
0.37
0.20
Yes, with short-term sexual partner
0.34
0.17
Yes, with commercial sexual partner
0.72
0.056
27
Pay for sex (%)
3.45
0.59
Get paid for sex (%)
0.48
0.10
Have erotic massage in the past 12 months (%)
4.53
0.12
Have viewed pornography in the past 12 months (%)
30.73
21.6
Have unwanted sex (%)
16.80
25.37
Have sex only to satisfy partner in the past 12 months
(0: never; 1: rarely; 2: sometimes; 3: often)
0.95
1.07
Frequency of orgasm
(0: never; 1: rarely; 2: sometimes; 3: often; 4: always)
2.49
2.06
Primary partner’s frequency of orgasm (answered by respondent)
(0: never; 1: rarely; 2: sometimes; 3: often; 4: always)
2.46
2.70
Sexual practices with primary partner (0: never; 1: sometimes; 2: often)
Kiss
1.00
0.90
Caress partner’s breast
Breast caressed by partner
1.09
Caress partner’s genitals
0.85
0.66
Genitals caressed by partner
0.92
0.87
Woman on top
0.52
0.43
From behind (‘doggy style’)
0.41
0.35
Give oral sex to partner
0.18
0.13
Oral sex provided by partner
0.21
0.21
Anal sex
0.29
0.04
Unpleasant demand by partner
0.18
0.24
Sexual satisfaction with primary partner
Physical satisfaction
(0: very dissatisfied; 1: dissatisfied; 2: satisfied; 3: very satisfied)
2.08
1.99
Emotional satisfaction
(0: very dissatisfied; 1: dissatisfied; 2: satisfied; 3: very satisfied)
2.15
2.08
Foreplay
(0: none; 1: not enough; 2: enough)
1.31
1.37
Partner knows how to please and satisfy respondent in sex
(0: never; 1: rarely; 2: sometimes; 3: often)
1.88
1.95
Feel ashamed during sex
(0: never; 1: sometimes; 2: often)
0.25
1.32
Thrilled by sex
(0: never; 1: sometimes; 2: often)
1.18
1.07
Feel dirty about sex
(0: never; 1: sometimes; 2: often)
0.41
0.49
Feel dirty about genital secretion, semen or menstrual blood during sex
(0: never; 1: sometimes; 2: often)
0.60
0.65
Fantasize about having sex with others during sex
(0: never; 1: sometimes; 2: often)
0.12
0.07
Express affection for primary partner
(0: no affection anymore; 1: not too deep; 2: fairly deep; 3: very deep)
2.35
2.33
Received affection from primary partner (answered by respondent)
(0: no affection anymore; 1: not too deep; 2: fairly deep; 3: very deep)
2.41
2.40
Survey design
Number of strata: 14
Number of primary survey units: 48
Number of observations: 3,821
Population size: 797,128,593
28
Table 2 Estimates for frequency of sex from happiness equations
All
Male
Female
Panel A: Frequency of sex with primary partner in the past 12 months
(Ref: once a day or more)
OLS
Three to six times a week
-0.0637
(-0.72)
-0.175
(-1.45)
0.0489
(0.38)
One to two times a week
-0.0602
(-0.72)
-0.154
(-1.32)
0.0294
(0.25)
Two to three times a month
-0.103
(-1.23)
-0.220*
(-1.86)
0.00570
(0.05)
Once or less a month
-0.194**
(-2.24)
-0.266**
(-2.18)
-0.124
(-1.01)
GMM
Three to six times a week
-0.0525
(-1.13)
-0.0930
(-1.35)
-0.0293
(-0.44)
One to two times a week
-0.0422
(-1.07)
-0.0801
(-1.38)
-0.0765
(-1.20)
Two to three times a month
-0.0779*
(-1.89)
-0.138**
(-2.24)
-0.0805
(-1.44)
Once or less a month
-0.173***
(-3.68)
-0.153**
(-2.25)
-0.168**
(-2.17)
Valuation (RMB)
Three to six times a week
n.s.
n.s.
n.s.
One to two times a week
n.s.
n.s.
n.s.
Two to three times a month
-1664
-2300
n.s.
Once or less a month
-3670
-2544
-3195
N
3137
1581
1556
Panel B: Have sex at least once a week with primary partner in the past 12 months
(Ref: No)
OLS
0.0688***
(3.25)
0.0783***
(2.64)
0.0599**
(1.99)
GMM
0.0661***
(3.26)
0.0773**
(2.70)
-0.0331
(-0.40)
Valuation (RMB)
1459
1184
n.s.
N
3137
1581
1556
Panel C: Frequency of sex is just right with primary partner
Ref: too much/few
OLS
0.124***
(6.21)
0.157***
(5.73)
0.0907***
(3.14)
GMM
0.125***
(6.34)
0.151***
(5.58)
-0.178
(-0.47)
Valuation (RMB)
2798
2401
n.s.
N
3136
1580
1556
Notes: t and z statistics in parentheses for OLS and GMM respectively; * p < 0.10, ** p < 0.05, *** p <
0.01; n.s. not significantly different from zero; All regressions include a full set of controls as given in
Table 1.
Mean monthly incomes are 812 RMB for full sample, 1017 RMB for males, and 610 RMB for females.
29
Table 3: Estimates for education, income and frequency of sex with primary partner from
happiness equations
All
Male
Female
Panel A: Income and frequency of sex
OLS
Have sex at least once a week
0.0831***
(3.69)
0.104***
(3.18)
0.0666**
(2.13)
Medium income
0.200***
(3.63)
0.208***
(3.03)
0.193**
(2.10)
High income
0.0710
(0.95)
0.128
(1.51)
-0.0614
(-0.39)
Have sex at least once a week × medium
income
-0.133*
(-1.89)
-0.119
(-1.37)
-0.175
(-1.43)
Have sex at least once a week × high income
-0.0143
(-0.15)
-0.0934
(-0.86)
0.219
(1.07)
GMM
Have sex at least once a week
0.0822***
(3.72)
0.0948***
(2.94)
0.0712**
(2.41)
Medium income
0.201***
(4.52)
0.216***
(3.80)
0.166***
(2.80)
High income
0.0666
(1.13)
0.110**
(1.96)
0.0479
(0.32)
Have sex at least once a week × medium
income
-0.157***
(-2.83)
-0.151**
(-2.17)
-0.127*
(-1.76)
Have sex at least once a week × high income
-0.0277
(-0.38)
-0.0290
(-0.41)
0.0976
(0.53)
N
3137
1581
1556
Panel B: Education and frequency of sex
OLS
Have sex at least once a week
0.0630***
(2.92)
0.0723**
(2.36)
0.0582*
(1.91)
University degree
-0.117*
(-1.68)
-0.106
(-1.25)
-0.0926
(-0.74)
Have sex at least once a week × university
degree
0.192**
(2.12)
0.214**
(1.97)
0.127
(0.78)
GMM
Have sex at least once a week
0.0621***
(2.98)
0.0707**
(2.41)
0.0638
(1.16)
University degree
-0.0773
(-1.54)
-0.0415
(-0.72)
-0.0198
(-0.27)
Have sex at least once a week × university
degree
0.144**
(2.34)
0.141**
(2.06)
0.0188
(0.19)
N
3137
1581
1556
Notes: t and z statistics in parentheses for OLS and GMM respectively; * p < 0.10, ** p < 0.05, *** p <
0.01; All regressions include a full set of controls as given in Table 1 (excluding ln income and
schooling.
Mean monthly income is 812 RMB. Low, medium and high incomes are defined as monthly income
lower than 1000 RMB, between 1000 and 2000 RMB, and higher than 2000 RMB respectively. Low
income is the reference group.
30
Table 4 Estimates for income and frequency of masturbation from happiness equations
All
Male
Female
OLS
Masturbate once a week or more
-0.0426
(-0.91)
-0.0294
(-0.55)
-0.0705
(-0.76)
Medium income
0.115***
(3.27)
0.118***
(2.75)
0.109*
(1.82)
High income
0.0685
(1.46)
0.0978*
(1.80)
0.0273
(0.30)
Masturbate once a week or more × medium
income
0.199
(1.42)
0.207
(1.35)
0.316
(0.94)
Masturbate once a week or more × high income
-0.0148
(-0.08)
-0.154
(-0.84)
0.980*
(1.73)
GMM
Masturbate once a week or more
-0.0517
(-1.36)
-0.00132
(-0.03)
-0.0570
(-0.88)
Medium income
0.117***
(3.95)
0.122***
(3.28)
0.0350
(0.90)
High income
0.0658*
(1.67)
0.0969**
(2.13)
0.0110
(0.14)
Masturbate once a week or more × medium
income
0.195**
(2.30)
0.134
(1.33)
0.488***
(3.08)
Masturbate once a week or more × high income
-0.00689
(-0.10)
-0.0266
(-0.30)
0.985***
(9.90)
N
3732
1844
1888
Notes: t and z statistics in parentheses for OLS and GMM respectively; * p < 0.10, ** p < 0.05, *** p <
0.01; All regressions include a full set of controls as given in Table 1 (excluding ln income).
Mean monthly income is 812 RMB. Low, medium and high incomes are defined as monthly income
lower than 1000, between 1000 and 2000, and higher than 2000 RMB respectively. Low income is the
reference group.
31
Table 5 Estimates for concurrent partners from happiness equations
All
Male
Female
Panel A: Have concurrent partners
OLS
One partner
0.141***
(4.26)
0.121**
(2.56)
0.130***
(2.81)
Two concurrent partners
-0.0818***
(-2.67)
-0.0595*
(-1.68)
-0.133**
(-2.22)
Three concurrent partners
-0.149***
(-2.73)
-0.144**
(-2.50)
-0.251
(-1.56)
GMM
One partner
0.128**
(2.45)
0.121*
(1.89)
0.152**
(2.20)
Two concurrent partners
-0.0841*
(-1.69)
-0.0500
(-1.06)
-0.190***
(-2.62)
Three concurrent partners
-0.130**
(-2.37)
-0.188***
(-3.28)
-0.0206
(-0.20)
Valuation (RMB)
One partner
2959
1953
3434
Two concurrent partners
-1940
n.s.
4282
Three concurrent partners
-2997
-3035
n.s.
N
3799
1889
1910
Panel B: Primary partner has another concurrent partner
(Ref: definitely yes)
OLS
Probably
0.198***
(3.25)
0.112
(1.20)
0.262***
(3.25)
Definitely no
0.306***
(5.35)
0.260***
(2.97)
0.346***
(4.57)
GMM
Probably
0.234**
(2.31)
-0.0280
(-0.31)
0.175**
(2.21)
Definitely no
0.320***
(3.66)
0.139**
(2.01)
0.239***
(3.07)
Evaluation (RMB)
Probably
5434
n.s.
4790
Definitely no
7436
1996
6552
N
3339
1665
1674
Panel C: Have extramarital sex
(Ref: no)
OLS
-0.109***
(-3.97)
-0.0746**
(-2.40)
-0.207***
(-3.72)
GMM
-0.110***
(-2.58)
-0.0949*
(-1.82)
-0.251***
(-3.96)
Evaluation (RMB)
-2232
-1213
-6300
N
3288
1604
1684
Notes: t and z statistics in parentheses for OLS and GMM respectively; * p < 0.10, ** p < 0.05, *** p <
0.01; n.s. not significantly different from zero; All regressions include a full set of controls as given in
Table 1.
32
Table 6 Estimates for frequency of orgasm from happiness equations
All
Male
Female
Panel A: Respondent’s frequency of orgasm
OLS
0.0472***
(4.24)
0.0244
(1.57)
0.0649***
(4.09)
GMM
0.0455***
(4.11)
0.0194
(1.25)
0.0703
(0.83)
Evaluation (RMB)
1060
n.s.
n.s.
N
3124
1573
1551
Panel B: Primary partner’s frequency of orgasm
OLS
0.0443***
(3.88)
0.0454***
(2.85)
0.0474***
(2.90)
GMM
0.0448***
(3.91)
0.0427***
(2.74)
0.0262
(0.29)
Evaluation (RMB)
1149
644
n.s.
N
2747
1368
1379
Notes: t and z statistics in parentheses for OLS and GMM respectively; * p < 0.10, ** p < 0.05, *** p <
0.01; n.s. not significantly different from zero; All regressions include a full set of controls as given in
Table 1.
33
Table 7 Estimates for sexual practices with primary partner in the past 12 months from
happiness equations
All
Male
Female
Panel A: Kiss
OLS
0.0640***
(4.04)
0.0410*
(1.83)
0.0829***
(3.72)
GMM
0.0659***
(4.29)
0.0509**
(2.32)
-0.0471
(-0.33)
N
3131
1579
1552
Panel B: Caress partner’s breast
OLS
0.0345
(1.47)
GMM
0.0388*
(1.73)
N
1577
Panel C: Breast caressed by partner
OLS
0.0628***
(2.75)
GMM
0.0119
(0.08)
N
1550
Panel D: Caress partner’s genitals
OLS
0.0166
(1.07)
-0.0399*
(-1.87)
0.0720***
(3.16)
GMM
0.0197
(1.28)
-0.0386*
(-1.86)
0.452**
(2.06)
N
3126
1576
1550
Panel E: Genitals caressed by partner
OLS
0.0172
(1.06)
0.00900
(0.39)
0.0241
(1.05)
GMM
0.0157
(1.00)
0.0143
(0.64)
0.0143
(0.08)
N
3126
1577
1549
Panel F: Woman on top
OLS
-0.00465
(-0.28)
-0.0395*
(-1.71)
0.0261
(1.12)
GMM
0.00271
(0.17)
-0.0344
(-1.58)
-0.0615
(-0.80)
N
3123
1573
1550
Panel G: From behind (‘doggy style’)
OLS
-0.0136
(-0.75)
0.00532
(0.21)
-0.0338
(-1.29)
GMM
-0.00397
(-0.23)
0.00177
(0.08)
0.0272
(0.28)
N
3124
1573
1551
Panel H: Give oral sex to partner
OLS
0.0237
(1.12)
-0.0170
(-0.61)
0.0690**
(2.14)
GMM
0.0183
(0.97)
-0.0292
(-1.16)
0.0639
(1.28)
N
3124
1573
1551
Panel I: Oral sex provided by partner
OLS
0.0157
(0.76)
0.00352
(0.12)
0.0291
(0.99)
GMM
0.0153
(0.80)
0.00101
(0.04)
0.00205
(0.04)
N
3127
1577
1550
Panel J: Anal sex
OLS
-0.0616
(-1.07)
-0.154*
(-1.70)
0.0101
(0.14)
GMM
-0.0309
(-0.74)
-0.0649
(-1.24)
0.0145
(0.25)
N
3127
1577
1550
Panel K: Unpleasant demand by
partner
OLS
-0.146***
(-5.80)
-0.106***
(-2.61)
-0.171***
(-5.31)
GMM
-0.142***
(-5.96)
-0.0802**
(-2.37)
-0.243***
(-2.78)
N
3128
1576
1552
Notes: t and z statistics in parentheses for OLS and GMM respectively; * p < 0.10, ** p < 0.05, *** p <
0.01; All regressions include a full set of controls as given in Table 1.
34
Table 8 Estimates for sexual satisfaction with primary sexual partner from happiness equations
All
Male
Female
Panel A: Physical satisfaction
OLS
0.156***
(9.84)
0.175***
(7.73)
0.143***
(6.46)
GMM
0.150***
(8.75)
0.179***
(7.60)
0.254**
(2.17)
N
3130
1578
1552
Panel B: Emotional satisfaction
OLS
0.176***
(11.09)
0.204***
(8.97)
0.151***
(6.89)
GMM
0.179***
(10.56)
0.209***
(8.78)
0.247*
(1.70)
N
3130
1579
1551
Panel C: Enough
foreplay
OLS
0.105***
(7.23)
0.111***
(5.35)
0.100***
(4.90)
GMM
0.102***
(6.92)
0.114***
(5.55)
0.0166
(0.19)
N
3132
1578
1554
Panel D: Partner knows how to please and satisfy respondent during sex
OLS
0.0716***
(5.25)
0.0850***
(4.50)
0.0562***
(2.86)
GMM
0.0728***
(5.19)
0.0878***
(4.51)
-0.0695
(-0.83)
N
3127
1576
1551
Panel E: Feel ashamed during sex
OLS
-0.118***
(4.97)
-0.0948**
(2.51)
-0.130***
(4.23)
GMM
-0.103***
(4.07)
-0.0817**
(2.19)
-0.0547
(0.71)
N
3132
1578
1554
Panel F: Thrilled by sex
OLS
0.0821***
(4.90)
0.0340
(1.51)
0.142***
(5.68)
GMM
0.0840***
(5.18)
0.0397*
(1.83)
0.251
(1.30)
N
3130
1576
1554
Panel G: Feel dirty about sex
OLS
-0.0747***
(4.15)
-0.0427
(1.52)
-0.101***
(4.28)
GMM
-0.0694***
(3.74)
-0.0152
(0.58)
0.0260
(-0.29)
N
3131
1577
1554
Panel H: Feel dirty about genital secretion, semen or menstrual blood during sex
OLS
-0.0745***
(4.90)
-0.0706***
(3.16)
-0.0749***
(3.60)
GMM
-0.0708***
(4.63)
-0.0504**
(2.44)
0.0411
(-0.25)
N
3129
1576
1553
Panel I: Fantasize about having sex with others during sex
OLS
-0.0745***
(4.90)
-0.0706***
(3.16)
-0.0749***
(3.60)
GMM
-0.0708***
(4.63)
-0.0504**
(2.44)
0.0411
(-0.25)
N
3129
1576
1553
Notes: t and z statistics in parentheses for OLS and GMM respectively; * p < 0.10, ** p < 0.05, *** p <
0.01; All regressions include a full set of controls as given in Table 1.
35
Table 9 Estimates for partner affection from happiness equations
All
Male
Female
Panel A: Express affection for primary partner
OLS
0.194***
(14.12)
0.154***
(7.82)
0.229***
(11.99)
GMM
0.126
(1.63)
0.0160
(0.19)
0.228***
(11.47)
Evaluation (RMB)
n.s.
n.s.
8647
N
3351
1671
1680
Panel B: Receive affection from primary partner
OLS
0.202***
(14.29)
0.177***
(8.80)
0.221***
(11.21)
GMM
0.126
(1.63)
0.0160
(0.19)
0.228***
(11.47)
Evaluation (RMB)
n.s.
n.s.
7256
N
3350
1670
1680
Notes: t and z statistics in parentheses for OLS and GMM respectively; * p < 0.10, ** p < 0.05, *** p <
0.01; n.s. not significantly different from zero; All regressions include a full set of controls as given in
Table 1.
36
Table 10 Estimates for commodification of sex from happiness equations
All
Male
Female
Panel A: Paid for sex
OLS
-0.116**
(-2.53)
-0.126***
(-2.75)
-0.000586
(-0.00)
GMM
-0.137***
(-2.95)
-0.220***
(-3.83)
0.0618
(1.20)
N
3760
1863
1897
Panel B: Been paid for sex
OLS
-0.0581
(-0.62)
0.00673
(0.07)
-0.455*
(-1.84)
GMM
0.0458
(0.59)
0.0801
(1.33)
-0.444***
(-7.19)
N
3738
1846
1892
Panel C: Had erotic massage in the past 12 months
OLS
-0.0462
(-1.20)
-0.0417
(-1.06)
-0.310
(-1.58)
GMM
-0.0284
(-0.70)
-0.0918*
(-1.86)
-0.296***
(-3.51)
N
3747
1851
1896
Panel D: Viewed pornography in the past 12 months
OLS
-0.0365*
(-1.79)
-0.0554**
(-2.06)
-0.0106
(-0.34)
GMM
-0.0821
(-1.41)
-0.0870
(-0.99)
-0.0157
(-0.29)
N
3747
1851
1896
Notes: t and z statistics in parentheses for OLS and GMM respectively; * p < 0.10, ** p < 0.05, *** p <
0.01; All regressions include a full set of controls as given in Table 1.
37
Table 11: Estimates for premarital sex from happiness equations
All
Male
Female
Have premarital sex
(Ref: no)
OLS
Yes, and then married
0.0943
(0.60)
0.0484
(0.22)
0.205
(0.93)
Yes, and now engaged
-0.0514
(-0.70)
-0.0364
(-0.38)
-0.0937
(-0.80)
Yes, but did not marry
0.0623
(0.75)
0.137
(1.31)
-0.0667
(-0.49)
Yes, with short-term sexual
partner
-0.0608
(-0.59)
0.0724
(0.60)
-0.372*
(-1.95)
Yes, with commercial sexual
partner
-0.155*
(-1.74)
-0.0961
(-1.00)
-0.728**
(-2.28)
GMM
Yes, and then married
0.131
(1.36)
-0.0330
(-0.18)
0.0308
(0.18)
Yes, and now engaged
-0.00868
(-0.12)
0.0228
(0.26)
-0.0795
(-1.27)
Yes, but did not marry
0.0518
(1.17)
0.147**
(2.38)
-0.0607
(-1.25)
Yes, with short-term sexual
partner
-0.0994
(-1.40)
0.0993
(1.45)
-0.275***
(-3.39)
Yes, with commercial sexual
partner
-0.183**
(-2.09)
-0.106
(-1.27)
-0.977***
(-4.31)
N
3799
1889
1910
Notes: t and z statistics in parentheses for OLS and GMM respectively; * p < 0.10, ** p < 0.05, *** p <
0.01; All regressions include a full set of controls as given in Table 1.
38
Table 12 Estimates for unwilling sex from happiness equations
All
Male
Female
Panel A: Have unwanted sex
(Ref: no)
OLS
-0.114***
(-4.87)
-0.105**
(-2.55)
-0.129***
(-4.45)
GMM
-0.140***
(-2.59)
-0.0316
(-0.76)
-0.145**
(-2.24)
N
3710
1837
1873
Panel B: Have sex only to satisfy partner in the past 12 months
OLS
-0.0727***
(-6.58)
-0.0826***
(-5.05)
-0.0631***
(-4.19)
GMM
-0.0706***
(-6.45)
-0.0782***
(-4.98)
0.0517
(0.48)
N
3135
1580
1555
Notes: t and z statistics in parentheses for OLS and GMM respectively; * p < 0.10, ** p < 0.05, *** p <
0.01; All regressions include a full set of controls as given in Table 1.
Article
Full-text available
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.
Article
Full-text available
Purpose The objective of this study was to examine whether sexual satisfaction is associated with subjective well-being longitudinally. Methods Data from 2002, 2008 and 2011 were drawn from a nationally representative study among individuals residing in private households aged 40 and over (n = 12,105 in regression analysis). The established Satisfaction with Life Scale was used to assess life satisfaction. The well-recognized Positive and Negative Affect Schedule was used to quantify positive and negative affect, respectively. A single-item measure was used to quantify sexual satisfaction, ranging from 1 to 5 (higher values correspond to higher sexual satisfaction). The analysis was stratified by sex and age group (40–59 years, 60 years and over). Results The mean sexual satisfaction score was 3.4 (± 1.0) in men and 3.5 (± 0.9) in women. Fixed-effects regressions revealed that sexual satisfaction was positively associated with life satisfaction (total sample: β = .08, p < .001; men: β = .08, p < .001; women: β = .07, p < .001) and positive affect (total sample: β = .04, p < .001; men: β = .04, p < .001; women: β = .03, p < .01) as well as was negatively associated with negative affect (total sample: β = − .05, p < .001; men: β = − .05, p < .001; women: β = − .04, p < .001). Conclusion This study emphasizes the longitudinal association between sexual satisfaction and subjective well-being both in men and women. Life satisfaction is associated with satisfaction with sex life in both age groups. We conclude that sexual satisfaction is a life domain related to life satisfaction among older men and women. Thus, maintaining or improvement of sexual satisfaction could have an impact on successful ageing. Research is required to further elucidate the underlying mechanisms.
Article
Full-text available
Hypotheses concerning possible correlates of sexual satisfaction in marriage were tested using the replies of 797 married women and men of diverse ages to a 70-item mailed questionnaire that contained seven Likert-type sub-scales measuring different sexual and non-sexual variables. Multiple regression analysis, using sexual satisfaction as the dependent variable, yielded a five-variable model that accounted for a significant portion of the variation in sexual satisfaction (Adjusted R Squared = .602). The variable 'overall satisfaction with marriage' had the highest correlation with sexual satisfaction (r = .622), followed by 'satisfaction with non-sexual aspects of the relationship' (r = .609), frequency of spouse/partner orgasm per sexual encounter (r = .529), frequency of sexual activity (r = .370), and 'sexual uninhibitedness' (r = .230). None of three measures of religiosity made a significant contribution to explaining the variation on self reported sexual satisfaction. Men and women did not differ in level of sexual satisfaction, and adding gender to the regression model did not increase the level of explained variation. The results indicate that sexual satisfaction in these married respondents could not be compartmentalized to their sexual interactions, but was strongly associated with non-sexual aspects of the overall marital relationship as well.
Article
Full-text available
To cite this paper: Vinod Mishra Russell Smyth , (2015),"Workplace policies and training in China: evidence from matched employee-employer data", International Journal of Manpower, Vol. 36 Iss 7 pp. 986 - 1011 Purpose – The purpose of the study is to examine the extent to which workplace policies and practices are related to participation in, and frequency and duration of, workplace training, controlling for worker and workplace characteristics. Design/methodology/approach - We regress variables depicting participation, frequency and duration of workplace training on workplace policies and control variables. In the case of participation in training, the dependent variable is binary; hence, we use a logit model. To examine the number of times which employees participate in training and the number of days they spend training we use a Tobit model. The Lewbel (2012) method is used to examine whether there is a causal relationship between workplace policies and the frequency, and duration, of training. Findings – The findings suggest that about half of the workplace policies considered are positively correlated with the incidence and breadth of workplace training. There is also some support for the view that bundling of policies is positively correlated with the provision of workplace training. The Lewbel (2012) results suggest a causal relationship between a bundle of workplace policies and the frequency, and duration, of workplace training. There is, however, no evidence that workplace policies designed to devolve responsibilities to workers and incentivize staff polarizes skills through resulting in more training for professional staff over others. Originality/value –We use matched employer and employee cross-sectional data for Shanghai in China. To this point most studies that have examined the determinants of training use data for Europe or the United States. There are few studies of this sort for countries in other regions and, in particular, developing or transition countries. There are no studies at all on the relationship between workplace policies and practices designed to promote organizational performance and training in developing or transitional countries. This study addresses this gap in our understanding of the factors related to on-the-job training in transitional countries, such as China.
Technical Report
Full-text available
This paper uses a novel identification strategy proposed by Lewbel (2012, J. Bus. Econ. Stat.) to illustrate how causation between job satisfaction and life satisfaction can be established with cross-sectional data. In addition to examining the relationship between composite job satisfaction and life satisfaction, we consider the relationship between life satisfaction and different facets of job satisfaction. We find evidence of bidirectional causality between job satisfaction and life satisfaction and mixed evidence of causation between life satisfaction and different facets of job satisfaction.
Book
Revisits and updates the centrality of the social construction of sexuality, especially in the age of Viagra, FSD (female sexual dysfunction) and the media saturation of sex. Leonore Tiefer is one of the foremost sexologists working in the United States today; she is a well-known and respected scholar who writes engagingly and humorously about a wide array of topics in sexuality to appeal to both students and general readers. Revised and updated with new pieces on the medicalization of sex, FSD (female sexual dysfunction) and the politics of sex, as well as classic pieces found in the original edition, such as “Am I Normal?: The Question of Sex.”
Article
The links between income, sexual behavior and reported happiness are studied using recent data on a sample of 16,000 adult Americans. The paper finds that sexual activity enters strongly positively in happiness equations. Higher income does not buy more sex or more sexual partners. Married people have more sex than those who are single, divorced, widowed or separated. The happiness-maximizing number of sexual partners in the previous year is calculated to be 1. Highly educated females tend to have fewer sexual partners. Homosexuality has no statistically significant effect on happiness. © The editors of the Scandinavian Journal of Economics 2004. Published by Blackwell Publishing.
Article
We used data on 20,000 ever-married adults in the General Social Survey to examine the relationship between watching pornographic films and various measures of marital well-being. We found that adults who had watched an X-rated movie in the past year were more likely to be divorced, more likely to have had an extramarital affair, and less likely to report being happy with their marriage or happy overall. We also found that, for men, pornography use reduced the positive relationship between frequency of sex and happiness. Finally, we found that the negative relationship between pornography use and marital well-being has, if anything, grown stronger over time, during a period in which pornography has become both more explicit and more easily available.
Article
IN SOME APPLICATIONS of the general linear model, the usual assumptions of homoscedastic disturbances and fixed coefficients may be questioned. When these requirements are not met, the loss in efficiency in using ordinary least squares (OLS) may be substantial and, more importantly, the biases in estimated standard errors may lead to invalid inferences. This has caused a number of writers to propose models which relax these conditions and to devise estimators for their more general specifications, e.g., Goldfeld and Quandt (8) for heteroscedasticity and Hildreth and Houck (11) for random coefficients. However, because the effect of introducing random coefficient variation is to give the dependent variable a different variance at each observation, models with this feature can be considered as particular heteroscedastic formulations for the purpose of detecting departure from the standard linear model. A test for heteroscedasticity with the same asymptotic properties as the likelihood ratio test in standard situations, but which can be computed by two least squares regressions, thereby avoiding the iterative calculations necessary to obtain maximum likelihood estimates of the parameters in the full model, is considered in this paper. The approach is based on the Lagrangian multiplier
Article
Smokers tend to be more impatient and prefer immediate benefits compared with non-smokers. If people follow their primitive instincts they will engage in sexual behavior. Impatient people are more likely to have sex. However, for married people, having sex with a spouse is considered to be an investment in family life. Based on individual level data of Japan General Social Surveys, this paper investigated how sexual behavior differs between smokers and non-smokers. In addition to it, determinants of life satisfaction are also investigated. The important findings through the interval regression model and the ordered probit model are: (1) frequency of sex is positively associated with family satisfaction; (2) unmarried smokers are more likely to have sex than unmarried non-smokers; (3) married smokers are less likely to have sex than married non-smokers. It follows from estimation results that smoking behavior alerts us to his unobserved characteristics before marriage, thus reducing information asymmetry.