In which it is argued that the founding virtues are inextricably bound up with the ways in which human beings acquire deep satisfactions in life—the ways in which they pursue happiness. Evidence about self-reported happiness is presented to support that position, and to document a steep drop in self-reported happiness in Fishtown.
THE DETERIORATION OF social capital in lower-class white America strips the people who live there of one of the main resources through which Americans have pursued happiness. The same may be said of the deterioration in marriage, industriousness, honesty, and religiosity. These are not aspects of human life that may or may not be important, depending on personal preferences. Together, they make up the stuff of life.
Using the word happiness may seem to be asking for trouble—doesn’t happiness mean many different things to different people? But the core nature of human happiness is widely agreed upon in the West. It goes all the way back to Aristotle’s views about happiness in the Nicomachean Ethics. Distilling his discussion of happiness into a short definition leaves out a lot, but this captures the sense of Aristotle’s argument well enough for our purposes: Happiness consists of lasting and justified satisfaction with life as a whole. The definition in effect says that when you decide how happy you are, you are thinking of aspects of your life that tend to define your life (not just bits and pieces of it); that you base your assessment of your happiness on deep satisfactions with the way things have gone, not passing pleasures; and that you believe in your heart of hearts that those satisfactions have been worth achieving. It is not really a controversial definition—try to imagine a definition of happiness you could apply to your own life that is much different.
What are these deep satisfactions that let us reach old age happy? We can begin by saying what they aren’t. Few people reach old age satisfied with their lives because they were rich or famous. Film and music producer David Geffen—and a billionaire—once said in a television interview, a sad smile on his face, “Show me someone who thinks that money buys happiness, and I’ll show you someone who has never had a lot of money.”1 Here’s a variant: Show me someone who thinks deep satisfactions in old age come from having been rich or famous, and I’ll show you someone who’s never been old. Rich and famous people can reach old age deeply satisfied with their lives, but because of how they got the money or how they got the fame.
Once you start to think through the kinds of accomplishments that do lead people to reach old age satisfied with who they have been and what they have done, you will find (I propose) that the accomplishments you have in mind have three things in common. First, the source of satisfaction involves something important. We can get pleasure from trivial things, but pleasure is different from deep satisfaction. Second, the source of satisfaction has involved effort, probably over an extended period of time. The cliché “Nothing worth having comes easily” is true. Third, some level of personal responsibility for the outcome is essential. In the case of events close to home, you have to be able to say, “If it hadn’t been for me, this good thing wouldn’t have come about as it did.”
There aren’t many activities in life that satisfy the three requirements of importance, effort, and responsibility. Having been a good parent qualifies. Being part of a good marriage qualifies. Having done your job well qualifies. Having been a faithful adherent of one of the great religions qualifies. Having been a good neighbor and good friend to those whose lives intersected with yours qualifies. But what else?
Let me put it formally: If we ask what are the domains through which human beings achieve deep satisfactions in life—achieve happiness—the answer is that there are just four: family, vocation, community, and faith, with these provisos: Community can embrace people who are scattered geographically. Vocation can include avocations or causes.
It is not necessary for any individual to make use of all four domains, nor do I array them in a hierarchy. I merely assert that these four are all there are. The stuff of life occurs within those four domains.
The simplest way of making the point that the four domains are in fact related to happiness is to use the social scientists’ measure of happiness: asking people “How happy are you?” It is not a perfect measure—I’m sure that many people tell an interviewer they are happy for reasons that have nothing to do with “lasting and justified satisfaction with life as a whole.” But the empirical relationship between self-reported happiness and the four domains is worth examining. The data come from the General Social Survey, combining the surveys conducted from 1990 through 2010.2
FIGURE 15.1. “HOW HAPPY ARE YOU?”
Source: Author’s analysis of the GSS surveys from 1990–2008. Sample limited to whites ages 30–49.
People were given the choice of answering “not too happy,” “pretty happy,” and “very happy.” Figure 15.1 shows the results.
About one out of three whites ages 30–49 said they were “very happy”—the only answer of the three that means much. You might answer “pretty happy” when what you really mean is “I’m doing okay, can’t complain” or “Things could be a lot worse.” People don’t answer “very happy” so haphazardly. It is likely to be a sign that a person really does assess his life positively.
The relationship of marriage to happiness is simple as can be. There’s hardly anything better than a good marriage for promoting happiness and nothing worse than a bad one. Fifty-eight percent of white prime-age GSS respondents who said they were in very happy marriages also said their lives were very happy, compared to 10 percent who said their marriages were “pretty happy” and 8 percent who said their marriages were “not too happy.”
Even without asking whether the marriage itself is happy, marriage is still a good bet for achieving happiness. Figure 15.2 shows the breakdown for marriage versus singlehood in its various forms.
Forty percent of married prime-age whites reported that they were very happy compared to 18 percent of everyone else (weighted average). Among those who were not married, widows were the happiest and never-marrieds were the unhappiest.
During three surveys in the 2000s, the GSS also asked about cohabitation. In terms of happiness, cohabitation is a little bit better than living alone, but not much. In those three surveys, 43 percent of all prime-age whites who were legally married said they were very happy, compared to 29 percent of those who were cohabiting with a partner and 22 percent of those who had a partner and were not cohabiting. Even the 29 percent has an artifact in it. It is a lot easier to end an unhappy cohabitation than to end an unhappy marriage. In effect, the pool of cohabitating people is drained of unhappy potential respondents much more quickly than is the pool of married people.
FIGURE 15.2. SELF-REPORTED HAPPINESS AND MARITAL STATUS
Source: GSS surveys from 1990 to 2010. Sample limited to whites ages 30–49.
Do children make people happy? As any parent can testify, that’s a complicated question. Infants are a source of joy, but they are a lot of work, especially for the mother, and they disrupt pleasant patterns of life that prevailed before the birth. Teenagers are notoriously a source of anxiety and unhappiness for the parents. And yet it is also true that most parents see their children as a defining aspect of their lives, often the defining aspect. When the children turn out well, they are also the source of perhaps the deepest of all human satisfactions. These many complications account for the fact that married whites in their thirties and forties report that they are about equally happy whether or not they have children. Among unmarried people (combining those who are separated, divorced, widowed, and never married), those with children are notably less happy than those without children—a finding that reflects the many economic and emotional difficulties of being a single parent.
Before leaving the topic, I must emphasize that the statistical relationship between marriage and happiness is not completely causal. To some degree, happy people self-select into marriage and unhappy people self-select out of it.3 However, the causal role that marriage plays in producing happiness is also indisputable. If you don’t know that from your own life, just ask people who are happily married. They will seldom have any hesitation in identifying their marriages as a primary cause of their lasting and justified satisfaction with life as a whole.
Direct evidence for the relationship of vocation to happiness comes via a GSS question that asks, “On the whole, how satisfied are you with the work you do?” The relationship of the answers to that question and self-reported happiness is unambiguous and strong, as shown in Figure 15.3.
FIGURE 15.3. SATISFACTION WITH WORK AND HAPPINESS
Source: GSS surveys from 1990 to 2010. Sample limited to whites ages 30–49.
The results support using the concept of vocation instead of job when thinking about what makes people happy. The highest proportion of people with high work satisfaction who also reported that they were very happy consisted of women whom the GSS called housewives. I have substituted the old-fashioned term homemakers in the figure, reflecting my assumption that the source of their satisfaction had a lot more to do with making a home than with keeping house.
Whether the satisfaction came from making a home or working at paid employment, the percentages of very happy people halved for those who were only moderately satisfied with their work, and halved again for those who were dissatisfied with their work.
Figure 15.4 sums up the bemusing situation facing social scientists, who as a group are predominantly secular. It shows the GSS data regarding religious attendance and self-reported happiness.
FIGURE 15.4. ATTENDANCE AT RELIGIOUS SERVICES AND HAPPINESS
Source: GSS surveys from 1990 to 2010. Sample limited to whites ages 30–49.
Social scientists rarely find such an orderly relationship over so many categories. At the top, 49 percent of those who attend worship services more than once a week report they are very happy. At the bottom, only 23 percent of the white adults who never attend worship services report they are very happy.
It is hard for me to find an artifact that might explain this result. It does not seem plausible (by any logic I can think of) that people who are already happy are more likely to attend worship services than unhappy people. If anything, any artifact in the data would seem to work the other way—people who are unhappy go to church in search of solace.
Is it the act of going to worship services or the content of the religious faith that is associated with the happiness? The answer appears to be that you have to believe and attend. Forty-three percent of people who are believers and attend at least once a week said they were happy, about twice the percentage who say they believe but never attend church. People who attend services without believing don’t get much advantage either—the percentage of nonbelievers reporting they were very happy stayed stuck at around 20 percent regardless of their attendance.
Explaining why this relationship between religiosity and happiness persists using a nonreligious explanation is a problem. Is it a matter of self-delusion? It might be argued that in America religion still refers overwhelmingly to Christianity, Christianity promises believers salvation and eternal life, and impressionable people buy into it. They’re happy because they think they are saved and will go to heaven, but there’s no substance to that happiness.
The data are not consistent with that hypothesis. First, believing in salvation and heaven isn’t enough. People who self-identify as fundamentalists—meaning that they definitely believe in salvation and heaven—but who attend church no more than once a year have a “very happy” percentage (22 percent) that is almost as low as for nonbelievers. Second, the relationship of religious attendance to self-reported happiness is almost as strong for people who identify themselves as religious moderates or liberals (meaning that their confidence in salvation and heaven is likely to be dodgy) as it is for fundamentalists—47 percent of fundamentalists who attend church weekly or more report they are very happy, compared to 42 percent of religious moderates and 41 percent of religious liberals.
The GSS has items measuring the level of community activity, as reported in the previous chapter, but those items were seldom given to the same respondents who were asked how happy they are for the GSS surveys conducted during the 1990s and 2000s (any given subject of the GSS is not asked about all the items for that year’s survey). Happily, there is an alternative source of data in the form of the Social Capital Benchmark Survey (SCBS) conducted under the auspices of the Saguaro Seminar organized by Robert Putnam. The survey was conducted in 2000. Its total sample of 29,233 included 8,895 whites ages 30–49, and it contained a comprehensive set of measures of level of community activity along with a question about self-reported happiness.4 The Social Capital Benchmark Survey also created indexes for different types of social capital. I used five of them for the comparison with happiness, made up of indicators described in the box.5
The group involvement index counts memberships in fraternal, ethnic, political, sports, youth, literary, veterans, or other kinds of clubs or organization other than religious ones.
The organized group interactions index combines measures of actual attendance at public meetings, club meetings, and local community events.
The giving and volunteering index combines indicators of volunteering for various organizations, frequency of volunteering, and charitable contributions.
The informal social interactions index combines measures of visits with relatives, having friends to the home, socializing with coworkers, hanging out with friends in public places, and playing cards and board games.
The electoral politics index combines indicators of voting, voter registration, interest in politics and national affairs, political knowledge, and frequency of newspaper reading.
In table 15.1, I grouped these indexes into five categories running from “very low” to “very high.” For indexes with many values, the cutoff points for the categories were the 10th, 33rd, 67th, and 90th centiles of the distribution. For indexes with fewer values, I followed those guidelines as closely as possible.6
TABLE 15.1. PERCENTAGE OF WHITES AGES 30-49 WHO REPORT THAT THEY ARE VERY HAPPY
Index category | ||
Social capital index | Very low | Very high |
Group involvement | 32% | 47% |
Organized group interactions | 29% | 49% |
Giving and volunteering | 32% | 57% |
informal social interactions | 29% | 48% |
Electoral politics | 29% | 48% |
Source: Social Capital Benchmark Survey. Sample limited to whites ages 30-49.
High levels of community involvement were consistently associated with much higher levels of “very happy” people than low levels of community involvement. Furthermore, each of the different types of involvement seemed to be about equally related to happiness, with “giving and volunteering” having a modest edge.7
The Social Capital Benchmark Survey also created an index of social trust combining responses to questions about trusting neighbors, coworkers, congregants, store workers, local police, and others in general. The relationship of self-reported happiness to their level of social trust was unusually high, as shown in Figure 15.5.
FIGURE 15.5. RELATIONSHIP OF SOCIAL TRUST TO HAPPINESS
Source: Social Capital Benchmark Survey. Sample limited to whites ages 30–49.
If people in the Social Capital Benchmark Survey were very high in social trust, 62 percent of them reported being very happy. If they were very low in social trust, only 21 percent of them reported being very happy.
Each of the four domains—family, vocation, faith, and community—has a direct and strong relationship to self-reported happiness. But which is the most important? Multivariate analysis can help answer that question. Appendix G gives the details, but essentially we are asking what the role of each is after controlling for the others and also asking about how they interact with one another.8
Figure 15.6 shows some of the results when the effects of the “high” level of each measure (a very happy marriage, high work satisfaction, strong religious involvement, and high social trust) are added to the probability that people say they are very happy.
At baseline—unmarried, dissatisfied with one’s work, professing no religion, and with very low social trust—the probability that a white person aged 30–49 responded “very happy” to the question about his life in general was only 10 percent. Having either a very satisfying job or a very happy marriage raised that percentage by almost equal amounts, to about 19 percent, with the effect of a very satisfying job being fractionally greater. Then came the big interaction effect: having a very satisfying job and a very happy marriage jumped the probability to 55 percent. Having high social trust pushed the percentage to 69 percent, and adding strong religious involvement raised the probability to 76 percent.
FIGURE 15.6. THINGS THAT INCREASE THE LIKELIHOOD THAT SOMEONE REPORTS BEING VERY HAPPY, IN ORDER OF THEIR IMPORTANCE
Source: GSS surveys from 1990–2010. Sample limited to white adults ages 30–49. Results of a logit analysis, fitted for a person age 40.
The details are not etched in stone. Different specifications of the model (for example, using three categories for the variables instead of four) sometimes gave a very happy marriage a greater independent effect than a very satisfying job, and gave strong religion a greater independent effect than high social trust. What remained unchanged under all the variations was the primacy of marriage and vocation and the secondary role of social trust and religion in raising the probability of responding “very happy.”
How much difference does it make to be in a happy marriage instead of just being married? In one sense, a lot, as I described earlier in the discussion of family. Unhappy marriages were associated with a low probability of being happy with life in general, whereas happy marriages were associated with high probabilities. Similarly, the multivariate analyses say that you have a fractionally greater likelihood of being happy if you are single than if you are in a marriage that is less than “very happy.” But among the people in the analysis reported in Figure 15.6, 67 percent said their marriages were “very happy,” and the payoff for that happy marriage was extremely large. Let’s ignore social trust and religion. In an analysis that includes just age, marriage, work, and the interactions between marriage and work, an unmarried person with very dissatisfying work more than triples his probability of reporting “very happy” if he can enter a happy marriage, from 9 percent to 30 percent. Even the lucky unmarried person with very satisfying work more than doubles his probability of reporting “very happy” if he marries, from 28 percent to 63 percent. So marriage is a risk, but the downside is much smaller than the upside, and for most people the risk pays off.
The analyses reported so far have not included income as a control variable, and yet it has consistently been found that, at any given slice of time, rich people are more likely than poor people to say they are happy. I have ignored that consistent relationship because it has just as consistently been found that the relationship is not causal after abject poverty has been left behind. Longitudinal evidence reveals that people don’t get happier as they go from a modest income to affluence.9
The relationship exists in a cross section of the population because the qualities in individuals that make them happy in their marriages, satisfied with their work, socially trusting, and strongly involved with their religion are also qualities that are likely to make them successful in their jobs. In addition, marriage itself, independent of the personal qualities that produced the marriage, increases income—both by combining two incomes in some cases and because of the marriage premium discussed in chapter 9. Conversely, people who have a failed family life, are dissatisfied with their jobs, are disengaged from their communities, and have no spiritual life tend not to be happy because of those failures—and the same qualities that produced those failures also mean that, as a group, they are likely to have depressed incomes. Controlling for income as an explanation of happiness is as likely to mislead as to inform, wrongly attributing to income effects that are actually the result of qualities that produce both happiness and high income.
For the record, however, all of the relationships between the four domains and happiness that I describe in this chapter, bivariate and jointly in the multivariate analysis, persist after controlling for family income.10 The absolute magnitude of the relationships are attenuated because of the cross-sectional correlation that exists between happiness and income. But the incremental effects on the probability of being very happy as shown in Figure 15.6 are about the same. For example, in Figure 15.6, having very satisfying work increased the baseline probability of answering “very happy” by 9 percentage points. For someone with the median income, the same change in work satisfaction increased the baseline probability by even more, 16 percentage points.11 Without controlling for income, adding a happy marriage to very satisfying work boosted the probability by 36 percentage points. For a person at the median income, the comparable boost was 31 percentage points.
THE ARGUMENT UNDERLYING these many graphs and analyses has been that the founding virtues are instrumental to the domains for achieving deep satisfactions in life. Decay in the founding virtues is problematic for human flourishing. Those statements may have been self-evident to many readers without all the graphs and tables, but empirical support for them is readily available.
We are now in a position to ask what happened to self-reported happiness over time and by neighborhood.
The roles of the four domains of happiness are somewhat different in Belmont and Fishtown, but they add up to a remarkably similar total. In Figure 15.7, I repeat the exercise of Figure 15.6, showing the increment in the probability of being very happy as the effects of each domain are added in the order of their importance, but reporting the results separately for Belmont and Fishtown.
Belmont and Fishtown have somewhat different profiles. Give the people of Belmont very satisfying work and a very happy marriage, and social trust doesn’t add much to the probability of reporting being “very happy,” and strong religion actually lowers the probability—statistically at least (it is appropriate to accept counterintuitive results of complex quantitative analyses provisionally). For the people of Fishtown, the effects of the four domains are more evenly spread, with the addition of a happy marriage, high social trust, and strong religion each adding a roughly equal increment to the probability of reporting being “very happy.”
FIGURE 15.7. DIFFERENCES AND SIMILARITIES IN THE ACHIEVEMENT OF HAPPINESS IN BELMONT AND FISHTOWN
Source: GSS surveys from 1990–2010. Sample limited to white adults ages 30–49. Results of a logit analysis, fitted for a person age 40.
In the end, people who are high on all four measures have a remarkably similar probability of reporting they are very happy, regardless of whether they belong to Belmont or Fishtown. This is worth pondering. There is no inherent barrier to happiness for a person with a low level of education holding a low-skill job. The domains for achieving happiness can work as well for the people of Fishtown as for the people of Belmont. But they haven’t been working as well over the course of the last half century, which leads to a predictable result when we examine the trendline of happiness in Figure 15.8.
FIGURE 15.8. SELF-REPORTED HAPPINESS OVER TIME IN BELMONT AND FISHTOWN
Source: GSS. Sample limited to whites ages 30–49. Data smoothed using locally estimated regression (LOESS).
When the GSS surveys began in the first half of the 1970s, the percentages of people in Belmont and Fishtown who reported they were very happy already showed a substantial gap of about 15 percentage points. There is no way of knowing what the proportions were in 1960.
In Belmont, the pattern is familiar from the trendlines in part 2: deterioration during the 1970s, stabilization thereafter. In Fishtown, self-reported happiness dropped from about 33 percent in the 1970s to an average of 22 percent in the 2000s. It is not a surprising finding, given the trendlines for Fishtown presented in part 2 and the testimony of the people of the real Fishtown in chapter 12, but it is an important finding. The trendlines for the founding virtues were not merely showing changes in social institutions and norms. They were saying things about the deterioration of life in Fishtown at the level of human happiness.