The most incisive guide to issues facing the American family today . . . An invaluable resource for anyone wishing to stay on the cutting edge of research on family trends.

-W. Bradford Wilcox
Associate Professor of Sociology, University of Virginia 

Retreat from Marriage Gives America the Blues

Bryce J. Christensen and Robert W. Patterson

If Americans with higher incomes report higher levels of happiness, why have reported levels of happiness declined during the past fifty years when living standards and incomes have increased? Among economists, prevailing theories to explain the paradox rest on one of two claims: 1) that happiness depends more on whether one’s income matches that of peers than whether it matches or exceeds that of their parents at the same point in the life cycle; or 2) that happiness has declined for many as household income has grown more unequal. A study by three Italian economists, however, offers a more compelling explanation, finding that a decline in “social connections,” particularly in marriage, has played a significant role in depressing the American mood.

Bartolini of the University of Siena and his colleagues came to their conclusion after discovering, in their first set of regressions using U.S. General Social Survey data between 1972 and 2004, that several variables—including changes in household income, reference income (a measure of income relative to peers), work status, and demographic characteristics—did not satisfactorily account, in their most sophisticated models, for the decline in self-reported happiness over the same time period. However, their second set of regressions explored the relationship between happiness and variables that measure what they call “social connections and confidence in institutions” or SCC for short. In these tests, the economists were able to find a “more precise prediction of the [un]happiness trend.”

The SCC variables included marital status, number of children, social contacts, trust in individuals, group membership, and confidence in institutions. In later analysis, the researchers found that all these measures had declined between 1975 and 2004. But in the second set of regressions, all the SCC variables save number of children were found to be significantly linked with reduced happiness. More to the point, marital status was found to be the most determinative factor in these regressions, consistently correlating with happiness (p

Moreover, the economists found that Americans in second marriages were not as happy as their peers in first marriages, “even without consideration the happiness reduction due to a divorce.” Although the Italian scholars don’t explicitly make the connection, their findings related to “social contacts” explain, in part, why marital status represents the trump card of their study, as marriage appears to direct individuals to a more fulfilling social life centered around family. As they report:

Spending evenings with relatives, neighbors, or friends correlates with a greater happiness, while spending evenings in a bar is associated with a lower level of happiness. More precisely, spending at least one evening with relatives goes with twice the happiness of spending one evening with friends or neighbors. Spending at least one evening at a bar is associated with less happiness, as much as spending evenings with relatives correlates with greater happiness.

Yet the study does more than simply confirm the connections between marriage and happiness. The Italian economists make an impressive, empirical case that the decline in American happiness related to declines in social connections—coupled with changes in reference income—“more than offsets” any increase in happiness arising from increases in household income since the 1970s. Their findings therefore go a long way to explain a paradox that has puzzled their fellow economists while confirming that money, indeed, does not buy happiness.

(Stefano Bartolini, Ennio Bilancini, and Maurizio Pugno, “Did the Decline in Social Connections Depress Americans’ Happiness?” forthcoming in Social Indicators Research.)