“WHY NOT SETTLE DOWN ALREADY?” A QUANTITATIVE ANALYSIS OF THE DELAY IN MARRIAGE

• Author: David Weiss,Cezar Santos
• DOI: http://doi.org/10.1111/iere.12163
• Date: 01 May 2016
• Published date: 01 May 2016

INTERNATIONAL ECONOMIC REVIEW

Vol. 57, No. 2, May 2016

“WHY NOT SETTLE DOWN ALREADY?” A QUANTITATIVE ANALYSIS

OF THE DELAY IN MARRIAGE∗

BYCEZAR SANTOS AND DAVID WEISS1

FGV/EPGE—Escola Brasileira de Economia e Financ¸as, Brazil; Tel Aviv University, Israel

A striking change in American society in the last 40 years has been the decline and delay in marriage. The fraction of

young adults who have never been married increased significantly between 1970 and 2000. Idiosyncratic labor income

volatility also rose. We establish a quantitatively important link between these facts. If marriage involves consumption

commitments, then a rise in income volatility delays marriage. We quantitatively assess this hypothesis vis-`

a-vis others

in the literature. Increased volatility accounts for about 20% of the observed delay in marriage and is strong relative to

other mechanisms.

1. INTRODUCTION

One of the most striking changes in American society over the last 40 years has been the

decline and delay in first-time marriage. The fraction of young men and women who have

never been married increased significantly between 1970 and 2000. This trend has captured

the attention of both academic researchers and the general public.2The question here, in the

vernacular, is: Why not settle down already? The answer we propose relies on the increased

labor income volatility observed in this period. In order to quantitatively assess this hypothesis,

we build and estimate a structural equilibrium search model of the marriage market.

Figure 1 shows the fraction of never-married American white males, by age, for both 1970

and 2000. This graph illustrates how the onset of marriage has been delayed. The numbers are

striking. In 1970, only 26% of 25-year-old white males had never been married. By 2000, this

number had more than doubled to 57%. At age 35, only 8% of white males were single in 1970,

whereas this number increased to 21% in 2000.3These numbers clearly illustrate the decline

and delay in marriage observed in this period.4

The economics literature has documented a rise in idiosyncratic labor income volatility over

the same period. Gottschalk and Moffitt (1994) and Heathcote et al. (2010), among others, find

∗Manuscript received February 2013; revised December 2014.

1We are very grateful to Jeremy Greenwood for his advice. The editor Jes´

us Fern´

andez-Villaverde and three

anonymous referees provided invaluable comments. We would also like to thank Nezih Guner, Moshe Hazan, Greg

Kaplan, Fatih Karahan, John Knowles, Dirk Krueger, Iourii Manovskii, Serdar Ozkan, Mich`

ele Tertilt, and seminar

participants at Penn, U of Montreal, PUC-Rio, EPGE-FGV/RJ, EESP-FGV/SP, the U of S˜

ao Paulo, ITAM, Mannheim,

Konstanz, the 2011 Midwest Macro, the 2011 Minerva Summer Institute at The Hebrew University, the SED-Ghent,

the 2011 Cologne Macro Workshop, and the 2011 Asian Meeting of the Econometric Society, the 2012 Royal Economic

Society Meeting (Cambridge), St. Louis Fed, and the 2013 North American Meeting of the Econometric Society for their

comments. We thank The Pinhas Sapir Center for Development for financial assistance and Unicredit, and Universities

for selecting this article for the UWIN Best Paper Awardon Gender Economics. The usual disclaimer applies. Please

address correspondence to: Cezar Santos, FGV/EPGE, Escola Brasileira de Economia e Financ¸as, Praia de Botafogo,

190/1100, Rio de Janeiro 22250-900, Brazil. E-mail: cezar.santos@fgv.br.

2For an excellent review of the academic literature, see Stevenson and Wolfers (2007).

3Decline and delay of marriage are two similar but distinct concepts. The decline in marriage is captured by the

fraction of people who never marry. There has been an increase in this fraction in the data. The delay in marriage can

be expressed as the fraction of people who are married by each age. As will be clear later, the model is consistent with

both.

4The graph for white women looks very similar. For data on cohabitation and by education groups, see Appendix

A.2. Detailed explanations about the data sources are contained in Appendix A.1.

425

C

(2016) by the Economics Department of the University of Pennsylvania and the Osaka University Institute of Social

and Economic Research Association

426 SANTOS AND WEISS

FIGURE 1

FRACTION OF WHITE MALES NEVER MARRIED,BY AGE

an increase in the variance of persistent and transitory shocks to income between the late 1960s

and 2000. Various effects of this changing labor market have drawn the attention of a wide

body of literature.5However, to the best of our knowledge, no quantitative work has been done

relating changes in income volatility with changing marriage decisions of young adults.

Figure 2 shows the increase in the median age of marriage for males and the increase in

labor income volatility as measured by the standard deviation of persistent income shocks. It

is interesting to note that both series exhibit a very similar increase between the late 1960s

and 2000. In fact, the correlation between the two series is 0.96. Some empirical papers have

also provided suggestive evidence of the impact of certain aspects of labor market volatility on

marriage.6

The contribution of this article is to quantitatively establish the effect of rising labor income

volatility on the delay in marriage. We do this by exploring three channels through which

income volatility can affect marriage timing. The first and novel effect that we explore in this

article arises from the presence of consumption commitments within marriage. Consumption

commitments emerge when households consume goods for which adjustments are costly. These

consumption commitments aggravate the effects of income fluctuations: Households must cover

these commitments, or face costly adjustments, following a bad income realization. To do so,

they might need to cut their discretionary consumption substantially, causing a large utility

loss. In this article, we provide evidence that married individuals, compared to their single

counterparts, have more consumption commitments; in particular, more married households

have children. Therefore, a rise in the volatility of income results in a delay in marriage as these

commitments become less desirable. That is, singles might find it preferable to wait until one

5For an excellent overview of this literature with a specific focus on welfare, see Heathcote et al. (2012).

6For example, using U.S. data, Oppenheimer et al. (1997) argue that difficulties in starting careers in a period of

higher volatility have delayed marriage. Ahn and Mira (2001) show that employment risk has caused delay in marriage

in Spain. Southall and Gilbert (1996) study the impact of economic distress in the 19th century United Kingdom and

find that periods to more uncertainty are related to fewer marriages overall as well as higher variability in marriage

rates for workers in more volatile occupations.