Demography

, Volume 48, Issue 3, pp 983–1004

Cohabitation History, Marriage, and Wealth Accumulation

Authors

    • Center for Human Resource ResearchThe Ohio State University
  • Matthew A. PainterII
    • University of Wyoming
Article

DOI: 10.1007/s13524-011-0043-2

Cite this article as:
Vespa, J. & Painter, M.A. Demography (2011) 48: 983. doi:10.1007/s13524-011-0043-2

Abstract

This study extends research on the relationship between wealth accumulation and union experiences, such as marriage and cohabitation. Using data from the National Longitudinal Survey of Youth 1979, we explore the wealth trajectories of married individuals in light of their premarital cohabitation histories. Over time, marriage positively correlates with wealth accumulation. Most married persons with a premarital cohabitation history have wealth trajectories that are indistinguishable from those without cohabitation experience, with one exception: individuals who marry their one and only cohabiting partner experience a wealth premium that is twice as large as that for married individuals who never cohabited prior to marrying. Results remain robust over time despite cohabiters’ selection out of marriage, yet vary by race/ethnicity. We conclude that relationship history may shape long-term wealth accumulation, and contrary to existing literature, individuals who marry their only cohabiting partners experience a beneficial marital outcome. It is therefore important to understand the diversity of cohabitation experiences among the married.

Keywords

MarriageCohabitationRace/ethnicityWealthMultilevel growth curve models

Introduction

Substantial wealth inequality characterizes the United States. The wealthiest 1%, the next 9%, and the remaining 90% of Americans each holds about one-third of total household wealth (Kennickell 2006). In 2004, nearly 17% of Americans had zero or negative net worth, and if they owned assets, they carried substantial debt (Wolff 2007). Many factors contribute to this wealth inequality, including education, income, inheritance, and family structure (Avery and Rendall 2002; Keister 2005; Wolff 2002). Relationship history also matters (Hao 1996; Waite and Gallagher 2000). Married individuals save more, have higher likelihoods of attaining affluence, and accumulate more wealth than cohabiting, divorced, and other nonmarried persons (Grinstein-Weiss et al. 2008; Hirschl et al. 2003; Holden and Kuo 1996; Wilmoth and Koso 2002).

To date, research on relationship history and wealth has typically compared the married with either nonmarried or remarried individuals. This approach treats those who marry for the first time as a homogenous group with respect to their relationship history. Yet, relationship experiences among first-time married persons are quite diverse if we consider premarital cohabitation. Cohabitation—living together in a romantic relationship without being married—is the most common first coresidential union in the United States. Indeed, two-thirds of recently married individuals have cohabited (Bumpass and Lu 2000; Schoen et al. 2007). Thus, by the time of marriage, many individuals have had a history of cohabiting whether they did so multiple times or only with their future spouse. Our goal is to explore the patterns of wealth accumulation among the married and to assess whether these patterns vary by individuals’ premarital cohabitation histories.

This study presents a detailed descriptive analysis of married persons’ wealth accumulation over time. It highlights the importance of examining the diversity of cohabitation histories among the married in order to understand marital outcomes. In doing so, our study challenges the current scholarly convention that cohabitation is detrimental to marriage. By examining wealth accumulation, our study also addresses broader issues in stratification. Wealth matters not only because it insulates against job loss and financial stress but also because it improves life chances by creating cycles of advantage through access to better schools, neighborhoods, and inheritances. Given the wealth advantages of marriage and the prevalence of cohabitation, it is important to more fully understand how cohabitation fits into the relationship between marriage and wealth accumulation over time.

Background

Marriage and Wealth

Several factors shape wealth accumulation, including educational attainment and income (Keister 2005). Individuals with higher education have higher earnings potential, save more, and accrue less debt (Oliver and Shapiro 2006). Inheritance also matters (Avery and Rendall 2002), in large part because it accounts for at least one-half of Americans’ wealth (Gale and Scholz 1994). Moreover, inheritance greatly facilitates homeownership, a core asset of many Americans (McNamee and Miller 1998). Apart from these and other factors, relationship history shapes wealth accumulation, with marriage sometimes described as a wealth-enhancing institution (Hao 1996).

Married individuals benefit from several institutionalized characteristics of marriage that facilitate savings and enhance wealth (Wilmoth and Koso 2002). They enjoy economies of scale and often benefit from financial transfers (Hao 1996; Waite and Gallagher 2000), wage premiums (Cohen 2002; Stratton 2002), and dual incomes. These factors help generate disposable income usable for greater savings and investments. Over time, resource pooling and joint investment further augment married persons’ wealth through capital appreciation and asset acquisition (Avellar and Smock 2005; Winkler 1997). Marriage is legally defined and benefits from normative expectations of permanence that increase trust, allowing married couples to feel comfortable investing together (Heimdal and Houseknecht 2003). Marriage even contains specific norms for investments, such as homeownership (Townsend 2002; Waite and Gallagher 2000).

Existing gender and racial/ethnic wealth inequality affect individuals’ wealth accumulation. Women typically accumulate less wealth than men (Schmidt and Sevak 2006; Yamokoski and Keister 2006) because of inequalities in labor force participation and earnings (O’Neill 2003; Warren et al. 2001). Blacks and Hispanics typically accumulate less wealth than whites, partly because of a historical legacy of inequality that limits educational, occupational, and financial opportunities (Campbell and Kaufman 2006; Conley 1999; Oliver and Shapiro 2006; Shapiro 2004). Consequently, nonwhites receive smaller inheritances from family than do whites, which contributes substantially to racial/ethnic wealth inequality (Avery and Rendall 2002). Discriminatory lending, residential segregation, and smaller intergenerational transfers further hinder nonwhites from owning homes of comparable value and with similar potential for appreciation as whites (Conley 1999; Krivo and Kaufman 2004; Oliver and Shapiro 2006). Thus, although wealth may grow over the course of marriage, nonwhites face unique challenges in improving their net worth that stem from social and structural processes of racial/ethnic stratification. Consequently, analyses of marital wealth should be sensitive to gender and race/ethnicity.

Cohabitation and Marital Wealth

Selection

Cohabitation is selective in a number of ways. First, it is selective of economically disadvantaged persons. Those with low income and education tend to cohabit whereas more-advantaged persons marry (Oppenheimer 2003; Xie et al. 2003). Cohabiters who transition to marriage have greater earnings potential than other cohabiters (Oppenheimer 2003; Xie et al. 2003) but are still economically disadvantaged compared with married individuals who never cohabited. Second, cohabitation is selective of race/ethnicity. Blacks, especially women, are more likely to cohabit than whites or Hispanics, and cohabiters who marry are disproportionately white (Bumpass and Lu 2000; Carlson et al. 2004; Manning et al. 2004). Thus, the roughly one-half of cohabitations that transition to marriage (Bumpass and Lu 2000) are selective of whites and individuals with higher earnings potential. Third, marriage grows more selective over time. Married cohabiters disproportionately divorce (Dush et al. 2003), as do blacks and U.S.-born Hispanics (Phillips and Sweeney 2005). Married individuals with lower incomes and education also disproportionately divorce compared with their higher-earning and higher-educated counterparts (Teachman and Polonko 1990). Over time, therefore, marriage is selective of individuals with the greatest wealth potential: whites, those who never cohabited, and those with higher incomes and education. These patterns stress the importance of considering selection out of marriage and exploring marital wealth over time.

Economic Characteristics

Besides selection, cohabitation has distinctive economic characteristics and arrangements that may help shape patterns of wealth accumulation. Cohabiters receive as large of a boost in income as married couples (Light 2004), and cohabiting men gain wage premiums (Stratton 2002). In addition, cohabiting women make greater economic contributions to the household and have fewer exits from the labor force than married women (Avellar and Smock 2005; Kenney 2004). Although only one-half of cohabiters pool resources compared with two-thirds of married couples (Heimdal and Houseknecht 2003; Kenney 2004), separate finances do not necessarily preclude financial advantage. For one, cohabiters benefit from economies of scale that may foster higher savings than maintaining separate residences. Time spent in cohabitation may shape marital wealth, as well. For married and cohabiting couples alike, the duration of coresidence positively relates to joint investing (Heimdal 2008).

Cohabitation Histories

Cohabitation is a heterogeneous experience, its meaning and function varying from a stepping-stone into marriage to a marital alternative. Moreover, these meanings differ by gender and race/ethnicity (Brown 2000; Manning 2001; Wildsmith and Raley 2006). One way to capture some of this heterogeneity is by measuring the number of cohabiting unions and dissolutions that individuals experience. For example, some married individuals cohabited just once and then married that same cohabiting partner; these “spousal cohabiters” have no experience with a cohabitation that dissolves. In contrast, other married individuals with cohabitation experience did not necessarily marry their cohabiting partner. Some experienced one cohabitation that dissolved before marrying (“one-time cohabiters”), and others cohabited and broke up multiple times before marrying (“serial cohabiters”).

Distinguishing these cohabitation types is important because research has shown that they are related to distinct marital outcomes. For example, spousal cohabiters have likelihoods of divorce (Teachman 2003) and levels of marital quality (Brown and Booth 1996) that are similar to those of married couples who never cohabited. In contrast, serial cohabiters have higher likelihoods of divorce (Lichter and Qian 2008; Teachman 2003) and lower marital quality (Brown and Booth 1996). Spousal cohabiters, therefore, experience marital outcomes that are often distinct from the outcomes of individuals whose cohabitations dissolved.

It is plausible that cohabitation dissolution influences marital wealth. Individuals who experienced a dissolved cohabitation gain firsthand knowledge of relationship instability and the difficulty of untangling commingled assets in the absence of legal safeguards for cohabiting unions. While other couples may continue saving and investing, individuals whose cohabitations dissolved are starting over with potentially diluted assets. Dissolution experience may even inhibit future joint investments and asset acquisition. Married persons with a history of cohabiting dissolution experience may be reluctant to invest because relationship instability means that joint investments are risky and financial returns are uncertain (van Lange et al. 1997).

Fertility

Undoubtedly, children are expensive. The estimated cost of raising a child can total $300,000 (Lino 2008), not counting lost wages (Raley et al. 2006) that may drain savings and increase debt. Consequently, parents are less likely to attain affluence over their life course than childless couples (Hirschl et al. 2003). Nonmarital fertility may be particularly important for wealth. Nearly 40% of nonmarital births occur within cohabitation (Bumpass and Lu 2000), and compared with married parents, cohabiting parents receive fewer financial transfers from family (Hao 1996) and spend more on personal luxuries while saving less (DeLeire and Kalil 2005). Considering that one-half of cohabiting births are unplanned (Musick 2002), many cohabiting parents may be unprepared for the financial burdens of childbirth. Nonmarital fertility is not evenly distributed across cohabiters, though. Compared with white women, black and Hispanic women have higher levels of unplanned and nonmarital fertility (Manning 2001; Ventura and Bachrach 2000). Thus, not only may nonmarital fertility hamper wealth as parents struggle to recoup lost savings and foregone investments, but nonwhites may be disproportionately affected.

Current Study

Past research on wealth and relationship history has typically compared wealth differences between the married and either nonmarried or remarried individuals. This study explores wealth differences among the married in light of a common relationship experience: premarital cohabitation. Because of cohabitation’s heterogeneity and selectivity, we face challenges in exploring marital wealth by cohabitation history. First, such an analysis requires longitudinal data on assets, marriage, and cohabitation, and also requires an analytic strategy that models change over time. Second, an analysis should model not only factors important for wealth (e.g., education, income, inheritance) but also characteristics of cohabitation that may influence wealth accumulation during marriage (e.g., length of cohabitating unions, dissolution experience, nonmarital fertility). Last, an analysis must consider selection: whether the results are robust across race/ethnicity, considering their selectivity in union experiences and differences in wealth accumulation, and whether the results are robust across time as marriages with cohabitation history disproportionately divorce.

Method

Data and Sample

We use the National Longitudinal Survey of Youth 1979 (NLSY79), a longitudinal study of a nationally representative U.S. cohort born between 1957 and 1964. Respondents were interviewed annually from 1979 until 1994, and biennially thereafter. The survey has had a retention rate of about 80% of respondents still eligible for interview through 2004. We use these data because they capture a period of adulthood when cohabitation and marriage are common. They also contain extensive information on relationship history and wealth. One limitation of the NLSY79 is that the data do not capture cohabitations between waves, which may underestimate short-term cohabitations. Nevertheless, research has used these data to study cohabitation to good effect (Avellar and Smock 2005; Oppenheimer 2003).

The sample consists of individuals who married for the first time in 1985 or later.1 Using the household roster and a count of respondents’ total partners/spouses, we identify four unique cohabitation histories among the married (summarized in Table 1). The most common is direct marriage, in which individuals have no prior cohabitation experience and transition directly from never married to married without cohabiting. Spousal cohabiters have only one cohabiting partner who then becomes their spouse (the directly married and spousal cohabiters have the same total number of partners). Before marrying, one-time cohabiters have had one cohabitation that ends in dissolution, and serial cohabiters have had at least two unique cohabitations.
Table 1

Operationalization of cohabitation histories, NLSY79 (N = 4,205)

Cohabitation Histories

Number of Cohabiting Partners

Total Partners (including spouse)

N

Marital Years

Directly Married

0

1

2,701

22,390

Spousal Cohabiters

1

1

933

6,884

One-Time Cohabiters

1

2

321

1,834

Serial Cohabiters

2 or more

2 or more

250

1,265

We construct a person-year sample using every year of data from 1985 to 2004. Respondents contribute marital years to the sample beginning with the first year of their first marriage. They continue contributing until the year of their marital dissolution or attrition from the sample. The total sample contains 4,205 married individuals (32,373 marital years).2 Of this sample, 2,701 were directly married (i.e., had no premarital cohabitation experience). The remaining 1,504 married individuals had some kind of cohabitation history prior to marriage: 933 were spousal cohabiters, 321 were one-time cohabiters, and 250 were serial cohabiters. The sample includes 2,270 whites, 1,180 blacks, and 755 Hispanics.3

Outcome Variable

The outcome variable is net worth (the value of assets less debts) measured in U.S. dollars for each survey year in the sample.4 Assets include the value of financial investments, such as checking and savings accounts, retirement accounts, and stocks. Assets also include the value of nonfinancial holdings, such as homes, automobiles, and other valuable possessions. We weigh the value of these assets against total debts (credit cards, hospitals bills, student loans, mortgages, liens). To correct for skew, we add a constant and then log net worth.

Explanatory Variables

The primary explanatory variables are cohabitation histories that we measure with four dichotomous variables, noted in Table 1 (directly married is the reference). Two continuous variables measure the years spent in spousal cohabitation and years spent in one-time and serial cohabitations. We measure gender (male is the reference) and race/ethnicity (white is the reference; black and Hispanic are alternate categories). For nonmarital fertility, we include a continuous measure of the number of nonmarital births by the time of marriage. Other explanatory variables include a continuous indicator of time (marital duration) measured in years, and its square to account for curvilinearity. Last, we use logged family income to account for the resources of both partners at the time of marriage.

Controls

Unless otherwise noted, controls are measured in 1979 and are time constant. A dichotomous variable measures childhood family structure (two biological parents is the reference compared with other family forms). Childhood religious affiliation affects adult wealth, so we control for Catholic, mainline Protestant, and black Protestant affiliations (Keister 2003; see also Steensland et al. 2000). Because the NLSY79 does not contain parental wealth, we use proxies including parent’s education which is measured with dichotomous variables: no high school diploma (reference), high school diploma, some college, bachelor’s degree, and advanced degree. We include logged parents’ income and a dichotomous variable for parental work status (non-full-time employment is the reference), both measured in 1978.

Other characteristics include respondent’s age5 and education at marriage,6 the latter measuring the highest attainment of either spouse (constructed in the same way as parental education). We use a time-varying continuous variable and its square to measure the effects of the number of children in the household. To control geographic advantage (Hao 1996), we include a time-varying dichotomous variable for rural/urban residency (rural is the reference) and four time-varying dichotomous variables capturing census regions (Northeast is the reference; alternate categories are North Central, South, and West). For financial resources, we use a time-varying variable for logged family income. A time-constant dichotomous variable measures whether respondents ever received an inheritance, and a time-varying control captures the logged value of received financial transfers. We control for Hispanic nativity because native-born Hispanics cohabit more often than their foreign-born counterparts (Brown et al. 2006).

Analytical Strategy

We model marital wealth using multilevel growth curves (Raudenbush and Bryk 2002; Singer and Willett 2003). Growth curves use a hierarchical specification and nest time (marital duration) within individuals. This method lets us assess how time-varying (marital duration) and time-constant (cohabitation history) variables relate to wealth trajectories over the duration of marriage. We interact cohabitation histories with marital duration and its square to account for nonlinear patterns of wealth accumulation. Results therefore illustrate (1) the relationship between cohabitation and marital wealth at the beginning of marriage (other covariates controlled), and (2) the relationship between cohabitation and marital wealth over the duration of marriage (other covariates controlled). To correct for heteroskedastic and correlated measurement errors across time, we use robust standard errors and assume a first-order autoregressive structure.

We specify four models. Model 1 (Table 3) shows the relationship between cohabitation histories and marital wealth. We interact several variables with marital duration, including cohabitation histories and nonmarital children. Because cohabitation and marriage are selective, we conduct two sensitivity tests. First, we reestimate Model 1 by race/ethnicity (in Models 2–4). We do this because wealth accumulation may differ within and between each racial/ethnic group by cohabitation history7 and because many of the predictors vary by race/ethnicity.8 To aid comparisons, we conduct statistical tests that evaluate regression coefficients between racial/ethnic groups (Clogg et al. 1995). Second, we explore wealth accumulation across marital duration to better account for selection out of marriage. Wealth accumulation is a long-term process, so this strategy identifies whether those who remain married bias results. To explore this possibility, we reestimate Model 1 stratifying by the number of years married across marriages lasting 3, 5, 7, and 10 or more years.

Results

Descriptives: Cohabitation History and Marital Characteristics

Table 2 presents descriptive statistics for the explanatory variables and wealth (Table 5 in the appendix contains descriptive statistics for the controls). About one-third of married individuals have cohabitation experience. Although this figure is smaller than current estimates, NLSY79 data are older and may underestimate short cohabitations, as noted. Of those with cohabitation history, spousal cohabiters make up the largest group (22% of the sample), followed by one-time cohabiters (8%) and serial cohabiters (6%). Average marital duration for the entire sample and the directly married is almost 7 years, whereas married individuals with cohabitation history have been married for less time. Race/ethnicity and gender are relatively evenly distributed across cohabitation histories. On average, those with cohabitation history have more nonmarital children than the directly married. Serial cohabiters also cohabited for the longest time prior to marriage (5 years), followed by spousal (3 years) and one-time (2 years) cohabiters.
Table 2

Means, medians, and standard deviations (in parentheses) for explanatory variables and wealth accumulation, NLSY79 (N = 4,205)

 

Total Sample

Directly Married

Spousal Cohabiters

One-time Cohabiters

Serial Cohabiters

Proportion of Sample

0.64

0.22

0.08

0.06

Marital Duration (years)

6.75

6.97

6.50

5.88

5.50

(4.92)

(5.02)

(4.75)

(4.50)

(4.30)

White

0.54

0.54

0.58

0.45

0.51

Black

0.28

0.28

0.26

0.34

0.32

Hispanic

0.18

0.19

0.16

0.19

0.18

Female

0.45

0.44

0.47

0.47

0.46

At Time of Marriage

 Age (years)

28.72

27.84

29.09

31.63

33.09

(5.06)

(4.76)

(4.64)

(5.58)

(5.31)

 Family incomea ($)

65,716

63,990

66,559

70,629

74,752

(119,518)

(120,650)

(105,771)

(132,169)

(138,546)

 Bachelor’s degree or higher

0.34

0.37

0.29

0.31

0.26

 Number of nonmarital children

0.60

0.50

0.77

0.77

0.88

(0.97)

(0.89)

(1.07)

(1.05)

(1.14)

 Years of spousal cohabitation

0.79

2.99

(1.95)

 

(2.76)

  

 Years of other cohabitation

0.34

2.25

5.00

(1.22)

  

(2.09)

(2.63)

Wealth Across Marital Duration ($)

 1st year mean

58,707

53,964

65,190

77,011

64,864

(180,822)

(153,880)

(214,040)

(293,690)

(130,807)

 1st year median

15,032

14,966

14,484

14,453

18,059

 3rd year mean

87,322

83,043

92,485

105,837

93,750

(228,265)

(220,865)

(254,726)

(240,642)

(165,006)

 3rd year median

23,632

23,703

21,017

33,530

35,517

 5th year mean

110,987

104,538

118,993

118,208

155,729

(26,373)

(229,123)

(315,242)

(252,234)

(419,588)

 5th year median

36,735

37,867

32,781

35,603

47,161

 10th + year mean

237,846

226,755

285,045

180,538

278,216

(502,474)

(451,108)

(683,954)

(293,535)

(453,113)

 10th + year median

99,179

98,728

97,488

86,278

130,541

N

4,205

2,701

933

321

250

aConverted to US$ 2004 with the Consumer Price Index.

Descriptive statistics for wealth accumulation show some intriguing patterns. Those with one-time and serial cohabitation histories appear advantaged early in marriage over spousal cohabiters and the directly married. Yet among marriages of the longest duration, spousal cohabiters have the highest average wealth, thus outpacing other married individuals. Several factors may shape these patterns. First, at the time of marriage, one-time and serial cohabiters are older and have higher incomes than the directly married or spousal cohabiters, which may boost their wealth accumulation early in marriage. Second, individuals with a cohabitation history and poor economic forecasts disproportionately divorce over time. Therefore, marriages of the longest duration are selective of individuals with the greatest wealth potential. These factors may explain one-time cohabiters’ early edge, which then dissipates. Curiously, serial cohabiters do not follow the same trend. Their greater income at marriage may offset the effects of their higher likelihood of divorce, leaving a small pool of wealthy former serial cohabiters among those still married after 10 years. These patterns emphasize the importance of examining wealth trajectories by cohabitation history while accounting for resources at the start of marriage and selection out of marriage over time.

Regression Results: Growth Curves

Table 3 shows results from growth curve analyses. To conserve space, we present the explanatory variables (results are from the full model; Table 6 in the appendix shows the analyses for the controls). Parameter interpretation depends on whether the explanatory variable is time varying, time constant, or an interaction. Time-varying slope parameters (marital duration and its square) describe the average rate of change in marital wealth across individuals per year of marriage. Interactions with marital duration allow the slopes of wealth trajectories to change over time. Other time-varying variables (logged family income) represent average differences over time in the slope of wealth trajectories. Time-constant variables (cohabitation histories) and their interactions describe wealth differences in the intercept or initial status at the first year of marriage that separate the trajectories across all years of marriage.
Table 3

Growth curve parameter estimates of cohabitation histories on marital wealth accumulation (logged dollars), NLSY79 (N = 4,205)

 

Model 1

Model 2

Model 3

Model 4

Full Sample

Whites

Blacks

Hispanics

Initial Status

Cohabitation history (ref. = directly married)

Spousal cohabiters

–0.051***

–0.060**

–0.014

–0.043

(0.014)

(0.019)

(0.017)

(0.023)

One-time cohabiter

0.001

0.012

0.019

–0.015

(0.019)

(0.032)

(0.014)

(0.025)

Serial cohabiters

–0.017

–0.045a

0.029

–0.019

(0.018)

(0.028)

(0.021)

(0.037)

Personal traits

Black (ref. = white)

–0.043***

(0.009)

   

Hispanic (ref. = white)

–0.024*

(0.011)

   

Female (ref. = male)

–0.004

–0.006

0.000

0.001

(0.006)

(0.009)

(0.008)

(0.011)

Number of nonmarital children

0.007

–0.002

0.010

0.015

(0.005)

(0.009)

(0.005)

(0.008)

Income at time of marriage

Directly married

0.037***

0.059***a

0.007

0.024**

(0.008)

(0.016)

(0.006)

(0.009)

Spousal cohabiters

0.026

0.015

0.047

0.009

(0.021)

(0.033)

(0.025)

(0.016)

One-time cohabiters

–0.020

–0.048†a

0.030

0.010

(0.023)

(0.026)

(0.017)

(0.031)

Serial cohabiters

–0.019

0.018

0.015

–0.020

(0.012)

(0.028)

(0.019)

(0.013)

Cohabitation experience

Years of spousal cohabitation

0.004

0.004

0.004

0.011

(0.003)

(0.005)

(0.004)

(0.007)

Years of nonspousal cohabitation

–0.005*

–0.006

–0.002

–0.002

(0.003)

(0.006)

(0.003)

(0.004)

Rate of Change

Marital duration (years)

0.010***

0.012***

0.011***

0.015***

(0.002)

(0.002)

(0.002)

(0.002)

Squared marital duration

0.000***

0.000*

(0.000)

(0.000)

  

Marital duration  ×  Nonmarital children

–0.004***

–0.004**

–0.003***

–0.004***

(0.001)

(0.001)

(0.001)

(0.001)

Interaction with marital duration (ref. = directly married)

Spousal cohabiters

0.010**

0.013*b

0.002

–0.001

(0.004)

(0.005)

(0.003)

(0.002)

One-time cohabiters

0.002

0.002

0.002

–0.005

(0.007)

(0.010)

(0.004)

(0.003)

Serial cohabiters

0.002

0.007

–0.002

0.000

(0.004)

(0.007)

(0.004)

(0.004)

Interaction with squared marital duration (ref. = directly married)

Spousal cohabiters

–0.000**

–0.000*

(0.000)

(0.000)

  

One-time cohabiters

0.000

0.000

(0.000)

(0.001)

  

Serial cohabiters

0.000

0.000

(0.000)

(0.000)

  

N

4,205

2,270

1,180

755

Notes: Standard errors are in parentheses. See Table 6 in the appendix for the analysis of the controls. For determining significance across models, see the analytic strategy.

aWhites and blacks differ at p  <  .05.

bWhites and Hispanics differ at p  <  .05.

p  <  .10; *p  <  .05; **p  <  .01; ***p  <  .001 (two-tailed tests).

In Model 1, the coefficient for marital duration represents the growth rate for the directly married who have no nonmarital births. Each year of marriage for this group significantly correlates with a 1% (= e0.01 – 1) increase in net worth. The positive coefficient for squared marital duration suggests that the wealth accumulation of the directly married increases at an accelerating rate over time. Individuals with spousal cohabitation experience are correlated with 5% (= e–0.051 – 1) less net worth at the time of marriage than the directly married. However, spousal cohabiters have the only cohabitation history that significantly interacts with marital duration: individuals who married their only cohabiting partner experience a 1% (= e0.01 – 1) wealth premium per marital year over the directly married. Thus, for each year of marriage, spousal cohabiters accumulate wealth at double the rate of directly married individuals (2% [= e0.01 + 0.01 – 1] and 1%, respectively). This premium allows married individuals with a history of spousal cohabitation to close the initial wealth gap with the directly married over time. The coefficient for the interaction between spousal cohabitation and the square of marital duration is negative, however. Thus, spousal cohabiters initially accumulate wealth more rapidly than the directly married, but the rate decelerates over time. For married individuals with one-time and serial cohabitation histories, neither the initial status nor rate of change is significant. Married persons with these cohabitation histories, therefore, experience wealth trajectories that are indistinguishable from those of the directly married.

Apart from cohabitation history, other factors shape marital wealth. Blacks and Hispanics begin marriage with less wealth than whites, although there is no difference between men and women. (In supplemental analyses, interactions between female and cohabitation history were not significant.) The rate of wealth accumulation for married individuals with nonmarital children is only one-half that of married individuals who did not have a nonmarital birth (0.6% [= e0.01 – 0.004 –1] and 1%, respectively). Notably, time spent cohabiting detracts from initial levels of wealth at the time of marriage but only for those with a history of cohabitation dissolution. It is not these types of cohabitation per se that hinder wealth accumulation; if it were, then the variables for one-time and serial cohabitation would be significant. Rather, what is negatively correlated with marital wealth is how long individuals spent in premarital cohabitations that dissolved.

Sensitivity Analysis: Race and Ethnicity

Models 2–4 estimate separate analyses by race/ethnicity (Table 3). Across racial/ethnic groups, married individuals remain associated with similar levels of wealth accumulation (coefficients for marital duration are statistically indistinguishable across race/ethnicity). Over time, though, the wealth of married whites outpaces that of married nonwhites (positive and significant squared term).9 We also observe wealth equality by cohabitation histories for blacks at the beginning of marriage, but not for whites or Hispanics. Among whites and Hispanics, spousal cohabiters have less wealth at the time of marriage than their directly married counterparts. Notably, only among whites do spousal cohabiters continue to experience a wealth premium after marriage (a premium of 1.3% [= e0.013 – 1] per marital year over directly married whites). This premium more than doubles spousal cohabiters’ rate of wealth accumulation to 2.5% (= e0.012 + 0.013 – 1) per marital year versus 1.2% (= e0.012 – 1) for the directly married. As in the full sample, this advantage erodes slowly over time (negative and significant interaction between spousal cohabitation among whites and squared marital duration). In contrast, wealth accumulation does not significantly differ over time by cohabitation history among blacks or Hispanics.

Figure 1 shows predicted values from Models 2–4 (Table 3) of net worth by cohabitation history and race/ethnicity (holding all other variables at their group-centered means). The figure presents trajectories for the first 10 years of marriage; after that time, the data become thin, with less than one-quarter of the sample’s total marital years remaining. We include the directly married from each racial/ethnic group for reference and spousal cohabiter trajectories for whites and Hispanics.10 On average, married individuals gain wealth over time regardless of cohabitation history or race/ethnicity. Nevertheless, Fig. 1 illustrates substantial racial/ethnic wealth inequality at the beginning and over the duration of marriage. At the beginning of marriage, directly married whites have the highest net worth, Hispanic spousal cohabiters have the least, and directly married blacks are in between. Only white spousal cohabiters show a significantly greater trajectory of wealth accumulation than directly married whites. This advantage tempers over time so that after 15 years the wealth trajectory of white spousal cohabiters begins diverging from the trajectory of the directly married. The data are thin for marriages of longer duration, though, with less than 9% of the sample’s total marital years remaining after 15 years. In contrast to whites, Hispanics show similar rates of wealth accumulation regardless of cohabitation history. Therefore, Hispanic spousal cohabiters do not close the wealth gap over time with their directly married counterparts.
https://static-content.springer.com/image/art%3A10.1007%2Fs13524-011-0043-2/MediaObjects/13524_2011_43_Fig1_HTML.gif
Fig. 1

Predicted values of net worth (logged dollars), by cohabitation history and race/ethnicity (NLSY79)

Sensitivity Analysis: Marital Duration

Table 4 reestimates Model 1 (Table 3) stratified by marital duration. For brevity, we present explanatory variables (results are from the full model). The bottom panel of Table 4 shows attrition across marital duration; about one-half of the analytic sample remains after 10 years. Spousal cohabiters and the directly married have similar rates of attrition, but one-time and serial cohabiters have higher rates. The small number of respondents with marriages lasting 10 years or longer (particularly one-time and serial cohabiters) urges caution when interpreting trends among this group. Across marital duration categories, findings are robust compared with Model 1 (Table 3). Married individuals remain positively correlated with wealth, and spousal cohabiters’ wealth premium over the directly married remains significant. Even when accounting for cohabiters’ disproportionate selection out of marriage, spousal cohabiters retain a wealth premium over the directly married, while the wealth patterns of individuals with other cohabitation histories remain no different from those of the directly married.
Table 4

Sample composition and marital wealth trajectories (logged dollars) by cohabitation history and marital duration, NLSY79 (N = 4,205)

 

Marital Duration

3 Years

5 Years

7 Years

10+ Years

Initial Status

Cohabitation history (ref. = directly married)

Spousal cohabiters

–0.048***

–0.052***

–0.049**

–0.067***

(0.014)

(0.015)

(0.016)

(0.020)

One-time cohabiters

0.004

0.021

0.026

0.008

(0.020)

(0.024)

(0.028)

(0.020)

Serial cohabiters

–0.007

–0.009

–0.011

–0.025

(0.020)

(0.020)

(0.021)

(0.026)

Rate of Change

 Marital duration (years)

0.010***

0.010***

0.011***

0.011***

(0.001)

(0.001)

(0.001)

(0.002)

 Squared marital duration

0.001**

0.001**

0.000***

0.000***

(0.000)

(0.000)

(0.000)

(0.000)

Interaction with marital duration (ref. = directly married)

Spousal cohabiters

0.010**

0.010**

0.011**

0.014***

(0.003)

(0.003)

(0.003)

(0.004)

One-time cohabiters

0.001

–0.001

–0.002

–0.006

(0.006)

(0.006)

(0.003)

(0.005)

Serial cohabiters

0.001

0.002

0.001

0.000

(0.004)

(0.004)

(0.004)

(0.006)

Interaction with squared marital duration (ref. = directly married)

Spousal cohabiters

–0.000**

–0.000**

–0.000**

–0.000**

(0.000)

(0.000)

(0.000)

(0.000)

One-time cohabiters

–0.000

0.000

0.000

0.000

(0.000)

(0.000)

(0.000)

(0.000)

Serial cohabiters

0.000

0.000

0.000

0.000

(0.000)

(0.000)

(0.000)

(0.000)

Sample Composition (N  =  4,205)

3,609

3,168

2,811

2,272

 (%)

(86)

(75)

(67)

(54)

Cohabitation History

 Directly married (n  =  2,701)

2,338

2,079

1,871

1,583

 (%)

(87)

(77)

(69)

(59)

 Spousal cohabiters (n  =  933)

809

716

641

494

 (%)

(87)

(77)

(69)

(53)

 One-time cohabiters (n  =  321)

262

208

168

121

 (%)

(82)

(65)

(52)

(38)

 Serial cohabiters (n  =  250)

200

165

131

74

 (%)

(80)

(66)

(52)

(30)

Notes: Standard errors are in parentheses. For brevity, controls are not shown but the results are from the full model.

p  <  .10; *p  <  .05; **p  <  .01; ***p  <  .001 (two-tailed tests)

Summary and Discussion

Building from past research on wealth and relationship history (Hao 1996; Hirschl et al. 2003; Holden and Kuo 1996; Waite and Gallagher 2000; Wilmoth and Koso 2002), our study examines married persons’ patterns of wealth accumulation across time in light of their cohabitation histories. We differentiate married individuals by the number of premarital cohabitations they have experienced and whether they married their only cohabiting partner (i.e., “spousal” cohabitation).

To date, scholarly findings imply that cohabitation is “bad” for marriage because it negatively affects several marital outcomes. Our findings depart from this convention. Marriage has a strong, positive relationship with wealth accumulation regardless of race/ethnicity or cohabitation history. This is not to say that all marriages generate equal levels of wealth, especially given existing racial/ethnic wealth inequality. Rather, the married tend to accumulate wealth at similar rates over time. This finding suggests that a common financial experience characterizes marriage and that the patterns underlying wealth accumulation within marriage operate similarly for whites, Hispanics, and blacks. Moreover, individuals who married their only cohabiting partner experience a rate of wealth accumulation that is twice as large as the rate of married individuals who never cohabited, even when we account for education, income, fertility, cohabitation duration, and selection out of marriage. Spousal cohabiters’ wealth premium appears to attenuate over time and is evident only among whites. Thus, although cohabitation history does not shape the wealth patterns of married blacks and Hispanics, it does distinguish the wealth patterns of married whites.

The wealth premium for spousal cohabiters is an intriguing finding and has several implications for future research. To our knowledge, we present the first evidence that cohabitation is positively related to a marital outcome. We uncover this finding because we focus on distinct cohabitation histories among the married. Our results, therefore, highlight the importance of distinguishing spousal cohabiters from one-time and serial cohabiters (i.e., those with and without dissolution experience). Indeed, only spousal cohabiters experience a unique pattern of marital wealth accumulation, and time spent in one-time and serial cohabitation is correlated with a financial deficit at the start of marriage. The longer individuals spend in cohabitations that do not result in marriage, the lower their wealth at the start of marriage. One-time and serial cohabiters’ experience with relationship instability may discourage investment in future relationships (van Lange et al. 1997). The implication of these findings is that what shapes the consequences of cohabitation for marriage may not be whether one cohabits but whether one experiences the dissolution of a cohabiting union. Cohabitation dissolution is therefore important to consider when conceptualizing different histories or experiences of cohabiting unions. Interestingly, cohabiting dissolution experiences do not inhibit wealth during marriage, suggesting that one-time and serial cohabiters wait until marriage to adopt wealth-promoting behaviors. It appears that marriage may reorient their financial behaviors to match those of individuals who marry without prior cohabitation experience.

Although spousal cohabiters’ wealth premium is an intriguing finding, the mechanism driving the result remains elusive. It is possible that the premium relates to the experience of cohabitation itself. Cohabitation may provide a period before marriage that “lays the groundwork” for future wealth accumulation via the experience of managing a financial household. Many cohabiters, low-income couples, and single parents view marriage as something that happens after they achieve economic stability. To these groups, their perception of the feasibility of marriage rests on economic prerequisites that signal whether they are financially secure and thus ready for marriage. These prerequisites can include completing one’s education, attaining job stability, getting out of debt, and saving for homeownership or even a wedding (Gibson-Davis et al. 2005; Sassler 2004; Smock et al. 2005). Falling short of these prerequisites casts doubts over future relationship stability and is a reason for delaying marriage (Sassler 2004; Smock et al. 2005). The perceived need to meet these prerequisites may drive spousal cohabiters to change their financial behaviors accordingly and work toward meeting their economic goals. In doing so, spousal cohabiters may galvanize their wealth accumulation and thus experience an accelerated rate of accumulation during marriage.

Alongside these findings, we should consider the study’s limitations. We are limited in our ability to “look inside” cohabiting unions to understand how cohabiters’ commitment, trust, and financial behaviors may shape marital wealth outcomes. Although cohabiters’ expectation to marry does not vary greatly by race/ethnicity (Manning and Smock 2002), the meaning and function of cohabitation do tend to vary among blacks, whites, and Hispanics (Brown 2000; Manning 2001; Wildsmith and Raley 2006). Distinguishing cohabiters by their dissolution experience and separating analyses by race/ethnicity, however, are proxies for modeling the heterogeneity of cohabitation. The data also underestimate short-term cohabitations and may undercount serial cohabiters or include some cohabiters among the directly married. The relatively small sample size for serial cohabiters (i.e., 250) could explain why we observe a negative, although nonsignificant, relationship with wealth at the time of marriage. Last, the data contain few individuals with lengthy marriages (less than 25% of marital years remain after a decade, and less than 10% remain after 15 years). Thus, the data are underpowered for long marriages, and spousal cohabiters’ declining wealth premium should be interpreted cautiously because it may be a real phenomenon or an artifact of thin data.

Nonetheless, we can draw important conclusions from this study. Not only do marriages accumulate wealth over time regardless of race/ethnicity or cohabitation history, but spousal cohabitation is associated with a beneficial marital outcome. Exploring the diversity of cohabitation histories among the married is, therefore, critical for understanding marital outcomes—an importance that may grow as cohabitation becomes more normative. Moreover, experiencing a cohabiting dissolution, rather than the experience of cohabitation itself, may be the linchpin that shapes marital outcomes, at least financial ones. If so, then any marital financial advantage associated with cohabitation would be restricted to first-time cohabiters who transition to marriage, while serial cohabiters may, over time, face a cumulative disadvantage from multiple dissolutions. These findings urge us to better understand cohabitation’s varied meanings and roles, as well as its relationship with financial behaviors and wealth. Importantly, these findings encourage us to reconsider cohabitation’s negative framing and search for new ways that cohabitation could be beneficial.

Footnotes
1

NLSY79 did not collect wealth data until 1985. Compared with post-1985 marriages, respondents in pre-1985 marriages grew up in less-educated and lower-income households. At marriage, they were younger, had less cohabitation experience, and were less likely to have had a nonmarital birth. They also had lower incomes and were less likely to own a home or to have more than a high school education. Excluding these marriages likely biases results upward because it omits marriages with lower wealth potential.

 
2

Missing data reduced the sample size slightly. Experiments with multiple imputation did not change the results.

 
3

White refers to non-Hispanic white; black, to non-Hispanic black.

 
4

We adjust for inflation to 2004 dollars, using the Consumer Price Index.

 
5

In a supplemental analysis of the directly married and spousal cohabiters, we examined respondent’s age at union formation rather than age at marriage to better account for cohabiters’ longer coresidence with their partner. Results were equivalent, so we present results using age at marriage with the full sample.

 
6

Results were robust when we substituted respondent’s education for the household-level measure.

 
7

We also interacted gender with cohabitation histories and marital duration because of gendered differences in wealth. None of the interactions was significant. For parsimony, we excluded them from the final models.

 
8

Cohabitation also varies among ethnicities. Puerto Ricans are more accepting of cohabitation and cohabit more often than Mexican Americans (Oropesa 1996). In supplemental analyses, we examined Mexicans and Mexican Americans, the largest Hispanic groups in the data. Results remained unchanged from the pooled Hispanic model.

 
9

In additional analyses, we included the square of marital duration for blacks and Hispanics. These terms were not significant, so we estimated a linear specification.

 
10

We include these cohabitation histories because their wealth patterns significantly differ from their respective directly married peers. For this reason, we include only black directly married persons. The figure does account for the curvilinear results of spousal cohabitation among whites. Because the squared term is small, the curvature is not particularly pronounced during the first 10 years of marriage.

 

Acknowledgments

For their advice and time spent reviewing this article, we thank Kenneth Land, the anonymous reviewers of Demography, Zhenchao Qian, Bob Kaufman, Randy Olsen, Randy Hodson, Jamie Lynch, and Adrianne Frech. We would also like to acknowledge the helpful comments of Wending Manning, Pamela Smock, and Kelly Raley from the session on cohabitation during the 2009 annual meetings of the Population Association of America in Detroit, MI, where an earlier version of this research was presented.

Copyright information

© Population Association of America 2011