Gray Divorce During the COVID-19 Pandemic

Abstract

Objectives

Drawing on emerging evidence that the pandemic appears to have impeded both the divorce process and actual divorces, we examined whether the gray divorce rate (i.e., divorce among adults aged 50+) declined following the onset of the COVID-19 pandemic.

Methods

Data from the 2019 and 2021 American Community Survey (ACS) were used to track changes in gray divorce. With the 2021 ACS, we estimated pandemic-era gray divorce rates across sociodemographic subgroups for middle-aged and older adults. We then pooled the 2019 (N = 892,700) and 2021 (N = 898,828) data to examine whether the risk of divorce changed with the onset of the pandemic net of sociodemographic characteristics, distinguishing trends for middle-aged versus older adults.

Results

The gray divorce rate dropped following the onset of the pandemic. This drop was more pronounced among middle-aged than older adults. For older adults, the divorce rate essentially stalled.

Discussion

The gray divorce rate now mirrors the overall trend of modest decline in U.S. divorce patterns. Whether the gray divorce rate continues to shrink as society transitions to a postpandemic environment awaits future research.

Over the past half century, the risk of divorce among older adults has soared. The rate of gray divorce, which describes divorces that occur to adults aged 50 and older, rose from 3.7 divorcing persons per 1,000 married persons in 1970 to 4.9 per 1,000 in 1990 and then doubled to 10.1 per 1,000 in 2010 (Brown & Lin, 2022). By 2019, the gray divorce rate had stabilized, but only among middle-aged adults between 50 and 64 years of age. For adults aged 65+, the gray divorce rate continued its ascent. Although historically gray divorce was uncommon, more than one in three people divorcing these days is aged 50 or older, and one in ten is aged 65 and older, signaling the graying of divorce (Brown & Lin, 2022).

Whether and how the gray divorce rate has changed since the onset of the COVID-19 pandemic is unknown. The pandemic had an immediate disproportionate impact on older adults that continues even now. Older adults remain highly susceptible to severe COVID-19, comprising the vast majority of hospitalizations and deaths (U.S. Center for Disease Control and Prevention, n.d.-a). In fact, the high mortality rate among older adults translates into fewer remaining married adults who are eligible to get divorced, foretelling diminished chances of gray divorce in the pandemic era. Moreover, the severity of the health risk posed by COVID-19 has contributed to greater physical and social isolation among older adults (Peng & Roth, 2022). Both anxiety about contracting COVID-19 (Pearman et al., 2021) and expected pandemic-related income declines (Whitehead, 2021) have further elevated stress among older adults. Stress is often linked to lower marital quality (Harper et al., 2000; Karney & Bradbury, 1995), suggesting that gray divorce may have increased during the pandemic. However, a recent national study on older couples’ relationship dynamics during the pandemic revealed that for most couples, relationship quality either remained stable (66%) or improved (23%) compared with before the start of the pandemic (Wong et al., 2023). Given the robust negative association between marital quality and gray divorce (Karraker & Latham, 2015; Lin et al., 2018), these findings signal that gray divorce may have declined during the pandemic.

Research on divorce patterns during the pandemic is scant. Two recent assessments of divorce count at the state level revealed modest declines from 2018/2019 to 2020 in most states (Manning & Payne, 2021; Westrick-Payne et al., 2022). Family therapists pointed to the practical constraints associated with the pandemic that likely curtailed the divorce process for many couples, noting that initiating periods of marital separation became more difficult during lockdown (Lebow, 2020). Similarly, access to the legal system was hampered during the early months of the pandemic, leading to an ongoing backlog of cases, which in turn could forestall the divorce process. Even as the pandemic may have shifted the relationship dynamics for some couples such that they were more inclined to divorce, the shrinking practicability of divorce likely has led to fewer divorces since the pandemic began.

In this brief report, we test whether the gray divorce rate significantly declined during the pandemic (i.e., between 2019 and 2021) for adults aged 50 and older as well as separately for middle-aged (50–64) and older (65+) adults since the two groups exhibited distinctive patterns as of 2019. As Baby Boomers aged out of midlife and into older adulthood, the divorce rate of middle-aged adults stalled, whereas the uptick among older adults persisted (Brown & Lin, 2022), reflecting distinctly high levels of remarriage among Boomers (gray divorce is more than twice as common among remarriages than first marriages). Although older adults have been uniquely vulnerable during the pandemic, their experiences are likely not homogenous. For instance, Black and Hispanic adults have had higher morbidity and mortality from COVID-19 than their White counterparts (U.S. Center for Disease Control and Prevention, n.d.-b). Similarly, adults with fewer economic resources often lack adequate healthcare, worsening their COVID-19 outcomes. Sociodemographic factors not only shape how older adults experience the pandemic but also are correlates of gray divorce. Prior research has shown that gray divorce is more common among Black and Hispanic than White older adults. Older adults in remarriages (which also tend to be of shorter duration) are much more likely to experience a gray divorce than those in (long-term) first marriages. And economic resources (e.g., education, employment, and income) are protective against gray divorce (Brown & Lin, 2022; Lin et al., 2018). Thus, we pay attention to gray divorce rates across sociodemographic subgroups. Finally, we examine whether sociodemographic factors related to gray divorce account for any observed decline during the pandemic. Our study contributes to the small but growing literature on how the pandemic is reshaping family life and its implications for older adult well-being.

Method

Data came from the 2019 and 2021 American Community Survey (ACS), a large, nationally representative survey administered annually by the U.S. Census Bureau. Since 2008, the ACS has included marital history questions that gauge whether respondents divorced in the past year, which allowed us to calculate a divorce rate by age. Our analytic sample included middle-aged (50–64) and older (65+) adults at risk of divorce (defined in the next paragraph) to compare their trends since recent work shows divorce has stagnated among middle-aged adults but is still rising among older adults (Brown & Lin, 2022). In 2019, the ACS included 3,239,553 persons, of whom 892,714 were aged 50 and older and at risk of gray divorce. After removing the 14 cases for which marriage order could not be determined, the final sample size was 892,700, including 477,127 middle-aged and 415,573 older adults. In 2021, the ACS included 3,252,599 persons, of which 898,835 were aged 50 and older and at risk of gray divorce. After removing the 7 cases for which marriage order could not be determined, the final sample size was 898,828, including 462,005 middle-aged and 436,823 older adults. We did not include the 2020 ACS for two key reasons. First, the data collection process was severely disrupted by the onset of the pandemic, leading the Census to label the 2020 ACS an “Experimental Data Product” that can only yield “experimental estimates” because they were unable to collect data from various segments of the population (U.S. Census Bureau, 2021a). Second, even if data quality were not an issue, the design of the divorce question, which asks about divorces experienced during the past 12 months, means that more than half of the divorces reported during the 2020 interview year would have occurred either in 2019 or during the first couple months of 2020, prior to the start of the pandemic, whereas the remainder would have occurred at the height of the pandemic. Our reliance on 2019 and 2021 data not only ensures high data quality but also a clear demarcation between prepandemic divorces versus those that occurred during the pandemic.

We followed Brown and Lin’s (2012, 2022) approach to calculating the divorce rate using the ACS. Specifically, we divided the number of people who reported divorcing in the past 12 months by the number who could have possibly become divorced in the past 12 months (those who were married or separated or in the past 12 months got divorced or became widowed).

We began our analyses by charting the gray divorce rate for adults aged 50 and older as well as separately for middle-aged and older adults in 2019 versus 2021. Then, we estimated 2021 pandemic gray divorce rates across sociodemographic subgroups to assess how the risk of divorce differed by the marital biography (marriage order and marital duration), demographic characteristics (gender and race-ethnicity), and economic resources (education, employment, and income). See Supplementary Table 1 for a detailed description of the sociodemographic measures. Finally, using pooled 2019 and 2021 ACS data, we conducted multivariable logistic regression analyses predicting gray divorce to test whether the risk declined during the pandemic, net of sociodemographic factors. Supplementary Table 2 depicts the results from models estimated separately by survey year (i.e., 2019 and 2021). All analyses were performed using the Census-recommended replicate weighting techniques to account for the complex sampling design of the ACS and generate robust standard errors (U.S. Census Bureau, 2014, 2022).

Results

As shown in Figure 1, the gray divorce rate declined during the pandemic, falling from 9.62 divorcing persons per 1,000 married persons in 2019 to 9.12 per 1,000 in 2021. This modest drop was evident among both middle-aged and older adults, although the reduction was larger among the former than the latter group. For middle-aged adults, the gray divorce rate fell from 12.71 per 1,000 in 2019 to 12.01 per 1,000 in 2021, a statistically significant decline. In contrast, the comparatively small decline observed for older adults from a rate of 5.59 per 1,000 in 2019 to 5.45 in 2021 did not achieve statistical significance. In 2021, 38% of all divorces were gray divorces, signaling that the age composition of divorce during the pandemic has held steady as in 2019 the figure was 36% (results not shown). About 26% of all gray divorces were among older (vs middle-aged) adults in 2021, which is comparable with the prepandemic level (25% in 2019, results not shown).

Figure 1.

The gray divorce rate in 2019 and 2021, by age group.

Table 1 depicts gray divorce rates across sociodemographic subgroups for adults aged 50 and older as well as separately for middle-aged and older adults. Across all subgroups, the divorce rate was higher for middle-aged than older adults (statistical tests shown in the rightmost column of the table). Subgroup variation followed similar patterns for middle-aged and older adults and thus we focus on the findings obtained for adults aged 50 and older. The divorce rate was nearly twice as high for adults in higher order (13.22 per 1,000) versus first (7.33 per 1,000) marriages. Divorce rates diminished as marital duration rose. Women and men exhibited comparable divorce rates. Gray divorce rates were highest among adults identifying as Black (19.32 per 1,000), followed by Hispanic (11.78 per 1,000), or other races (11.43 per 1,000), and lastly by White (7.69 per 1,000) and Asian (6.56 per 1,000). Divorce rates were lowest for those with a college degree (7.37 per 1,000), followed by those with a high school diploma (9.68 per 1,000). Adults with either some college (10.27 per 1,000) or who did not complete high school (10.46 per 1,000) had higher divorce rates. Unemployed adults (18.27 per 1,000) faced a higher divorce rate than employed adults (11.20 per 1,000) and the rate was still lower for those not in the labor force (6.74 per 1,000). Across income groups, variation in divorce rates was modest with those in the lowest and highest income brackets having the lowest rates of divorce.

Table 1.

2021 Gray Divorce Rates Across Marital Biography, Demographic Characteristics, and Economic Resources, by Age Group

Variable	Age group			(2) vs (3)
	50+ (%)	50–64 (%)	65+ (%)
	(1)	(2)	(3)
Total	9.12	12.01	5.45	***
Marital biography
Marriage order
First marriage	7.33	9.86	3.96	***
Higher order marriage	13.22	17.24	8.59	***
Marital duration (years)
0–9	17.38	18.36	14.06	*
10–19	14.72	16.19	10.15	***
20–29	12.49	13.11	10.28	**
30–39	7.97	8.33	7.05	*
40+	3.53	5.93	3.09	***
Demographic characteristics
Gender
Women	9.22	11.89	5.58	***
Men	9.03	12.12	5.34	***
Race and ethnicity
White	7.69	10.67	4.37	***
Black	19.32	23.28	13.05	***
Hispanic	11.78	13.40	8.37	***
Asian	6.56	7.32	5.42	*
Others	11.43	13.52	7.71	***
Economic resources
Education
Less than high school	10.46	13.01	7.34	***
High school graduate	9.68	13.37	5.32	***
Some college	10.27	13.42	6.14	***
Bachelor’s degree or more	7.37	9.60	4.41	***
Employment
Employed	11.20	11.99	7.51	***
Unemployed	18.27	19.72	12.88	*
Not in labor force	6.74	11.19	4.86	***
Personal income
<10K	8.73	10.70	6.03	***
10–25K	9.57	15.24	5.94	***
25–40K	9.53	13.30	5.67	***
40–55K	9.68	13.01	4.90	***
55–70K	8.92	11.95	4.15	***
>70K	8.62	10.34	4.77	***
Unweighted N	898,828	462,005	436,823

Source: American Community Survey, 2021.

*p < .05. **p < .01. ***p < .001.

The odds ratios obtained from the multivariable logistic regression model predicting the likelihood of gray divorce are shown in Table 2. Consistent with our hypothesis, the odds of gray divorce were significantly lower during the pandemic than they were just prior to the pandemic. Relative to 2019, the odds of divorce in 2021 were only 0.95 times as large. The risk of gray divorce was lower for older than middle-aged adults and this differential was of similar magnitude in both 2019 and 2021 (i.e., an interaction between age group and survey year did not achieve statistical significance, result not shown). Adults in first marriages were less likely to divorce than their remarried counterparts. Marital duration was negatively associated with divorce. Women were more likely to divorce than men. Compared with those who identified as White, adults who identified in all other racial groups except Asian faced higher odds of divorce. Those with a college degree were less likely to divorce than those who completed high school. The odds of divorce were higher for both the employed and the unemployed compared with those not in the labor force. Income was negatively associated with divorce except among those with the lowest income level. A fully interactive model (results not shown) yielded no significant interaction effects between survey year and each covariate, except for race-ethnicity which is not readily interpretable as the measurement of race-ethnicity in the ACS changed in 2020 (U.S. Census Bureau, 2021b). Estimating the model shown in Table 2 separately by survey year provided additional confirmation that the covariates operated similarly at both time points (see Supplementary Table 2).

Table 2.

Odds Ratios From Logistic Regressions of the Likelihood of Gray Divorce in the Last 12 Months (Unweighted N = 1,791,541)

Variable	Odds ratio
Survey year
2019 (reference category)
2021	0.948	*
Age group
50–64 (reference category)
65+	0.682	***
Marital biography
Marriage order
Higher order marriage (reference category)
First marriage	0.722	***
Marital duration
0–9 years	2.817	***
10–19 years	2.504	***
20–29 years	2.466	***
30–39 years	1.637	***
40+ years (reference category)
Demographic characteristics
Gender
Women	1.067	*
Men (reference category)
Race and ethnicity
White (reference category)
Black	1.893	***
Hispanic	1.218	***
Asian	0.912
Others	1.259	***
Economic resources
Education
Less than high school	1.091
High school graduate (reference category)
Some college	1.021
Bachelor’s degree or more	0.826	***
Employment
Not in labor force (reference category)
Unemployed	1.175	***
Employed	1.794	***
Personal income
<10K	1.072
10–25K	1.276	***
25–40K	1.202	***
40–55K	1.103	*
55–70K	1.014
>70K (reference category)
Constant	0.005	***
	F(22, 58) = 192.89

Source: American Community Survey, 2019 and 2021.

*p < .05. **p < .01. ***p < .001.

Conclusion

The COVID-19 pandemic has not just reshaped individual health and well-being but also had myriad consequences across a range of life domains. Our study demonstrates that the pandemic may have contributed to a slowdown in gray divorce, particularly among middle-aged adults. The downward trend in gray divorce since the onset of the pandemic is at odds with the prepandemic trend that was marked by stagnation. However, the distinct patterns of change for middle-aged versus older adults are consonant with prior work (Brown & Lin, 2022) showing that between 2010 and 2019 divorce stalled among the former group even as it continued its ascent among the latter group. Now, middle-aged adults exhibit a decrease in divorce, whereas divorce among older adults has effectively stalled. Our study provides additional evidence that divorce among the Baby Boom cohort adheres to a unique pattern as demonstrated by the trend we uncovered for older adults. For middle-aged adults, a growing share are not Boomers but rather Generation X and their lifetime risk of divorce has trailed that of Boomers (Brown & Lin, 2022; Cohen, 2019). In sum, our expectation that the pandemic coincided with a decline in divorce was supported, albeit mainly for the middle-aged (who comprise three-quarters of gray divorces). We acknowledge that we cannot definitively determine whether the pandemic caused a decline in gray divorce, but our empirical evidence aligns with both larger pandemic-related trends in overall divorce levels (Manning & Payne, 2021; Manning et al., 2022) as well as ongoing theorizing about the ramifications of the pandemic for the initiation of the divorce process (Lebow, 2020).

The salience of sociodemographic variation in the risk of gray divorce appears to be largely impervious to the pandemic. We did not uncover any significant interactions between survey year and factors such as education, income, or employment (results not shown), suggesting their effects on gray divorce were largely unaltered during the pandemic. Unfortunately, the ACS changed their measure of race-ethnicity between 2019 and 2021, precluding direct comparisons of these measures over time (U.S. Census Bureau, 2021b). This is a notable limitation of our study because non-White adults tended to experience more pronounced pandemic-related stressors and thus also may have faced unique risks of gray divorce. Such differentials did not emerge for adults who were economically disadvantaged.

The pandemic appears to be a pivotal turning point in the trajectory of gray divorce. Prior to the pandemic, the gray divorce trend was distinctive, marked by a doubling between 1990 and 2010 followed by stasis through 2019, a trajectory that belied the dramatic descent in the divorce rate among younger adults during this period (Brown & Lin, 2022). Since the onset of the pandemic, the gray divorce trend has shifted and now aligns with the larger trend evident in the overall U.S. divorce rate, which has steadily declined in recent decades (Cohen, 2019; Kennedy & Ruggles, 2014). Now, divorce is declining across all age groups except those aged 65+, for whom we find that divorce has stagnated since 2019 (after rising sharply for the prior half century). An important task for future research is to assess whether this initial downward trend in gray divorce persists as we move out of the pandemic era.

Funding

This research was supported by a grant from the National Institute on Aging (R15AG047588). Additional support was provided by the Center for Family and Demographic Research, Bowling Green State University, which has core funding from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (P2CHD050959).

Conflict of Interest

None.