Perceptions of the Future and Pregnancy Avoidance in the U.S

Abstract

Despite low U.S. fertility rates since the Great Recession, two-child norms remain pervasive, suggesting individuals are unable to achieve their goals. To understand what may be driving the apparent mismatch between goals and behavior, we focus on pregnancy avoidance, as individuals may be deciding against births in the short term rather than deciding not to have any, or any more, children. Further, we incorporate subjective evaluations of the future related to economic and relational factors as well as objective socioeconomic indicators, drawing from the Narratives of the Future framework and Easterlin’s theory about expected standard of living. We use data from the 2018–2020 wave of the Toledo Adolescent Relationships Study (N = 880), a population-based dataset, to examine short-term pregnancy avoidance among adults aged 29–36. We find that higher levels of personal economic pessimism and concerns about having a good relationship in the future are associated with greater importance of avoiding a pregnancy in the short term, even when controlling for objective characteristics such as economic hardship, relationship status, and other sociodemographic covariates. The results highlight the need to incorporate both subjective and objective statuses in research on fertility decision-making, and the implications of these findings point to short-term pregnancy avoidance and fertility postponement as a potential mechanism underlying contemporary low birth rates in the U.S.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11113-025-09962-2.

Introduction

Declining fertility in the U.S. and other high-income countries since the Great Recession is well-documented but poorly understood. Although fertility typically falls during economic shocks, it also typically rebounds when the economy improves (Sobotka et al., 2011). Yet fertility did not rise after the Great Recession ended in 2009 and instead continued to fall. The U.S. Total Fertility Rate (TFR) has reached lows not previously recorded, surpassing the previous low of 1.74 in 1973 (Livingston, 2019) to reach 1.62 in 2023 (Hamilton et al., 2024). This puts the U.S. on par with European countries, which have long exhibited low fertility rates (Beaujouan, 2023; Kohler et al., 2002; Morgan & Taylor, 2006). As birth rates remain low—with worries about continued decline—efforts to understand how young adults of childbearing age make fertility decisions are increasingly important.

There is considerable research linking objective circumstances at both the macro- and micro-level to fertility. Aggregate economic conditions, such as unemployment rates, housing costs, and labor market structure (Comolli, 2017; Florida et al., 2021; Seltzer, 2019), and individuals’ own socioeconomic and demographic circumstances, such as employment status, educational attainment, and marital status (Alderotti et al., 2021; Guzzo, 2022; Nitsche & Hayford, 2020), affect both overall fertility levels and individual fertility behavior. Yet these factors cannot completely account for ongoing decline (Comolli et al., 2021; Kearney et al., 2022). An emerging line of both theoretical and empirical work suggests that subjective assessments about one’s current or future circumstances may help understand how individuals make childbearing decisions (Lappegård et al., 2022; Vignoli et al., 2020). And, indeed, prior work has found that subjective assessments related to specific economic micro- or macro-level conditions are important to consider (e.g., Alderotti et al., 2021; Brauner-Otto & Geist, 2018; Hofmann et al., 2017). We extend these arguments by incorporating additional elements: a generalized feeling of economic uncertainty and uncertainty about future relationships. Further, we consider whether individuals are trying to avoid becoming pregnant at a given time. Pregnancy avoidance is likely to be more sensitive to perceptions of people’s current and future circumstances than general goals (i.e., whether they would like to ever have children or how many they hope to eventually have), which have typically been the focus of other work. In particular, we suggest that contemporary low fertility could be driven primarily by pregnancy avoidance, with individuals continually postponing the entrance into parenthood or having another child until some nebulous point in the future when their life circumstances align with their schemas about childbearing and childrearing (Bachrach & Morgan, 2013; Johnson-Hanks et al., 2011).

In this paper, we use population-based data collected prior to the onset of the COVID-19 pandemic, the Toledo Adolescent Relationships Study (TARS), to consider the factors linked to pregnancy avoidance among a cohort of adults aged 29–36. We draw on the Narratives of the Future framework and Easterlin’s (1968, 1973) theory of expected standard of living to consider the role of perceived economic and relationship uncertainty, accounting for standard objective sociodemographic characteristics and statuses as well. The findings support arguments that perceptions of the future in different domains among young adults—those in the prime childbearing years—are associated with individuals’ assessment of the importance of avoiding a pregnancy at a given time. This work has implications for understanding current low fertility rates in the U.S. and other high-income countries because fertility delayed—and delayed again and again—often becomes fertility foregone (Quesnel-Vallée & Morgan, 2003).

Background

There is a long and robust literature in demography on fertility goals, an umbrella term for future orientations toward childbearing that includes specific plans/expectations/intentions as well as general norms/desires/preferences. Overall, fertility goals are strong but imperfect predictors of individual-level fertility (Morgan & Rackin, 2010); they seem to be somewhat better predictors at the aggregate level, though gaps remain (Beaujouan & Berghammer, 2019). There are several reasons for the lack of concordance at the individual level. One, responses to questions about fertility goals may more strongly reflect generalized social norms and ideals than personal preferences; for instance, the two-child ideal remains pervasive in Europe and the U.S. (Brenan, 2023; Sobotka & Beaujouan, 2014), and U.S. teens’ attitudes towards parenthood and having children have been remarkably stable since the late 1970s (Allendorf et al., 2023). Two, longitudinal surveys in the U.S. with information on fertility goals and subsequent fertility behavior (primarily the National Longitudinal Survey of Youth 1979 Cohort (NLSY79) but also the National Study of Adolescent to Adult Health (Add Health)), tend to collect information about goals that are nonspecific in terms of the time horizon. The NLSY79, for instance, asked about desired family size at baseline and in subsequent years asked how many children respondents expect to have (in addition to their current parity). Note that such a measure lacks a time dimension, yet we know from work in non-U.S. contexts that short-term fertility goals are, in fact, quite strongly related to short-term fertility behaviors—and much more strongly related to subsequent fertility than measures referencing longer timeframes or that reflect lifetime goals (Dommermuth et al., 2015; Yeatman et al., 2020).

This brings us to a third reason: fertility goals are dynamic over the life course, influenced by micro-, meso-, and macro-level factors. Childlessness expectations tend to unfold over time (Gemmill, 2019; Rybińska & Morgan, 2019), and intended parity declines modestly as individuals age (Berrington, 2021). The cognitive-social model and the Theory of Conjunctural Action argue that people have fertility schemas (Bachrach & Morgan, 2013; Johnson-Hanks et al., 2011)—a package of ideas about the conditions in which they would like to have children. In the absence of these conditions, people may still be desirous of (additional) childbearing but have no plans to enact those desires. A mismatch between individuals’ current circumstances and their fertility schema may explain why so many individuals are uncertain about their fertility goals at a given point in time (Ní Bhrolchain & Beaujouan, 2019) but also why uncertainty seems to be a transient state (Jones, 2017). As people’s life course conditions change (such as getting married) or the conditions in which childrearing would occur change (such as an economic downturn), they may alter their short-term plans in terms of timing and spacing rather than changing their ultimate fertility goals (Beaujouan & Berghammer, 2019).

The pattern in contemporary societies seems to be one of frequent fertility postponement, and so pregnancy avoidance at a given time could reflect a pattern of ongoing postponement. Birth rates in the U.S. are at a record low, and fertility is increasingly delayed, with women’s average age at first birth at an all-time high of 27.4 years old (Osterman et al. 2024). Yet Americans continue to report an ideal family size of two children, with a substantial minority reporting that three or more is ideal, and only five percent reporting that zero or one child is ideal (Brenan, 2023); similarly, individual fertility intentions remain around two children (Guzzo & Hayford, 2023). Taken together, this suggests that understanding why people are choosing not to have a child at a given point is likely more informative than examining whether people want children or how many they intend to have. The larger story about fertility behaviors in high-income countries seems to be that people are not rejecting parenthood, or rejecting having another child, so much as they are saying “not now.” Thus, we argue that focusing on pregnancy avoidance—how important it is for someone to not become pregnancy right now—is especially informative about contemporary fertility decision-making in an era of heightened social and economic change. These short-term fertility decisions are highly sensitive to fluctuations in individual circumstances and perceptions (Barber et al., 2019; Marteleto et al., 2023). A focus on pregnancy avoidance also harkens back to early fertility theories and models that note that individuals do not formulate—or stick to—set fertility goals decided upon early in the life course. Instead, fertility behavior is driven by a series of sequential short-term fertility decisions (Udry, 1983). In low-fertility settings, most people are repeatedly making a series of decisions to avoid a pregnancy at any given point, and, of course, a series of postponements or repeated decisions to avoid a(nother) birth can translate into foregone fertility (Quesnel-Vallée & Morgan, 2003). Put differently, people are not deciding that they do not want children (or another child) at all so much as they may find themselves repeatedly evaluating their current circumstances as “not the right time” and hoping that better circumstances will materialize in the future that they would consider more appropriate for having a child.

Subjective Perceptions and Pregnancy Avoidance

New theoretical developments coming out of Europe, such as the Narratives of the Future framework (Lappegård et al., 2022; Vignoli et al., 2020), provide important insights as to how a broader set of indicators about the future orientation may influence pregnancy avoidance. There is a growing body of work that finds that subjective notions of uncertainty about the future are linked to fertility goals, with most of this work focusing on economic uncertainty. Some studies have considered how individuals’ experiences in settings with precarious macro-level employment conditions or with potential instability in employment, unemployment, or temporary positions, are related to lower fertility intentions (Fahlén & Ohál, 2018; Mills & Blossfeld, 2013). Experimental vignette approaches that try to capture how potential shifts in macro-level economic contexts might influence childbearing decision-making have found that when individuals are presented with positive future economic conditions, they report higher fertility intentions (Lappegard et al., 2022; Vignoli et al., 2022). Another body of literature has examined individuals’ perceptions of uncertainty about economic aspects of their own lives (Brauner-Otto & Geist, 2018; Gatta et al., 2022; Guetto et al., 2022; Manning et al., 2022; Vignoli et al., 2022). Uncertainty about one’s job or employment status is a commonly used measure in this literature (Buh, 2023). It is often, but not always, linked to lower fertility intentions, and the strength of this association varies across context, though not necessarily across time (van Wijk & Billari, 2024). Related concepts, such as housing insecurity, are also relevant (Vignoli et al., 2013). These individual-level indicators largely seem to tap into what could be called “status uncertainty”—individuals’ perceptions about their jobs and their housing situation in the short term—but not tap into their general orientation about their financial futures.

Although we agree with the tenets of the Narratives of the Future Framework, we extend it by focusing on pregnancy avoidance, which should be more strongly related to individuals’ current circumstances than general fertility intentions. Further, we argue that empirical applications of the Narratives Framework have been perhaps too narrow. The focus on perceived status uncertainty—often driven by data availability—ignores the potential role of more global assessments of future economic well-being. More specifically, we suggest that another important aspect of economic uncertainty is how confident (or worried) people feel about meeting their own economic goals and standards of living. The influence of expected standard of living on fertility dates back to Easterlin (1968, 1973) and Butz and Ward (1979). At its most basic, the idea is that during childhood and drawing from their experiences in their parents’ household(s), individuals form the basis for their expected standard of living and material aspirations as adults. If they fail to reach those expectations and aspirations as adults, they are unlikely to have children, or to have as many children, as they would have otherwise; a key determinant in Easterlin’s theory of the likelihood of reaching expectations and aspirations was relative cohort size. Expected standard of living and aspirations were usually operationalized in objective terms using various aggregate measures of relative economic status across the parent and adult child generations (i.e., unemployment ratios, wage or income ratios, etc.) or comparisons of individual economic characteristics to aggregate characteristics (i.e., individual men’s wage relative to median men’s wage). Although there was usually support for a linkage between relative income and fertility, research was less clear about the link between material aspirations and fertility, in part because of variation in its measurement (Macunovich, 1998). Although Easterlin’s work has largely fallen out of favor as an influential paradigm in fertility research, we draw from this work to incorporate the notion that individuals’ perceptions of achieving their future economic goals are important for fertility decision-making. Young adults likely have ideas about what constitutes “enough” money or a “good” job—and an assessment about whether they are likely to reach those benchmarks; avoiding a pregnancy when one feels they are not currently well-positioned for their definition of success aligns with rational choice models of fertility. Individuals may not be changing their overall childbearing plans, but they are likely to believe it is important to not get pregnant right now given their circumstances. As both economic goals and assessments are almost certainly subjective, perhaps the best way to capture those is individuals’ own responses to direct queries.

Another area that has often been overlooked in both the theoretical and empirical work on uncertainty is relationship uncertainty, despite the persistent link between unions and fertility (Hayford, 2013). Relationship status is strongly associated with fertility goals—women who are partnered have higher odds of intending children than their single counterparts both in the U.S. and Europe (Guzzo, 2022; Sturm et al., 2023). Although about 40% of births in the U.S. are outside of marriage (Osterman et al., 2024), Americans are more approving of childrearing within marriage than outside of marriage (Parker & Minkin, 2023), suggesting the continued presence of norms promoting childbearing and childrearing within stable partnerships. Yet marriage rates have generally declined over time (Loo, 2023), and a declining share of adults have ever married (Payne, 2021). Moreover, the centrality of marriage seems to be waning; more than half of Americans view marriage and committed relationships as important but not essential to leading a fulfilling life (Barroso, 2020), and there is growing uncertainty among young people about whether they will marry in the future (Pepin & Cohen, 2024). Thus, to the extent that people continue to view childbearing as tied to relationship formation and stability, uncertainty about being able to form a stable relationship in the future is likely a salient factor in people’s evaluation of the importance of avoiding a pregnancy.

Current Investigation

We advance understanding of short-term fertility decision-making by focusing on a period following the Recession and prior to the COVID-19 pandemic to assess how subjective views about the future are related to pregnancy avoidance. We operationalize uncertainty in two ways: personal economic pessimism and stress about future relationship circumstances. We expect that individuals who have more negative evaluations of their economic prospects and more concerns about their future relationship circumstances will be more motivated to avoid a pregnancy in the short term. We expect these will be important net of objective economic, relationship, and psychosocial factors. We include covariates that have traditionally been associated with fertility intentions: sociodemographic measures (age, parity, race/ethnicity, education), economic measures (homeownership, income, and economic hardship), relationship status (single, dating, cohabiting, married), and depressive symptomology (Guzzo, 2022; Lazzari & Beaujouan, 2025; Musick et al., 2009).

Data and Methods

We draw on population-based data collected prior to the onset of the COVID-19 pandemic. The Toledo Adolescent Relationships Study (TARS) is a longitudinal data set that consists of a stratified random sample of adolescents based on the 2000 enrollment records of 7th, 9th, and 11th grade students in Lucas County, Ohio (N = 1,321). Although sampling was based on school rosters, school attendance was not a requirement for inclusion in the sample. The TARS data set includes oversamples of Black and Hispanic populations. The first wave of data was collected in 2001, and respondents were reinterviewed eight times. We rely on Wave 6 data because it is the first time questions about pregnancy avoidance, as well as the key independent variables, were included; subsequent waves did not include the focal future orientation measures. The data collection was conducted between April 2018 and March 2020, and the full sample included 990 respondents (mean age = 32). The analytic sample included only those respondents who reported their fertility expectations (n = 921). We further excluded respondents who did not provide data on their race/ethnicity (n = 3), parity (n = 21), education (n = 2), or depressive symptoms (n = 2). Individuals with missing data on our key measures (discussed below) related to personal economic pessimism (n = 3), stress about future relationship (n = 9), or economic hardship (n = 1) were also omitted from the current analysis. Thus, the final analytic sample consisted of 880 respondents.

Measures

The dependent variable is a measure of pregnancy avoidance in the short term. Women were asked the following question: “Thinking about your life right now, how important is it for you to avoid becoming pregnant?” Men were asked: “Thinking about your life right now, how important is it for you to avoid getting someone pregnant?” The responses ranged from (1) “not at all important” to (5) “very important.”

There are two focal measures of uncertainty about the future. First, to measure personal economic pessimism, respondents were asked “How concerned are you about…?” for the following three items: (1) “Not achieving the standard of living you desire,” (2) “Having a dead-end job,” and (3) “Not having enough money.” All responses were measured on a five-point scale from (1) “not at all concerned” to (5) “very concerned” and summed together (α = 0.84). Second, respondents’ stress about future relationship circumstances was measured with two separate questions based on whether respondents were currently in a romantic relationship. Those who were single were asked how stressed they were about not having a good relationship in the future, measured on a five-point scale ranging from (1) “not at all concerned” to (5) “very concerned,” and those who were in a relationship were asked “I feel uncertain about our prospects to make this relationship work for a lifetime” with responses ranging from (1) “strongly disagree” to (5) “strongly agree.” We combined these into a single measure of stress about future relationship circumstances. For both personal economic pessimism and stress about future relationship circumstances higher scores indicate more concern, and these measures were not strongly correlated with each other ( = 0.33).

We control for a set of individual-level covariates found to be related to fertility goals in prior work. These include sociodemographic and well-being measures as well as economic status indicators. Sociodemographic measures include age, gender, race/ethnicity, relationship status, and parity. Age was measured in years as a continuous variable. Gender was coded as a dichotomous variable for women, with men as the reference category. Respondents’ race/ethnicity was coded into four categories: non-Hispanic White (reference category), non-Hispanic Black, Hispanic, and “other.” Relationship status at the time of the survey has four categories: single, dating, cohabiting, and married (reference category). Parity is a categorical measure: no children, one child (reference category), and two or more children. Education is categorized as: high school or less (reference category), some college, and a college degree or more. We account for mental health by including an indicator of depressive symptoms. This was measured using an eight item version of the CES-D scale (Radloff, 1977). Respondents were asked how often each statement was true during the past week: (1) “You felt you just couldn’t get going,” (2) “You felt that you could not shake off the blues,” (3) “You had trouble keeping your mind on what you were doing,” (4) “You felt lonely,” (5) “You felt sad,” (6) “You had trouble getting to sleep or staying asleep,” (7) “You felt that everything was an effort,” (7) “You felt depressed.” Responses ranged from (1) “never” to (8) “several times a day” (α = 0.92). Homeownership was coded as a binary variable (homeowner = 1, not a homeowner = 0). Income was measured by asking respondents how much money they made from working in a typical week and how much money their partners made. The combined income variable was top-coded at the 95th percentile and logged in models to reduce the skewness of the distribution. Finally, economic hardship was created based on responses to six items. Respondents were asked whether, in the past two years, there was a time when they or someone in their household (1) “didn’t pay the full amount of rent or mortgage because there wasn’t enough money,” (2) “were evicted from the house or apartment for not paying the rent or mortgage,” (3) “ran out of money to buy food to make a meal,” (4) “needed to see a doctor or go to the hospital but didn’t because they didn’t have enough money,” (5) “were unable to pay the full gas, electric, or other utility bill because there wasn’t enough money,” (6) “were unable to make the minimum payment on their credit card because there wasn’t enough money.” Based on exploratory analyses, we created a three-category measure of economic hardship as no hardships (reference category), one hardship, and two or more hardships.

Analytic Strategy

We begin by presenting the sample characteristics. To examine the association between how important it was to avoid pregnancy in the short term and our measures of future orientation, we use ordinal logistic regression models and report log likelihood values to provide tests of model fit. Based on Wald tests, we find that the parallel regression assumption is not violated. Each measure of uncertainty about the future is entered into separate unconditional models (Models 1 and 2) to show the bivariate associations. Model 3 includes both measures of future orientation together, and Model 4 shows a full model with future orientation measures and the covariates. As sensitivity tests, we checked for interactions of the focal independent variables and gender, parenthood status, and objective economic measures. We also explored unconditional models by parity and considered alternative individual-level economic characteristics. We mention these results briefly in the text and present the associated tables in the Online Supplement.

Results

Descriptive Results

Table 1 presents the sample descriptives. Nearly half (48%) of respondents reported that it was pretty or very important to avoid a pregnancy at the time of the survey, and about a third (36%) of respondents considered avoiding pregnancy “not too” or “not at all” important. The mean score for personal economic pessimism (mean = 2.42), measured on a scale from 1 to 5, approached the midpoint of the scale. Stress about future relationship circumstances had a mean score of 2.19, also measured on a scale of 1 to 5, suggesting that respondents were not too concerned about having a good relationship in the future. Thus, respondents seemed to feel more optimistic than pessimistic about their economic futures and reported, on average, feeling low stress levels about having their future relationship circumstances.

Table 1.

Descriptive statistics

Source Toledo Adolescent Relationships Study (TARS), Wave 6

	Mean (SD) or %
Avoiding Pregnancy
Not at all important	23.98%
Not too important	12.50%
Somewhat important	16.14%
Pretty important	14.89%
Very important	32.50%
Future Orientation
Personal economic pessimism scale (1–5)	2.42 (1.08)
Stress about future relationship circumstances (1–5)	2.19 (1.27)
Sociodemographic & Well-Being
Age (29–36)	32.45 (1.78)
Gender
Men a	43.30%
Women	56.70%
Race/ethnicity
non-Hispanic White a	65.00%
non-Hispanic Black	21.14%
Hispanic	11.70%
Other	2.16%
Relationship status
Single	21.70%
Dating	11.36%
Cohabiting	21.25%
Married a	45.68%
Parity
0 children	33.30%
1 child a	21.02%
2 or more children	45.68%
Education
High school or less a	22.93%
Some college	31.59%
College or more	46.48%
Depressive symptoms (1–8)	2.22 (1.49)
Objective Economic Factors
Homeownership
No a	57.39%
Yes	42.61%
Weekly income b	1,528.90 (1,362.33)
Economic hardship
0 hardships a	59.09%
1 hardship	13.07%
2 or more hardships	27.84%

N = 880; ^a Serves as the reference group in multivariate analyses. ^b Logged value included in multivariate analyses

In terms of the other sample characteristics, the mean age of the respondents was 32 years old, and 57% were women. Sixty-five percent were non-Hispanic White, 21% were non-Hispanic Black, 12% were Hispanic, and 2% identified as some other racial/ethnic group. Twenty-two percent of respondents were single, 11% were dating, 22% were cohabiting, and 45% were married. About two-thirds (67%) of respondents already had at least one child. About one-fifth of the sample (22%) reported completing high school or less, 32% of respondents had some college education, and 46% indicated having a college degree. Depressive symptoms were measured on a scale from 1 to 8, and the mean score among respondents was 2.21. Regarding objective economic factors, 42% of respondents were homeowners, and the mean weekly household income was about $1,529. Finally, 59% of respondents reported no economic hardship, 13% reported one and 28% reported two or more hardships.

Multivariable Results

Table 2 show coefficients from multivariable ordinal logistic regression predicting pregnancy avoidance. As a reminder, a positive coefficient means respondents reported it is more important to avoid a pregnancy at the time of the survey, whereas a negative coefficient indicates it is less important to avoid a pregnancy. Looking first at unconditional associations and starting with personal economic pessimism in Model 1, we see that as personal economic pessimism increases, respondents report greater importance of avoiding a pregnancy (b = 0.266). We observed a similar association with each separate indicator of the personal economic pessimism scale (not shown). Similarly, greater stress about future relationship circumstances (Model 2) is associated with increased importance of avoiding a pregnancy (b = 0.221). When the subjective measures are in the same model (Model 3), both economic pessimism and relationship uncertainty remain significantly and positively associated with pregnancy avoidance.

Table 2.

Results from ordinal logistic multivariable models predicting importance of avoiding pregnancy

Source Toledo Adolescent Relationships Study (TARS), Wave 6

	Model 1		Model 2		Model 3		Model 4
	Bivariate		Bivariate		Subjective		Full Model
Future Orientation
Personal economic pessimism	0.266***				0.204***		0.174*
Stress about future relationship circumstances			0.221***		0.163***		0.107†
Sociodemographics & Well-Being
Age							0.001
Woman							-0.012
Race/ethnicity
(Non-Hispanic White)
Non-Hispanic Black							−0.120
Hispanic							−0.448*
Other							−0.249
Relationship status
Single							0.844***
Dating							0.707***
Cohabiting							0.568**
(Married)
Parity
0 children							−0.152
(1 child)
2 or more children							0.724***
Education
(High school or less)
Some college							0.055
College or more							0.211
Depressive symptoms							0.066
Objective Economic Factors
Homeownership							−0.278†
Logged weekly income							−0.001
Economic hardship
(0 hardships)
1 hardship							0.152
2 or more hardships							-0.256
Model Fit
χ²	22.48***		20.97***		32.62***		93.72***
Log likelihood	−1348.81		−1349.56		−1343.73		−1313.18

N = 880; the model cutpoints are omitted; † p <.10, * p <.05, ** p <.01, *** p <.001

The full model, Model 4, includes both subjective measures about the future and the full set of covariates including the objective economic indicators. Both personal economic pessimism and future relationship uncertainty remain key predictors of the importance of avoiding a pregnancy, with those who are more pessimistic about their economic circumstances in the future (p = 0.014) or more concerned about not having a good relationship (p = 0.055) reporting greater importance of avoiding a pregnancy. In this model, we see that only one objective economic measure approaches significance: homeownership is marginally and negatively associated with pregnancy avoidance (p = 0.071). Similarly, objective relationship status is independently linked to pregnancy avoidance even when controlling for stress about future relationship circumstances; being in a less serious relationship is positively associated with the importance of pregnancy avoidance relative to marriage. In terms of the other covariates, respondents with two or more children report greater importance of pregnancy avoidance than those with only one child. Compared to non-Hispanic White respondents, Hispanic respondents report less importance of pregnancy avoidance. Thus, it appears that these subjective factors are independent predictors of short-term pregnancy avoidance even with the inclusion of the sociodemographic measures, depressive symptoms, and objective economic measures.

In a model with all covariates except the two subjective measures of personal economic pessimism and stress about future relationship circumstances (see Online Supplement Table S1), homeownership was more strongly and significantly linked to lower importance of avoiding a pregnancy than in Model 4 of Table 2. Depressive symptoms were also significantly and positively linked to pregnancy avoidance when not accounting for subjective factors. Based on model fit, the subjective indicators improved the explanatory power of the model when comparing Model 4 to this supplementary model without personal economic pessimism and relationship uncertainty (p = 0.007). As such, the results do indeed support our argument that uncertainty about the future is linked to short-term fertility goals net of objective statuses and characteristics. To demonstrate the relative magnitude of the subjective perceptions of the future, Online Supplement Table S2 shows the average marginal effects of each level of importance of avoiding a pregnancy for the two indicators of future orientation and shows relatively small effects. For each unit change in the personal economic pessimism indicator, respondents are three percentage points less likely to report it is not all important to avoid a pregnancy. Each unit change in stress about future relationship circumstances reduces the likelihood that respondents indicate that avoiding pregnancy is not at all important by two percentage points.

Sensitivity Analyses

We tested whether the association of the future orientation measures and pregnancy avoidance differed by key sociodemographic and economic factors by specifying interaction terms. First, given some evidence that the influences on fertility decision-making vary by gender (Novelli et al., 2021), we tested whether the objective economic indicators (homeowner, income, and hardship) and future orientation measures (personal economic pessimism and relationship uncertainty) operated in the same way for both men and women. These models indicated there were no statistically significant differences in the associations between any of these indicators for men and women (see Online Supplement Table S3). Next, we considered variation by parenthood. Full models that included interactions between a dichotomous indicator of parenthood status and each of our two future orientation measures and our three objective economic indicators did not find evidence that any of the measures were differentially associated with pregnancy avoidance for those with and without children (see Online Supplement Table S4). Next, we ran models separately by parity to see if the link between pregnancy avoidance and our two subjective measures varied across those with zero, one, or two or more children; note that these are unconditional models due to small parity-specific sample sizes. These unconditional models suggest that economic pessimism and relationship uncertainty were not differentially associated with pregnancy avoidance across parity, though some associations did not reach statistical significance at conventional levels likely due to the smaller sample size (see Online Supplement Table S5).

We also further explored whether the link between personal economic pessimism and fertility outcomes varied by objective economic measures. We found that for personal economic pessimism, the association with pregnancy avoidance varied by income; specifically, economic pessimism was more strongly associated with avoidance of pregnancy for those with lower incomes than higher incomes (see Online Supplement Table S6). No other interactions between economic pessimism and objective economic indicators were statistically significant.

Finally, in another set of sensitivity tests, we considered two sets of alternative indicators of objective economic circumstances: specific types of debt (see Online Supplement Table S7) and debt-to-income ratio and poverty status (see Online Supplement S8). For specific types of debt, the main association between our measures of future orientation and pregnancy avoidance did not change when including any type of debt (car, credit card, payday loan/cash, court, or educational loan debt) in the models instead of homeownership, logged income, or economic hardship. Of the specific types of debt, only credit card debt was associated with pregnancy avoidance; it lowered pregnancy avoidance, perhaps because credit card debt is associated with income in a curvilinear pattern, with the share of households with credit card debt highest in the middle of the income distribution and lowest at the tails of the distribution (Federal Reserve Bank of St. Louis, 2024). Looking at debt-to-income and poverty status, neither of these measures were significantly associated with pregnancy avoidance, nor did using these measures instead of the original economic indicators alter the association between the future orientation measures and pregnancy avoidance. Neither of the models with alternative specifications of objective economic conditions improved model fit.

Discussion and Conclusion

As fertility rates have continued to fall in the post-Recession years in the U.S. and elsewhere, defying earlier patterns of fertility recovery during periods of economic recovery, researchers have explored potential explanations for declines. Most work has focused on either aggregate- or individual-level economic indicators. Although this work has indeed found that objective indicators are important for fertility levels and behaviors, a full understanding of the decision-making underlying fertility has remained elusive. Further, most work has considered general fertility goals—whether people want children ever and how many they plan to have—yet contemporary fertility patterns may be largely driven not by changes in such goals but by short-term decisions to postpone having a child at a given moment. The Narratives of the Future framework (Vignoli et al., 2020) provides a helpful approach to understand the more subjective aspects that may be linked to fertility decision-making, though applications to date have largely focused on European contexts. It explicitly incorporates a future orientation in which individuals draw from their past experiences and interpret their current circumstances to make evaluations of what they think is likely in the future. In its attention to uncertainty about the future, it reorients away from current conditions to anticipated conditions, which are inherently subjective. It aligns well with fertility decision-making, which is inherently about the future given that childrearing is a long-term, and irreversible, endeavor.

We take the Narratives of the Future framework and incorporate concepts from longstanding fertility theories and focus on pregnancy avoidance, advancing the literature in two ways. First, most work following the uncertainty framework has focused on what might be considered “status uncertainty”—how individuals evaluate the stability of their current employment or housing characteristics. Far less attention has been paid to relational uncertainty, although Americans continue to support the importance of raising children in a stable relationship (Parker & Minkin, 2023), and short-term pregnancy intentions fluctuate considerably based on changes in relationship factors (Barber et al., 2019). Moreover, drawing from Easterlin’s work (1968, 1973), which posits that expected standard of living and material aspirations—and whether one has achieved them—is a key predictor of fertility, we focused on how individuals perceive their economic and relational futures in a general sense. Will they have “enough” money and “good” jobs? Will they have “good” relationships that endure?

Second, although most work in this area considers fertility goals more broadly—looking at desires or intentions for the future, often without a time horizon—we argue that how individuals feel at a given moment is probably most strongly related to their views on having a (nother) baby at that moment. Pregnancy avoidance taps into the mechanism that seems to be the driver of contemporary fertility decline: postponement. Ideal family size in the U.S. remains high and stable, as do the number of children Americans intend to have (Brenan, 2023; Guzzo & Hayford, 2023). Yet fertility rates are declining, suggesting low fertility rates may be driven largely by a series of short-term decisions to not have a child. As such, understanding why people are specifically avoiding pregnancy at a specific moment in time might provide insights that studies of generalized fertility goals cannot provide. Considering how individuals perceive their economic and relational situations are likely key drivers of whether people feel they should avoid getting pregnant right now.

Among our sample of young adults in the prime childbearing years, just under half reported it was pretty or very important to avoid becoming pregnancy, though this is undoubtedly related to the fact that roughly the same share of respondents already had at least two or more children. Respondents were, in general, more optimistic than pessimistic about their economic futures. Nonetheless, respondents with greater economic pessimism—achieving their desired standard of living, having a “good” job, and making “enough” money—found it more important to avoid a pregnancy in the short term. This was true even when we accounted for more objective economic indicators, such as economic hardship and income, and was robust to several sensitivity tests that explored other ways of accounting for objective economic conditions. The emphasis on specific aspects of one’s view of their economic circumstances supports Easterlin’s earlier contention that how individuals assess their own likelihood of meeting their economic aspirations is a key driver of fertility behavior. Further, our expectations that relationship evaluations would be important were supported as well. We find that individuals who expressed greater concern about having a good relationship in the future reported that it was more important to avoid getting pregnant in the short term. This was true even when controlling for respondents’ actual relationship status, which is typically the only relationship measure included in studies of fertility decision-making.

Our findings suggest that the Narratives of the Future framework is not only applicable to the U.S. but can be expanded as well. For instance, the notion of uncertainty should move beyond specific economic and financial concerns linked to income, employment and housing to tap into people’s feelings of generalized economic angst and uncertainty (like those underlying Easterlin’s arguments). Further, it can be expanded to incorporate non-economic domains, such as romantic partnerships, which have emerged in these analyses as important. Additional domains may include other types of relationships, such as parents and friends, that could be linked to social support for parents (Bernardi & Klӓrner, 2014), or aspects of health (Lazzari & Beaujouan, 2025). Although the U.S. has a weaker social safety net than many peer nations, there is little reason to believe that the findings here—that broad subjective evaluations of one’s economic and intimate relationship futures—would not be relevant elsewhere, nor is there any reason to suspect that concerns over potentially other aspects of one’s life, such as health or social support, would not be important. Childbearing is inherently future-oriented; having and raising children is a long-term and irreversible endeavor, and so how people evaluate their current and future lives across a broad set of domains will undoubtedly influence how they formulate fertility goals. This is likely to be true across settings, even those that offer more support for parents.

Limitations

Although we provide new insights into the role of subjective economic and relationship indicators, there are a few limitations. First, the analyses are based on cross-sectional data so we can only identify associations and cannot make causal inferences. Relatedly, we only capture future orientation at one point in time; such evaluations of the future are likely fluid and dynamic. Third, TARS is not nationally representative. However, comparisons of the Wave 6 sample with an analogous age group (30–34 year olds) in the American Community Survey demonstrated that the TARS sample closely matches the national population (e.g., in the TARS sample, 38% of respondents were racial/ethnic minorities, compared to 35% of the U.S. population; and 36% of respondents were college graduates, compared to 40% of the U.S. population). Still, there is some attrition based on gender and race/ethnicity that may bias the results. Finally, because TARS does not include information on sterilization or reproductive health (either for the respondent or, if they are in a relationship, their partner), we are unable to limit our analyses to those who remain able to have children, nor can we examine how fertility goals and pregnancy avoidance may be linked to sterilization decisions. It is possible that some people do not feel a strong desire to avoid a pregnancy because they are no longer at risk given sterilization or other fecundity issues.

Conclusion

The low fertility rate of the U.S. emerged later than in many of its peer nations, and, perhaps unsurprisingly, there is comparatively little theoretical or empirical research in the U.S. that addresses why individuals are not having children. Drawing from one of the few U.S. data sources that include both objective and subjective measures of economic and relational indicators as well as fertility goal indicators, our work points to the importance of perceptual factors for fertility decision-making as well as the importance of considering individuals’ views on pregnancy avoidance at a given point in time, rather than more abstract fertility goals for one’s overall life course plans. An important next step will be to assess how these key domains operate together by examining the indirect and direct pathways through which these subjective assessments work together to influence fertility decision-making. We are not arguing that objective factors are irrelevant but, instead, that incorporating how individuals perceive their circumstances provides additional insight into fertility decision-making. We join many researchers’ calls to move beyond the “economic gaze” (Comolli et al., 2021) and consider subjective indicators across a wide range of life course domains to understand shifts in fertility. Further, given relative stability in general fertility goals about having children and the number of children, it may be more informative to focus on short-term pregnancy avoidance. Contemporary low fertility seems to be driven not so much by individuals deciding not to have any (more) children as it is by individuals deciding not to have a child now.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Data availability

The data used in this study, Wave 6 of the Toledo Adolescent Relationships Study, is available through the Inter-University Consortium for Political and Social Research (ICPSR), study 38016.