Traumatic Experience and Self-Control in Old Age

1. Introduction

The predictive validity of self-control for adult life outcomes has been demonstrated in a variety of domains. For instance, pre-school children who could wait longer for their preferred rewards in a controlled experimental setting developed into more academically and socially competent adolescents, achieving higher SAT scores and coping maturely with frustration and stress (Mischel et al., 1988; Shoda et al., 1990). In follow-up research, this self-control test proved to have significant predictive validity for educational and health outcomes over the life span (Moffitt et al., 2011). Economists have modeled self-control as a parameter of “willpower” in the intertemporal choice model (Thaler and Shefrin, 1981) and have demonstrated its association with savings behavior, occupational achievements, and financial well-being (Daly et al., 2015; Kaur et al., 2015; Laibson, 1998; O’Donoghue and Rabin, 2000; Strömbäck et al., 2017).

Numerous studies have investigated why some individuals are better at self-control than others. This research stream has revealed significant individual heterogeneity across various demographic and socioeconomic characteristics (Calero et al., 2007; Cobb-Clark et al., 2019; Silverman, 2003; Matthews et al., 2009). Much of this heterogeneity has been attributed to differences in parental and family environments that could determine the socialization process during childhood (Gottfredson and Hirschi, 1990). However, this claim has been challenged by the finding that measures of parenting have little explanatory power for self-control in empirical analyses (Beaver, 2009). Some studies have found that parenting is no longer a significant determinant of self-control once the model accounts for genetic factors and prenatal conditions (Beaver et al., 2009; Wright and Beaver, 2005).

It is likely that other salient factors that could determine the learning of self-control have not been identified by the literature. This study focuses on the lasting negative influence of traumatic experiences on self-control. The human brain is known to encode external stimuli as threatening or non-threatening and to adjust its behavior to the perceived threat level (McEwen, 2007). When experiencing a potentially threatening event, it enters a state of “mental combat mode” and redirects energy and resources away from advanced reasoning and toward a lower level of basic bodily processes necessary for survival (Arnsten et al., 2015). Sustained exposure to threats and the related hormonal changes have lasting effects on brain architecture; these lead to an under-activation of the brain’s modulatory functioning and thus undermine its ability to resist temptation (National Scientific Council on the Developing Child, 2005).

Most studies on traumatic experience and impulsive behavior have examined adolescents or young adults, focusing on the role of domestic violence and childhood abuse in the socialization process. As a result, current knowledge is limited to the relatively short-term effects of traumatic events experienced in early childhood (Simmen-Janevska et al., 2015). The question of whether shifts in self-control persist into later life has significant implications for economic behaviors and poverty dynamics, given their salience for predicting later life outcomes. This study examines the association between traumatic life experiences and self-control in older adults using retrospective data drawn from the Health and Retirement Study (HRS). Our results show that the experience of serious physical attacks or assaults is associated with about a 3.1% decline in self-control up to three decades later. The estimated trauma effect is roughly comparable to the effect of being unemployed and three times greater than the gender gradient in self-control, indicating that early life trauma is a major determinant of self-control with lasting consequences.

This study contributes to three strands of research. The first is research on early life experiences and human capital accumulation, which reports the association between adverse conditions (e.g., illness, in-utero alcohol exposure, poor nutrition) during early childhood (the period from in-utero to age five) and health and economic outcomes in later life (Almond and Currie, 2011; Ampaabeng and Tan, 2013; Blackwell et al., 2001; Hayward and Gorman, 2004). Although self-control is generally not considered a form of human capital, it is a valuable psychological resource that enables people to follow through on their commitments to human capital investments and expedites their accumulation (Cadena and Keys, 2015; Griesdorn and Durband, 2016).

Our work also adds to the growing body of literature that estimates the effect of traumatic experiences on preference features, such as risk aversion and time preference. Evidence that preferences shift after exposure to a traumatic event has been reported for survivors of major natural disasters (Ahsan, 2014; Callen, 2015; Cameron and Shah, 2015; Cassar et al., 2017; Eckel et al., 2009; Hanaoka et al., 2018), terrorist attacks (Sacco et al., 2003), and civil conflicts and wars (Bellucci et al., 2020; Brown et al., 2019; Callen et al., 2014; Jakiela and Ozier, 2019; Kim and Lee, 2014; Moya, 2018), exploiting the event as a source of identification for the trauma’s causal effect. A similar finding has also been reported in a study of retrospective data on negative life events and risk preference (Banks et al., 2020; Bucciol and Zarri, 2015; Hetschko and Preuss, 2020). Despite this mounting evidence, however, the direction of the preference shifts remains unclear, with the findings differing according to the study site, source of traumatic events, and preference outcomes examined (Abatayo and Lynham, 2020).

Finally, our study addresses the literature in psychology and human development that explores how self-control is formed and evolves over a lifetime. The conventional theory has been that self-control is established early in life and remains relatively stable afterward (Gottfredson and Hirschi, 1990). Hence, current knowledge is unable to help us determine why some individuals continue to exhibit impulsive behaviors long after exposure to toxic stress (e.g., among veterans home from war) and what this implies for their long-term economic well-being. We depart from this conventional practice by taking a longer, lifetime perspective of self-control development.

The remainder of the paper is organized as follows. Section 2 reviews the research that links traumatic experiences to self-control and explains how this issue has been modeled in studies of economic preferences (e.g., risk aversion, time preference). Section 3 describes our data, measures, and empirical models. Section 4 presents our estimation results. Section 5 concludes with a discussion of the study’s findings and implications, and potential future research directions.

2. Previous research

3. Methods

3.3 Measures of traumatic experiences

Traumatic experiences were assessed using the lifetime trauma inventory developed by Krause et al. (2004). The traumatic events considered in this study include (a) a serious physical attack or assault (“physical attack” hereafter); (b) a major fire, flood, earthquake, or other natural disaster (“natural disaster” hereafter); and (c) a life-threatening illness or accident (“life-threatening illness” hereafter). For each event, respondents were asked to indicate whether it had occurred to them and when it had happened last. Using these responses, we construct binary indicators of occurrence and of whether the event occurred before or after the threshold age of brain maturity. Approximately 7% of our sample reported a physical attack, 17% reported exposure to a natural disaster, and 25% reported suffering from a life-threatening illness. The mean number of years elapsed since the event is 30.3, 27.6, and 19.2 for physical attacks, natural disasters, and life-threatening illnesses, respectively (see Figure 1). The survey questions are provided in Table A1.

View this table:

• View inline
• View popup
• Download powerpoint

Table A1.

Description of variables

• Download figure
• Open in new tab

Figure 1.

Frequency histogram of years elapsed after traumatic events

4. Results

Table 1 shows the average descriptive statistics for the study sample. The sample characteristics are presented for the full sample and for subsamples divided according to trauma exposure (a sample of 3,471 persons with an experience of physical attack, natural disaster, or life-threatening illness and a sample of 5,296 without). We find that the average respondent is 63.2 years old, has an average of 2.66 living children, owns 0.75 private health insurance plans, and holds a net worth of $559,858. Most participants are married (72%) and rate their health condition as good or better (81% good or better; 19% poor or fair). In terms of lifetime trauma, 7% report having experienced a physical attack, 17% report having experienced a natural disaster, and 25% report a life-threatening illness. Comparing the self-control scores by trauma exposure shows that those who experienced a traumatic event have lower self-control than those who have not (p < 0.05).

Table 2 presents the main regression results. We begin with the null model controlling for only trauma indicators and gradually expand the model with additional control variables. Coefficient estimates for the month-of-survey dummies, year-of-survey dummies, and census region fixed effects are omitted for brevity. All standard errors are clustered at a combination of individual and age levels (two-way clustering).

The null model in column 1 shows a negative correlation of self-control with experience of a physical attack (β= − 0.182; p < 0.01) and a life-threatening illness (β= − 0.081; p < 0.01). After we adjust for the baseline covariates and fixed effects (column 2), the coefficient estimates on the two trauma indicators are lower but remain negatively significant at the 5% level. A rough comparison to the sample mean of self-control shows, on average, a 3.7% decline (= − 0.162 / 4.36) and a 0.94% decline (= − 0.041 / 4.36) for the experience of an attack and illness, respectively. Further conditioning on early life conditions (column 3) leads to a coefficient estimate of − 0.137, or a 3.1% decline in self-control in response to an attack (= − 0.137 / 4.36), and an estimate of − 0.037, or a 0.8% decline in self-control due to illness (= − 0.037 / 4.36), which remains statistically significant at the 5% level. The estimate of a 3.1% decline is roughly comparable to the effect of being unemployed and is half the effect of having a college degree or being an ethnic minority. The fact that a memory of almost 30 years ago explains as much as the influence of education and race bolsters our hypothesis that the influence of traumatic experience is long-lasting.

The regression in column 4 additionally controls for measures reflecting the Big Five personality traits (neuroticism, extraversion, openness to experience, agreeableness, and conscientiousness). This specification allows us to test whether the results from the previous specification are driven by a shift in personality or a shift in self-control skills beyond the respondents’ personality. The results show that the estimate on physical attacks remains negatively significant (β= − 0.078; p < 0.01), while the estimate on life-threatening illnesses is no longer different from zero at the 5% level. This result suggests that the influence of physical attacks was not limited to personality shifts but led to a meaningful reduction in one’s ability to resist temptation.

The regression in column 5 is weighted by the propensity score weight from matching the sample by exposure to any traumatic event. Matching on “observables” has the advantage of balancing out differences in the observed characteristics between the treated (with trauma) and untreated (without trauma) groups, thereby reducing the bias resulting from non-random exposure to a traumatic event. We estimate a logistic regression for having been exposed to any traumatic event (one for exposure to any event and zero otherwise) and use the kernel matching method with a bandwidth of 0.05. Similar to the unmatched results, re-weighting regression using the propensity score weight leads to a negative estimate on physical attack (β= − 0.084; p < 0.01).

Next, we examine whether the association between a traumatic event and self-control differs according to the exposure timing (see Table 3). For each traumatic event, we create a dummy variable for an event that occurred at age x or less (early life exposure) and another for an event that occurred after age x (late-life exposure), where x is the critical age of brain maturity. The age at which the brain stops developing and reaches its peak is not clear, with recent evidence showing that the human brain continues to mature into a person’s 20s or 30s (see Somerville, 2016 for a review). The fMRI study by Dosenbach et al. (2010) has shown, for instance, that the human brain reaches its adult volume by the age of 22, while some regions like the frontal lobe (the area of the brain concerned with self-control) continue to grow past this age, until the age of 24 or 32. Currently, there is no agreed-upon threshold for full brain maturity. Therefore, we consider the three ages posited in Dosenbach et al. (2010) as the likely threshold age of brain maturity when constructing the early-and late-life trauma exposure indicators.

The regressions in Table 3 are conditioned on the baseline covariates and fixed effects as specified in eq. (1). Across all columns, exposure to a traumatic event is negatively related to self-control regardless of when the event occurred (p < 0.05 for all estimates of a). The F-tests for the pair of estimates concerning experience at age ≤ x vs. experience at age > x fail to reject the null hypothesis at the 5% level (linear restrictions at the bottom of Table 3), indicating that the correlation between physical attacks and self-control does not differ according to the exposure timing. Overall, there is no evidence that an attack experienced at the critical age for brain development is more meaningful for self-control skills at older ages.

Finally, we evaluate the predictive validity of our self-control measure by examining its relationship with important life outcomes known to be correlated with self-control (see Table 4). Following the evidence that self-control predicts health, health behavior, and financial well-being, we estimate the association between our measure of self-control and body mass index (BMI), obesity (BMI > 30), number of chronic health conditions, depression (Center for Epidemiologic Studies Depression scale), exercise (one for moderate physical activity zero for no physical activity), log of financial net worth, log of total net worth, and life satisfaction (using the seven-point scale of Diener et al. [1985]). Similar to the previous regressions, these models all include the baseline covariates and fixed effects. O.verall, the results confirm the prior finding that self-control is positively related to health outcomes (lower BMI, less obesity, less chronic disease, and fewer depressive symptoms), health behavior (increased participation in physical activity), financial well-being (financial and total net worth), and life satisfaction. All the coefficient estimates on self-control are statistically significant at the 5% level and are large enough to be economically meaningful. Though suggestive, these results show that our self-control measure is a valid indicator of the ability to resist temptation over many domains of behavior.

5. Discussion

The willpower to override impulses and follow through on a plan is important for individual intertemporal choice. While the behavioral implications of self-control have been studied extensively, its determinants are yet to be fully understood. This study argues convincingly that self-control can be influenced by lifetime exposure to traumatic events. Our empirical analyses have considered various potential traumatic events and have examined their associations with self-control at older ages.

Our main results reveal an average 3.1% reduction in self-control in response to a serious physical attack or assault experience. The association between exposure to attack and self-control does not differ according to whether the attack was experienced before or after age 24. The mean of the years elapsed since the attack or assault is about 30, conditional on exposure, meaning that self-control skills never fully recover to their pre-exposure level. For other types of traumatic events (natural disasters and life-threatening illnesses), we have found no consistent evidence that self-control shifts in response to exposure. Our inference regarding the effects of physical attacks and assaults is robust to controlling for early life conditions and using the propensity score re-weighting to balance out underlying heterogeneity by trauma exposure.

The weak correlations of self-control with natural disasters and life-threatening illnesses might best be explained with reference to the strength of the emotional backlash each event triggers. It is well-known that events that are more shocking and salient are more durable and accessible in memory. Among the three events considered in our study, physical attacks and assaults may be more emotionally intense given their criminal nature. Though suggestive, our preliminary findings have revealed a strongly positive correlation between physical attacks or assaults and the number of depressive symptoms. This raises the possibility that attacks and their related emotional changes manifest as post-traumatic stress disorder and disrupt the brain’s modulatory core.

Our study has several limitations. First, the higher mortality in the trauma group raises the concern that our sample may under-represent severely traumatized individuals and over-represent those exposed to relatively low risk (Pak, 2021). Under the plausible assumption that those excluded from the sample are the most impulsive, addressing survivorship bias would lead to a stronger negative correlation between trauma exposure and self-control. Second, it remains unclear through which mechanism traumatic experience produces a lasting influence on self-control. Our estimates could be the result of mental health problems such as post-traumatic stress disorder or its influence through employment outcomes or portfolio reallocation. Future research needs to examine cohort data that could track a lifetime transition into old age and explore the various channels underlying the consequences of trauma. Finally, individuals might differ in terms of their propensity to experience traumatic events. Although our childhood covariates and matching estimation mitigate these effects, we cannot completely rule out the possibility that some aspects of mid-life are still unobserved and confound our estimation process. Finally, we were unable to identify a causal relationship between traumatic experience and self-control. It would be fruitful to exploit exogenous changes in traumatic experiences, such as natural disasters and armed conflicts, and examine their correlation with self-control outcomes at older ages.

Despite these limitations, our findings offer a useful guide for future researchers examining the formation of self-control skills. Specifically, it is important to note that the demographics, socioeconomic characteristics, and trauma indicators used in this study jointly explain less than 40% of an individual’s self-control. The limited explanatory power of the estimated models calls for a more rigorous and systematic investigation into how individual self-control is determined. Researchers in behavioral economics are encouraged to replicate or expand this study using multiple self-control assessments, different sample frames, and various domains of self-control, in both the experimental setting and the survey context.

Among its key implications, this study highlights the need for government intervention to treat traumatic experiences. While mental health programs are available for war veterans, those outside the military are often overlooked and are not provided with targeted intervention. As discussed above, having an ability to resist temptation has broad benefits for individuals, their family, and society they belong to. Our findings imply that a well-designed mental health program that explicitly focuses on traumatized individuals and provides support early in life would produce significant social benefits. Since those at the bottom of the economic scale are more likely to experience traumatic events, the benefits of such policy measures would disproportionately accrue to the least advantaged social groups.