Emotional Instability and Financial Decisions: How Neuroticism Fuels Panic Selling

Abstract

This study investigates the relationship between neuroticism and panic-selling behavior among investors, particularly during market downturns. Building on the theoretical framework of behavioral finance, we hypothesize that higher levels of neuroticism are positively associated with an increased likelihood of panic selling. The data for this research were derived from a comprehensive survey titled Survey on Life and Money, which was conducted by Rakuten Securities in collaboration with Hiroshima University in November and December 2023, with a total sample size of 189,524 participants. Our results reveal that 9.46% of the respondents fully or partially panic-sold their stocks during market volatility. Additionally, the respondents demonstrated a tendency toward neuroticism, with an average score of 2.95 out of 5 on the neuroticism scale. Using a probit regression analysis, we examined the dependent variable of panic selling in relation to neuroticism as the independent variable, controlling for various demographic, socioeconomic, and behavioral characteristics. Our findings robustly support this hypothesis, indicating that individuals with higher neuroticism scores are more likely to engage in panic selling during periods of market volatility, with significance at the 5% level. This study contributes to the behavioral finance literature by highlighting the significant role of personality traits in investment decision making and underscores the importance of understanding investor psychology in financial markets. This study emphasizes the need for a nuanced understanding of how individual psychological factors, particularly neuroticism, drive market behavior and influence broader economic stability.

1. Introduction

Panic selling refers to the impulsive and widespread sale of assets by investors reacting to sudden market declines, often driven by fear of further losses (Elkind et al. 2022; Huynh and Xia 2021; Meet et al. 2023). This behavior not only impacts individual portfolios, but also creates significant market volatility, leading to financial instability across sectors and even influencing broader economic conditions (Elkind et al. 2022; Barberis and Thaler 2013). When investors sell in anticipation of continued downturns, they prioritize immediate protection over the potential for asset recovery, effectively destabilizing the markets on a larger scale. The resulting large-scale selloffs contribute to systemic risk, amplify market disruptions, and affect financial stability during crises, underscoring the importance of strategies to mitigate panic-driven market reactions.

Panic selling is a behavioral phenomenon that has been largely explained through the lens of prospect theory, particularly the concepts of loss aversion and regret aversion (Elkind et al. 2022; Schmidt and Zank 2005; Kahneman and Tversky 1979; Tversky and Kahneman 1981; Wang et al. 2013; Sainsbury 2023; Zhou 2020). When large groups of investors react to negative signals by rapidly divesting assets, cascading effects drive prices downward and amplify market instability. This herding behavior is exacerbated by loss and regret aversion, which leads investors to prioritize immediate emotional relief over long-term strategy (Wang et al. 2013; Kuramoto et al. 2024; Lal et al. 2024). Other behavioral biases, such as overreaction, overconfidence, framing, and impulsivity, further influence tendencies toward panic selling (Meet et al. 2023; Kuramoto et al. 2024; Lal et al. 2024). These dynamics heighten market volatility, undermine investor confidence, and may contribute to prolonged economic slowdowns as risk aversion intensifies across markets. While behavioral theories offer valuable insights into the mechanisms behind panic selling, they do not fully explain why some individuals are more prone to these biases than others. This study addresses a critical gap in the literature by introducing personality traits, particularly neuroticism, as explanatory factors for panic selling. Personality traits have been widely used to explain general investment behavior but have not been applied to panic-selling behavior specifically. Neuroticism is especially relevant because it captures emotional volatility, anxiety, and sensitivity to negative stimuli, all of which align closely with the psychological drivers of panic selling. Neurotic individuals tend to emphasize negative information and are more reactive to perceived threats, making them more susceptible to panic selling during market downturns. Moreover, personality traits also help explain why some individuals exhibit a higher susceptibility to behavioral biases, such as loss and regret aversion. By linking personality traits to panic selling, this study offers a novel perspective on the intersection of personality psychology and behavioral finance, providing deeper insights into the individual-level factors that drive this phenomenon.

The factors that drive investors to engage in panic selling remain unclear despite several behavioral theories that have successfully predicted this behavior (Elkind et al. 2022; Sainsbury 2023; Kuramoto et al. 2024; Lal et al. 2024). During market crises, when panic selling surges, investors often respond emotionally rather than analytically to market changes, as fear forces them to sell stocks impulsively (Elkind et al. 2022; Sauer and Kramer 2022). This phenomenon often drives rational investors to make irrational decisions by triggering emotional biases that overpower their analytical approaches. When markets become turbulent, rational investors are influenced by fear-based signals, such as negative news, rumors, and herd behavior, which cloud judgment and override long-term strategies. Instead of focusing on fundamentals or market trends, they become preoccupied with avoiding losses or regret of not acting before price drops.

This shift from a calculated investment approach to an emotional response closely aligns with the principles of loss and regret aversion. Loss aversion, which suggests that losses inflict roughly twice the emotional pain of equivalent gains, leads investors to avoid losses at any cost, often resulting in impulsive selling rather than patience during market recovery (Schmidt and Zank 2005; Kahneman and Tversky 1979; Tversky and Kahneman 1981). Regret aversion heightens this effect by pressing investors to act preemptively, fearing that they will regret their inaction if prices fall further (Coricelli et al. 2005; Loomes and Sugden 1982). Together, these biases create a powerful force, distorting decision making and leading rational investors to overreact to short-term negative events, abandon their long-term strategies, and join in panic selling alongside other investors. This behavior often starts with a negative event or rumor that undermines investor confidence, spreads anxiety, and prompts some investors to sell off their assets. As these initial sales drive prices further down, others join, creating a cascading effect that amplifies market loss.

Behavioral finance describes panic selling as an irrational overreaction to adverse signals that erode initial optimism and push investors to quickly offload stocks (Sainsbury 2023). Research into behavioral finance documents excessive reactions to market downturns as a common pattern, particularly during crashes (Wang et al. 2013; De Bondt and Thaler 1985, 1987). Common behavioral biases associated with panic selling include loss and regret aversion, herd mentality, shifting sentiments, impulsivity, framing, and altered risk preferences (Sainsbury 2023; Kuramoto et al. 2024; Lal et al. 2024). Among these biases, loss aversion and regret aversion are especially influential, with loss aversion driving people to minimize losses through panic selling (Sainsbury 2023; Odean 1998; Weber and Camerer 1998), while regret aversion prompts investors to sell in anticipation of further declines (Sainsbury 2023; Odean 1998; Weber and Camerer 1998).

Personality traits play a crucial role in shaping individuals’ investment and economic decisions, influencing their risk tolerance, time horizons, and responses to market fluctuations (Donnelly et al. 2012; McCrae and Costa 1992, 1997; Vuković and Pivac 2024). Traits such as openness, conscientiousness, extraversion, agreeableness, and neuroticism can significantly impact financial behaviors and outcomes as they affect how investors perceive risk, manage stress, and pursue or avoid risk-based opportunities (Donnelly et al. 2012; McCrae and Costa 1997; Vuković and Pivac 2024; Durand et al. 2013; Rodrigues and B.V. 2024). For instance, high levels of conscientiousness often correlate with careful planning and long-term investment strategies, whereas high levels of openness can lead to greater flexibility in adapting to market changes (Donnelly et al. 2012).

However, neuroticism is particularly important in this context because it is associated with emotional volatility, anxiety, and sensitivity to negative events, all of which may amplify susceptibility to behavioral bias under market stress (Lin 2011; Hughes et al. 2020). Jiang et al. (2024) and Oehler et al. (2018) found that individuals with high emotional instability were reluctant to invest and were unable to hold stocks for extended periods. Jiang et al. (2024) reported that neurotic tendencies lead to lower estimates of future stock returns and higher estimates of the probability of market crashes. As far as panic selling is concerned, neuroticism is more relevant than the other four personality traits because it is closely tied to emotional reactions, whereas panic selling is often driven by emotional decisions that deviate from rationality. Neurotic individuals tend to experience higher levels of stress and are more reactive to losses or perceived threats, which can lead them to prioritize immediate emotional relief over long-term financial strategy (Barlow et al. 2021; Klein et al. 2011). This increased sensitivity to adverse market movements makes neurotic investors more vulnerable to behavioral phenomena, such as loss aversion, regret aversion, and overreaction, which can lead to impulsive decision making and deviation from rational financial planning.

In market downturns, investors high in neuroticism may exhibit a tendency toward panic selling, as the fear of losses outweighs potential long-term gains. While the influence of personality traits on financial behavior is widely acknowledged, the specific link between neuroticism and panic selling remains unexplored in academic research. Given the established association between neuroticism, risk aversion, and emotional reactivity, understanding their role in panic selling could offer valuable insights into investor behavior during crises and help develop targeted interventions to mitigate irrational market reactions.

Neuroticism, defined as a heightened sensitivity to stress and emotional instability, is theoretically linked to panic selling, as it intensifies reactions to perceived financial losses and uncertainties. As one of the Big Five personality traits, neuroticism is characterized by emotional volatility, anxiety, elevated response to stress, and negative outcomes (McCrae and Costa 1992, 1997). Individuals high in neuroticism are more prone to experiencing intense reactions to perceived threats, risk, uncertainty, or loss, which can profoundly affect their decision-making processes, particularly under financial stress (Hughes et al. 2020). Moreover, neurotic individuals are more likely to be influenced by the reference group because they are less confident in their investment decisions (Lin 2011; Bashir et al. 2013; Jamshidinavid et al. 2012). According to the prospect theory framework, individuals are more affected by potential losses than by equivalent gains, a tendency known as loss aversion (Kahneman and Tversky 1979). When combined with neuroticism, this aversion to loss becomes more pronounced as neurotic individuals are less equipped to tolerate market volatility and are more inclined toward emotionally driven decisions, such as prematurely selling assets to alleviate their anxiety. For example, Durand et al. (2019) found that neurotic individuals tend to have myopic loss aversion. Moreover, Fenton-O’Creevy and Furnham (2020a, 2020b) argued that people with higher neuroticism experience financial distress as they are less capable of planning ahead or choosing better financial products. Additionally, neurotic individuals often struggle with regret aversion (Winter 2019), fearing that failure to act might result in greater losses if markets decline further. These increased responses to loss and regret make neurotic investors more susceptible to panic selling during market downturns as their primary focus shifts from rational investment strategies to immediate emotional relief.

Based on this theoretical background, this study hypothesizes that higher levels of neuroticism are positively associated with an increased probability of panic selling during market downturns. This hypothesis posits that neuroticism, due to its association with increased loss and regret aversion, directly contributes to panic-driven divestment behavior. The primary objective of this study is to investigate the relationship between neuroticism and panic selling behavior among investors during market downturns, and specifically to assess how personality traits, particularly neuroticism, influence decision-making processes in the context of financial stress and emotional responses to market fluctuations. To address this, the study seeks to answer the following research question: “To what extent does neuroticism, as a key personality trait, predict panic-selling behavior among investors during market downturns?” By explicitly linking this research question to the study’s objectives, the analysis highlights the role of neuroticism in driving panic selling, providing valuable insights into the intersection of personality psychology and behavioral finance. This study offers a novel contribution to the existing literature by focusing on the link between neuroticism and panic selling, an area that has been underexplored in behavioral finance. While previous studies examined various personality traits related to investment behavior, the direct connection between neuroticism and the propensity for panic selling during market crises has not been systematically investigated. By leveraging a large and comprehensive dataset, an approach seldom adopted in prior research, this study provides a more robust and nuanced analysis of how neuroticism specifically influences panic-selling behavior. By integrating insights from psychology and behavioral finance, this study aims to fill a significant gap in understanding how individual differences in personality can lead to irrational financial decisions. These findings provide valuable insights for academics and practitioners in finance and psychology, expanding the theoretical framework of behavioral finance by linking personality traits to specific investment behaviors. Additionally, this study offers practical implications for financial advisors and investment firms by identifying the characteristics of investors prone to panic selling, thereby facilitating the development of targeted interventions, educational programs, and risk-management strategies that help mitigate emotional decision making and promote more rational investment behavior, ultimately contributing to greater market stability.

2. Data and Methods

2.1. Data

This study utilizes data from the Survey on Life and Money, conducted as an online survey by Rakuten Securities, Japan’s leading online securities firm, in association with Hiroshima University. The survey was administered in November and December 2023 and is one of the largest financial micro datasets in Japan, covering a broad spectrum of socioeconomic backgrounds to ensure representativeness across the population. The survey sampled active security users aged 18 years and older, encompassing 189,524 respondents, with an effective sample of 112,896 (59.56%) after data processing. The questionnaire addressed various aspects, including socioeconomic status, psychological factors, investment preferences, and cognitive biases. Given the large and representative sample, the data provided a solid foundation for reliable analysis.

2.2. Variables

The dependent variable in this study is panic selling during the COVID-19 market crash in March 2020. We identified panic selling based on the responses to the question, “How did you modify your investment position in March 2020 when the stock market plummeted due to the spread of the new coronavirus?” Respondents had five options: (1) I sold some of my stocks and investment trusts; (2) I sold all of my stocks and investment trusts; (3) I increased the amount of stocks and investment trusts; (4) I purchased new stocks and investment trusts; and (5) I made no changes to my investments. To align with the study’s objective, we created two binary dependent variables to capture panic-selling behavior:

Selling All: This variable equals 1 if the respondent sold all their stocks and investment trusts, and 0 otherwise.

Selling All or Partially: This variable equals 1 if the respondent sold all or some of their stocks and investment trusts, and 0 otherwise.

The primary independent variable was neuroticism, one of the Big Five personality traits associated with increased negative affectivity and psychological distress. Neuroticism was measured using two statements regarding the respondents’ personality characteristics. Respondents rated how well each statement described their personality on a five-point scale, ranging from “strongly agree” to “strongly disagree.” The two neuroticism elements were (1) “Anxious, easily upset” and (2) “Calm, emotionally stable” (reverse-coded). With a Cronbach’s alpha value of 0.89, the measurement of the variable demonstrated a high degree of reliability.

To isolate the impact of neuroticism on panic selling, we included controls for demographic and socioeconomic variables based on prior studies that have shown associations between panic selling and these factors. Specifically, gender, age, marital status, parenthood, employment status, education level, household income, household financial assets, risk aversion, and financial literacy have all been found to influence panic-selling tendencies in previous research (Sainsbury 2023; Kuramoto et al. 2024; Lal et al. 2024). Including these control variables ensures that the observed effects of neuroticism are not confounded by established demographic, socioeconomic, or behavioral influences.

Table 1. Definitions and measurements of variables.

Dependent Variables
Selling All	Binary variable:1 indicates a respondent selling all stocks and mutual funds, 0 otherwise
Selling All or Partially	Binary variable: 1 indicates a respondent selling all or part of their stocks and mutual funds, 0 otherwise
Main Independent Variables
Neuroticism	Continuous variable: average score for number of correct answers from the two personality traits questions related to neuroticism
Control Variables
Age	Continuous variable: age of the respondents
Age Squared	Continuous variable: Square of each age
Male	Binary variable: Equal to 1 if the respondent is male, 0 otherwise
University Degree	Binary variable: Equal to 1 if the respondents hold at least a university degree, 0 otherwise
Unemployed	Binary variable: Equal to 1 if the respondent is unemployed, 0 otherwise
Married	Binary variable: Equal to 1 if the respondent is married, 0 otherwise
Have Children	Binary variable: Equal to 1 if the respondent has at least one child, 0 otherwise
Log of Household Income	Continuous variable: Natural log of the respondent’s own income
Log of Household Asset	Continuous variable: Natural log of the respondent’s household financial assets
Financial Literacy	Continuous variable: average score for number of correct answers from the three financial literacy questions
Risk Aversion	Continuous variable: Risk of rain preference (percentage score from the question, “Usually when you go out, how high must the probability of rainfall be before you take an umbrella?”)

provides detailed definitions and measurements of all dependent and independent variables.

2.3. Descriptive Statistics

Table 2. Descriptive statistics.

Variable	Mean	Std. Dev	Min	Max
Dependent Variables
Selling All	0.0174	0.1309	0	1
Selling All or Partially	0.0772	0.2669	0	1
Main Independent Variables
Neuroticism	2.9488	0.8383	1	5
Control Variables
Age	46.2785	12.3175	18	94
Age Squared	2293.4210	1190.9740	324	8836
Male	0.6543	0.4755	0	1
University Degree	0.6372	0.4808	0	1
Unemployed	0.0746	0.2629	0	1
Married	0.6825	0.4654	0	1
Have Children	0.6021	0.4894	0	1
Household Income	7,229,614	3,594,268	1,000,000	20,000,000
Log of Household Income	15.6470	0.5850	13.8155	16.6487
Household Asset	19,927,628	17,686,348	2,500,000	100,000,000
Log of Household Asset	16.2297	1.0688	14.7318	18.4207
Financial Literacy	0.7879	0.3006	0	1
Risk Aversion	0.5423	0.2312	0	1
Observations	112,896

presents the descriptive statistics for the variables in this analysis based on a sample size of 112,896 observations. These descriptive statistics offer a comprehensive overview of the demographic, socioeconomic, and psychological characteristics of the sample, contributing to a nuanced understanding of the factors potentially associated with panic-selling behavior. Among the respondents, 1.74% sold all their stocks and investment trusts in March 2020 during the sharp stock market decline caused by the coronavirus pandemic, while 7.72% sold either all or part of their stocks and investment trusts. The mean neuroticism score of the respondents was 2.95 (SD = 0.84) on a 1-to-5 scale, with higher scores reflecting higher levels of neuroticism.

The average age of the respondents was 46.28 years (SD = 12.32), and 65.43% were male. The mean household income is 7.23 million yen (SD = 3.59 million yen), and the average household assets amount to 19.93 million yen (SD = 17.69 million yen). Educational attainment was high among the respondents, with 63.72% holding a university degree or higher and 7.72% reporting not being employed currently. Additionally, 68.25% of the respondents were married, and 60.21% had children. Regarding financial and behavioral variables, the mean risk aversion score was 0.54 (SD = 0.23), suggesting a moderate level of risk aversion in the sample. Financial literacy was high, with an average score of 0.79 (SD = 0.30), indicating a strong understanding of financial concepts.

2.4. Methods

This study examines the relationship between panic-selling behavior and neuroticism during a market downturn triggered by the COVID-19 pandemic. We used two binary dependent variables to measure panic selling: Selling All and Selling All or Partially. Both variables are indicators of panic selling based on the respondents’ self-reported actions during a market crash. Since these dependent variables are binary, probit regression models are an appropriate choice for the analysis. The primary independent variable in this study is neuroticism, and several demographic, socioeconomic, and behavioral factors were controlled to account for their potential influence on the relationship between neuroticism and panic selling. We employed two probit regression models to examine the likelihood of panic selling.

Model 1 (Selling All):

P (Selling All = 1) = β0 + β1 Neuroticism + β2 Gender + β3 Age + β4 Marital Status +
β5 university degree + β6 unemployed + β7 children + β8 log of household income +
β9 of household assets+ β10 financial literacy +β11 risk aversion + εi

Model 2 (Selling All or partially):

P (Selling All or partially = 1) = β0 + β1 Neuroticism + β2 Gender + β3 Age + β4 Marital Status +
β5 university degree + β6 unemployed + β7 children + β8 log of household income +
β9 of household assets+ β10 financial literacy +β11 risk aversion + εi

Probit regression is particularly well suited for this analysis, as the dependent variables are binary and probit models provide estimates of the probability that an event occurs based on the values of the independent variables. This approach accounts for the nonlinear relationship between the independent variables and the probability of panic selling, making it more appropriate than linear probability models.

To ensure the reliability of the regression results, we conducted multicollinearity tests using variance inflation factors (VIF) and a correlation matrix. We found that all VIF values were well below the threshold of 10, and the correlation coefficients between the variables were less than 0.70, indicating that multicollinearity was not a concern. For brevity, we do not present the VIF and correlation matrix in this study; however, the absence of multicollinearity supports the validity of the estimated coefficients.

3. Empirical Findings

We conducted a probit regression analysis to examine the relationship between panic selling and neuroticism during the COVID-19 market downturn.

Table 3. Probit regression analysis of panic selling (sold all securities).

	(1)	(2)	(3)	(4)
Variables	Model 1-1	Model 1-2	Model 1-3	Model 1-4
Neuroticism	0.0424 ***	0.0412 ***	0.0365 ***	0.0357 ***
	(0.0119)	(0.0120)	(0.0124)	(0.0124)
Male	0.3795 ***	0.3768 ***	0.4404 ***	0.4371 ***
	(0.0245)	(0.0245)	(0.0255)	(0.0254)
Age	−0.0210 ***	−0.0202 ***	−0.0152 ***	−0.0146 ***
	(0.0054)	(0.0054)	(0.0054)	(0.0055)
Age_squared	0.0002 ***	0.0002 ***	0.0002 ***	0.0002 ***
	(0.0001)	(0.0001)	(0.0001)	(0.0001)
University_degree	−0.1468 ***	−0.1514 ***	−0.1021 ***	−0.1061 ***
	(0.0208)	(0.0209)	(0.0213)	(0.0213)
Unemployed	0.0175	0.0193	0.0371	0.0385
	(0.0435)	(0.0435)	(0.0440)	(0.0440)
Married	0.0016	0.0025	−0.0092	−0.0086
	(0.0280)	(0.0280)	(0.0279)	(0.0280)
Have_children	0.0000	0.0014	−0.0073	−0.0061
	(0.0265)	(0.0266)	(0.0266)	(0.0266)
Log_hincome	−0.0917 ***	−0.0900 ***	−0.0673 ***	−0.0659 ***
	(0.0215)	(0.0214)	(0.0216)	(0.0216)
Log_hasset	−0.0865 ***	−0.0894 ***	−0.0617 ***	−0.0643 ***
	(0.0109)	(0.0110)	(0.0112)	(0.0113)
Risk_aversion		0.1755 ***		0.1419 ***
		(0.0435)		(0.0430)
Financial_literacy			−0.5253 ***	−0.5204 ***
			(0.0308)	(0.0306)
/cut1				−0.2263
				(0.3338)
Constant	0.8625 ***	0.7842 **	0.2892
	(0.3324)	(0.3315)	(0.3341)
Observations	102,561	102,561	102,561	102,561
Log likelihood	−8809	−8800	−8665	−8659
Chi2 statistics	446	452.1	735.2	736.2
p-value	0	0	0	0

Note: Robust standard errors are shown in parentheses. *** p < 0.01, ** p < 0.05.

presents the results of this analysis. Our approach involved incrementally adding sets of variables to assess the models’ robustness and identify consistent predictors across specifications. In Model 1-1, we included only demographic and socioeconomic variables to provide a baseline model for observing the foundational effects of these factors on panic selling. Moving to Model 1-2, we added risk aversion to capture the individual risk preferences that are strongly associated with investment behavior during market stress. In Model 1-3, we incorporated financial literacy, recognizing its role in shaping rational decision making. Finally, in Model 1-4, we included all the control variables (demographic, socioeconomic, and financial) to ensure a comprehensive coverage of factors that might impact the probability of panic selling.

Across all models, neuroticism consistently shows a statistically significant positive association with panic-selling behavior at the 1% level, indicating that higher neuroticism scores correlate with a greater propensity to sell off all investments during periods of market turmoil. This robust finding across the model specifications highlights that neuroticism is a stable predictor of panic sales. These control variables provide important insights. We found that being male and having a higher risk aversion score significantly increased the likelihood of panic selling, while age, holding a college degree, higher household income, and household financial assets were significantly associated with a reduced likelihood of panic selling. Variables such as employment status, marital status, and the presence of children were not statistically significant, suggesting that they do not play a meaningful role in predicting panic-selling behavior in our sample.

For model fit, we assessed the log-likelihood and chi-square statistics, which suggest that each successive model provides a better fit, with Model 1-4 showing the best overall fit owing to its comprehensive variable inclusion. These improvements in model fit further validate the robustness of our findings in all specifications. In summary, the probit regression results suggest a robust positive association between neuroticism and panic selling, with specific demographic and financial characteristics significantly contributing to the likelihood of panic-driven divestment.

To test the robustness of the initial findings presented in Table 3, we conducted an additional probit regression analysis using a more flexible definition of panic selling, which included both the partial and complete liquidation of stock holdings. This adjustment acknowledges that investors may react to market volatility with varying degrees of selling, rather than exclusively divesting all assets. The results of these analyses are shown in

Table 4. Probit regression analysis of panic selling (sold all or partially).

	(5)	(6)	(7)	(8)
Variables	Model 2-1	Model 2-2	Model 2-3	Model 2-4
Neuroticism	0.0511 ***	0.0501 ***	0.0460 ***	0.0452 ***
	(0.0069)	(0.0069)	(0.0070)	(0.0070)
Male	0.3453 ***	0.3432 ***	0.3923 ***	0.3900 ***
	(0.0137)	(0.0137)	(0.0141)	(0.0141)
Age	−0.0176 ***	−0.0169 ***	−0.0124 ***	−0.0118 ***
	(0.0033)	(0.0033)	(0.0033)	(0.0033)
Age_squared	0.0002 ***	0.0002 ***	0.0002 ***	0.0002 ***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)
University_degree	−0.0764 ***	−0.0808 ***	−0.0386 ***	−0.0427 ***
	(0.0126)	(0.0126)	(0.0128)	(0.0128)
Unemployed	0.0404	0.0420 *	0.0540 **	0.0554 **
	(0.0252)	(0.0252)	(0.0254)	(0.0254)
Married	−0.0060	−0.0056	−0.0115	−0.0113
	(0.0163)	(0.0163)	(0.0163)	(0.0163)
Have_children	−0.0001	0.0018	−0.0061	−0.0044
	(0.0153)	(0.0153)	(0.0153)	(0.0153)
Log_hincome	−0.0620 ***	−0.0608 ***	−0.0467 ***	−0.0456 ***
	(0.0121)	(0.0121)	(0.0121)	(0.0121)
Log_hasset	−0.0447 ***	−0.0471 ***	−0.0241 ***	−0.0264 ***
	(0.0063)	(0.0063)	(0.0064)	(0.0064)
Risk_aversion		0.1537 ***		0.1374 ***
		(0.0253)		(0.0252)
Financial_literacy			−0.4598 ***	−0.4568 ***
			(0.0193)	(0.0192)
/cut1				0.2328
				(0.1844)
Constant	0.2049	0.1393	−0.1752
	(0.1845)	(0.1845)	(0.1843)
Observations	106,951	106,951	106,951	106,951
Log likelihood	−28,553	−28,534	−28,273	−28,258
Chi2 statistics	1099	1131	1650	1670
p-value	0	0	0	0

Note: Robust standard errors are shown in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

, where each model incrementally adds controls to better isolate the influence of neuroticism on panic-selling behavior under this broader definition.

Across all the models, neuroticism remained significantly positively associated with selling all or part of one’s investments at the 1% level. This consistency across the models reinforces the finding that individuals with higher neuroticism scores are more likely to engage in panic-driven divestment. These control variables provide additional insights. In line with prior findings on panic-selling behavior, being male and risk aversion were significantly associated with a higher probability of panic selling. Conversely, age, university education, household income, and household financial assets were significantly negatively associated with the outcome, indicating that older, better-educated, and financially stable individuals are less likely to sell investments in response to market downturns. Employment status shows a positive association, with unemployed individuals more likely to engage in panic selling, although this effect is only weakly significant in Models 2-3 and 2-4. Other variables, such as marital status and having children, were not statistically significant, suggesting minimal influence on panic selling in our sample.

Regarding model fit, the log-likelihood values improved progressively from Model 2-1 to Model 2-4, while the chi-square statistics and associated p-values indicated strong overall model significance. These improvements in the model fit lend further support to the reliability of our findings. Overall, the probit regression results in Table 4 underscore the robustness of neuroticism as a predictor of partial or full panic selling, even after controlling for an array of demographic, socioeconomic, and behavioral factors. This finding highlights the importance of personality traits in financial decision making, particularly under stressful market conditions.

4. Discussion

Our findings demonstrate a robust association between neuroticism and panic selling, supporting the hypothesis that higher levels of neuroticism are positively related to the likelihood of full and partial panic selling during market downturns. These results align well with prospect theory (Kahneman and Tversky 1979), which suggests that individuals’ aversion to loss and potential regret can lead to emotionally driven, risk-averse behaviors during periods of market instability (Schmidt and Zank 2005; Coricelli et al. 2005). Neuroticism amplifies these tendencies, as it is associated with increased sensitivity to negative stimuli and reduced tolerance for uncertainty and loss (McCrae and Costa 1997; Vuković and Pivac 2024; Hughes et al. 2020; Bashir et al. 2013). In both regression models, neuroticism remains a significant predictor of panic selling at the 1% level, regardless of whether the behavior involves complete or partial divestment. This consistent significance underscores the role of neuroticism as a driving force of panic selling, even when the definition of panic selling is relaxed to include partial sales. This result aligns with previous findings indicating that emotionally volatile individuals, such as those with high neuroticism, are more likely to react strongly to negative market signals (Hughes et al. 2020; Jiang et al. 2024; Bashir et al. 2013). By responding to stress through immediate divestment, neurotic investors may prioritize short-term emotional relief over long-term investment goals, supporting the premise that neuroticism enhances susceptibility to loss and regret aversion, as prospect theory notes.

In addition to the COVID-19 market crash, it is important to consider how other market conditions or events might influence the observed behaviors. Market downturns caused by factors such as geopolitical tensions, interest rate changes, or economic recessions may elicit similar panic-selling tendencies, particularly among individuals with high neuroticism. Although our study focuses on the COVID-19 crash as a specific context, the increased sensitivity to loss and uncertainty associated with neuroticism is likely relevant for various types of market disruptions. These behaviors may vary in intensity depending on the nature, magnitude, and perceived unpredictability of the event. Furthermore, although the study is based on a Japanese sample, we argue that the findings are applicable across cultural contexts because neurotic individuals, regardless of cultural or economic background, tend to exhibit similar emotional responses (Kajonius and Giolla 2017). While the magnitude of panic selling might vary due to contextual factors, these variations are unlikely to materially affect the core tendencies associated with neuroticism. Future research could explore how different stress-inducing market scenarios interact with personality traits to provide a more comprehensive understanding of panic selling under diverse market conditions. This broader perspective would help contextualize our findings within the larger framework of investor behavior during market stress.

In addition to neuroticism, demographic and socioeconomic factors contribute significantly to the likelihood of panic selling. Consistent with previous research, men were more likely to engage in panic selling than women (Sainsbury 2023; Kuramoto et al. 2024), likely because of gender differences in risk tolerance and confidence levels. Age and age-squared coefficients suggest that younger investors are more prone to panic selling, a finding that aligns with those of previous studies showing age-related differences in risk perception and financial behavior (Kuramoto et al. 2024). Higher education levels, household income, and household financial assets have a significant negative relationship with panic selling. These findings align with theories suggesting that greater financial resources and higher levels of financial literacy are associated with reduced susceptibility to emotional and impulsive selling (Kuramoto et al. 2024; Lal et al. 2024). Individuals with more resources may have both the knowledge and security required to withstand temporary market downturns without succumbing to panic.

Risk aversion was consistently positively associated with panic selling, indicating that individuals who are more risk averse are also more prone to panic during market decline, corroborating the findings of Lal et al. (2024) and Kuramoto et al. (2024). However, financial literacy shows a significantly negative relationship, suggesting that financially literate individuals are better equipped to maintain composure and make rational decisions during crises. This observation is in line with previous findings that financial literacy enhances investors’ resilience to bias (Lal et al. 2024; Baker et al. 2019; Bucher-Koenen and Ziegelmeyer 2014).

These results contribute to the behavioral finance literature by highlighting how personality traits, particularly neuroticism, interact with market events to influence panic selling. While previous studies documented the role of biases, such as herd mentality, framing effects, impulsivity, and overconfidence, in influencing market behavior (Sainsbury 2023; Kuramoto et al. 2024; Lal et al. 2024), our findings provide evidence that individual personality traits significantly moderate these biases, especially under the stress of a market crisis. Neuroticism’s link to panic selling reinforces the theory that personal factors play a critical role in investment decisions, extending beyond behavioral and cognitive influences alone (Donnelly et al. 2012; Rodrigues and B.V. 2024; Jiang et al. 2024).

By directly examining the personality factors of neuroticism, this study fills a gap in the understanding of which individual traits increase susceptibility to emotional and impulsive responses during crises. These insights have practical implications for individuals and financial institutions. Understanding the impact of neuroticism and other individual differences on investment decisions can help financial advisors develop customized interventions and help investors maintain long-term investment strategies, even in volatile markets.

Advisors can help clients recognize emotional responses and employ strategies to mitigate impulsive decisions by educating them about the potential influence of personality traits on financial decisions. Moreover, financial literacy programs that address common biases, such as loss aversion and regret aversion, could further reduce the likelihood of panic selling by equipping investors with tools to better evaluate market downturns.

The insights gained from this study on neuroticism and panic selling highlight several limitations and suggest directions for future research. First, focusing solely on neuroticism may limit our understanding of how other personality traits, such as conscientiousness or openness, influence panic-selling behavior. A multidimensional approach could uncover how these traits interact under stress to shape financial decision making. Future studies could expand on the current findings by exploring how personality traits, such as neuroticism, interact with specific investor types and characteristics, including stock market experience, and investment objectives, and with influencing factors such as financial knowledge or market conditions. Second, the cross-sectional design limits our ability to make causal inferences about the relationship between neuroticism and panic selling. Although our findings support a strong association between neuroticism and the likelihood of panic-driven behavior, causality cannot be established. Longitudinal studies would allow researchers to observe whether changes in market conditions interact with personality over time to influence investment behavior. Third, the study’s findings may be limited in generalizability, as the dataset was derived exclusively from Rakuten Securities’ clientele. This sample may disproportionately represent digitally literate and financially engaged individuals, potentially skewing the results. However, the large sample size of more than 106,000 respondents reduces the potential for significant bias and provides a solid foundation for the analysis. Nonetheless, future studies should aim to validate these findings using more diverse datasets that encompass a broader range of investor demographics and behaviors to ensure wider applicability.

5. Conclusions

This study emphasizes that neuroticism is a key predictor of panic selling, particularly during periods of market instability. Integrating insights from loss and regret aversion under prospect theory (Schmidt and Zank 2005; Kahneman and Tversky 1979; Coricelli et al. 2005), we demonstrated that neurotic individuals are especially prone to panic-selling behavior. Across the models, neuroticism consistently showed a strong association with both full and partial panic-driven asset liquidation, illustrating how personality traits directly impact financial behavior during crises. Additionally, demographic and financial factors, such as gender, age, risk aversion, and financial literacy, also significantly shape panic-selling tendencies, suggesting that investor behavior is influenced by a complex interplay of personality and contextual factors.

Our findings have practical implications for financial advisors and policymakers, who may leverage these insights to support neurotic investors in adopting strategies to resist market-driven emotional impulses. Targeted interventions, such as educational programs on behavioral biases, can help mitigate emotional decision making and reinforce long-term investment perspectives that contribute to greater market stability.

These results indicate promising avenues for future research in this area. Although this study emphasizes neuroticism as a key predictor of panic selling during market instability, it is important to acknowledge certain limitations and outline potential directions for further research. For instance, focusing exclusively on neuroticism may not fully capture the influence of other personality traits, such as conscientiousness or openness, on panic-selling behavior. Adopting a multidimensional approach could provide deeper insights into how these traits interact under stress to shape financial decision making. Moreover, the cross-sectional design of this study limits the ability to draw causal inferences about the relationship between neuroticism and panic selling. Longitudinal research is necessary to explore how these dynamics evolve over time and across varying market conditions, further enriching our understanding of behavioral finance. By expanding studies to include diverse samples and exploring additional personality dimensions, future research can contribute significantly to this growing field. These directions, already discussed in detail earlier, underscore the complexity of investor behavior and provide a foundation for advancing both theoretical and practical applications in financial decision making.