Universitas Internasional Batam · Faculty of Business and Management
Candy · Viviona Priscillia · Isnaini Nuzula Agustin
Group Presentation · 6 Presenters
This study examines how two CEO psychological traits — narcissism and overconfidence — jointly influence firm risk-taking and firm performance, and whether board gender diversity moderates those relationships. The study is set within the framework of Upper Echelons Theory (UET) and draws on data from 60 industrial firms listed on the Indonesia Stock Exchange (IDX) during 2019–2023.
Previous studies have largely examined CEO narcissism or overconfidence in isolation. Limited research has explored their combined effects on firm risk-taking and performance, particularly under the influence of gender diversity.
This study contributes by integrating both psychological traits within a single framework and analyzing their interaction with board composition in the Indonesian context, where female executive representation remains limited.
The study extends UET by integrating dual psychological traits (narcissism and overconfidence) as well as an organizational context variable (gender diversity), offering new insight into how CEO personalities interact with governance structures to shape firm behavior.
Practically, the study provides guidance for boards of directors and policymakers to design executive selection processes, monitoring mechanisms, and risk management strategies.
Indonesia provides a valuable yet unique empirical setting. As an emerging market with evolving governance systems, cultural norms, and relatively low female representation in top management, Indonesia faces distinct conditions under which CEO behavior and gender diversity may influence strategic risk-taking and firm performance.
The timeframe captures firm behavior during significant global and domestic economic changes, including the global pandemic and the recovery that followed, which are expected to influence CEO decision-making and firm risk-taking dynamics.
Industrial firms are characterized by higher capital intensity, long-term investment horizons, and greater exposure to operational and strategic risks compared to firms in service or financial sectors. These characteristics make managerial traits, including narcissism and overconfidence, particularly salient in shaping corporate risk-taking and performance outcomes.
Gender diversity is an increasingly vital component of effective corporate governance. In Indonesia, diverse leadership tends to encourage more balanced perspectives and minimize excessive risk-taking behavior. Gender diversity serves as a moderating factor that can mitigate the negative effects of CEO narcissism and overconfidence.
Drawing from Upper Echelons Theory (UET), this study posits that firm outcomes are shaped by and reflective of the personal and cognitive attributes of top management. CEO narcissism and overconfidence, as key psychological traits, influence strategic risk-taking and performance outcomes. Gender diversity, informed by complementary theories, is introduced as a moderating factor that can mitigate excessive risk and enhance governance effectiveness.
Executives' psychological and cognitive attributes shape how they assess and implement solutions to strategic challenges. The primary backbone of the paper.
Women leaders, shaped by social expectations of empathy, collaboration, and caution, are more likely to exhibit risk-averse and ethically grounded decision-making patterns.
Diverse boards with female members introduce alternative perspectives that reduce managerial overconfidence and align decisions with collective long-term interests.
Gender diversity enriches the board's resources and perspectives, stabilizing decision-making and enhancing adaptability. It helps transform CEO narcissism and overconfidence into constructive leadership energy.
Eight hypotheses are tested. The state of each chip below is neutral at this stage. Findings are reported in Section 5.
H1 CEO narcissism affects corporate risk-taking.
H2 CEO narcissism affects firm performance.
H3 CEO overconfidence affects corporate risk-taking.
H4 CEO overconfidence affects firm performance.
H5 Gender diversity moderates the effect of CEO narcissism on firm risk-taking.
H6 Gender diversity moderates the effect of CEO overconfidence on firm risk-taking.
H7 Gender diversity moderates the effect of CEO narcissism on firm performance.
H8 Gender diversity moderates the effect of CEO overconfidence on firm performance.
Research Model — Figure 1
Figure 1 is the hypothesised research model. It shows the paths the study set out to test, not the final results. The moderation paths (H5–H8) are proposals that the study tests in Sections 5–6; they are not established findings.
The actual Figure 1 image is not present in the working file set. This SVG is a faithful structural recreation of the hypothesised model.
Secondary data from annual and financial reports, 2019–2023, sourced from the official IDX website, Yahoo Finance, and company websites.
Purposive sampling. Population: 941 IDX-listed firms. Sample: 60 industrial-sector firms consistently listed 2019–2023 with complete, publicly accessible annual and financial reports.
Panel data design. Ordinary Least Squares (OLS) regression and Moderated Multiple Regression (MMR). All variables standardized before testing. Controls: leverage, CEO gender, firm age.
| Category | Variable | Measurement | Source |
|---|---|---|---|
| Dependent | Firm Risk-taking | σRit = standard deviation of firm returns | Hesniati et al. (2024) |
| Dependent | Firm Performance | Net Income / Total Assets × 100% | Agnihotri & Bhattacharya (2024) |
| Independent | CEO Narcissism | 1–5 scale based on CEO photo prominence in annual reports: no photo; group photo with co-executives; <½ page; >½ page; full page.† | Candy & Delfina (2023) |
| Independent | CEO Overconfidence | Dummy: 1 if firm DER > industry median; 0 otherwise | Candy & Delfina (2023) |
| Moderator | Gender Diversity | Percentage of female directors | Amin et al. (2024) |
| Control | Leverage | Total Liabilities / Total Assets | Agnihotri & Bhattacharya (2024) |
| Control | CEO Gender | Dummy: 1 = female, 0 = male | She et al. (2020) |
| Control | Firm Age | Current year − year of establishment | Agnihotri & Bhattacharya (2024) |
Source: Processed data (2025). Financial data cross-checked against IDX official database. Inter-rater reliability assessed through independent coding by two raters.
† Coding note. The source paper's Table 1 lists the scale beginning at 1 (1 = no CEO photo); its own Table 2 reports a minimum of 0.000 and its descriptive narrative states "0 = no CEO picture." The two are inconsistent within the source. Verify the intended coding with the authors before presenting. This deck reproduces the scale without asserting the contested anchor value.
Variable construction vs. standardization
Table 1 defines how each variable is constructed (operational definition / formula). The paper separately states that all variables were standardized before statistical testing. These are two distinct steps: Table 1 = variable construction; standardization = a later data-preparation step (commonly a z-score, z = (x − mean) / SD); regression = the statistical test on the prepared variables. The paper does not print a standardization formula; do not present one as such.
| Variable | Minimum | Maximum | Mean | Std. Deviation |
|---|---|---|---|---|
| Risk-taking | 0.000 | 7.701 | 0.118 | 0.718 |
| Firm Performance | −12.184 | 1.607 | 0.019 | 0.737 |
| CEO Narcissism | 0.000 | 5.000 | 3.460 | 1.354 |
| CEO Overconfidence | 0.000 | 1.000 | 0.460 | 0.499 |
| Female Director | 0.000 | 0.667 | 0.103 | 0.148 |
| Leverage | 0.000 | 0.744 | 0.248 | 0.190 |
| CEO Gender | 0.000 | 1.000 | 0.050 | 0.211 |
| Firm Age | 5.000 | 153.000 | 32.000 | 22.604 |
Key Observations
Risk-taking mean is low (0.118), indicating a tendency toward small risks, with substantial variation (SD = 0.718). Firm performance mean is close to zero (0.019), indicating stable financial performance across most of the sample.
Governance Composition
Only ~10.34% of board seats are held by women. About 5% of the sample is led by female CEOs. CEO overconfidence: ~46% of the sample falls in the overconfident category (DER above industry median).
| RiskTak. | Perf. | Narciss. | Overconf. | Female Dir. | Leverage | |
|---|---|---|---|---|---|---|
| Risk-taking | 1.000 | |||||
| Firm Performance | 0.468*** | 1.000 | ||||
| CEO Narcissism | 0.115** | 0.146** | 1.000 | |||
| CEO Overconfidence | −0.076 | −0.193*** | 0.010 | 1.000 | ||
| Female Director | −0.059 | 0.121** | 0.017 | −0.207*** | 1.000 | |
| Leverage | −0.164*** | −0.291*** | −0.026 | 0.764*** | −0.202*** | 1.000 |
Full table available in source paper. * p<0.10, ** p<0.05, *** p<0.01.
Correlation is not causation
Table 3 shows bivariate associations only. It does not test H1–H8. The formal hypothesis tests are the regression and interaction models in Sections 5–6. Do not describe the correlation table as confirming or rejecting any hypothesis.
| Group | Obs | Mean | SD |
|---|---|---|---|
| Low overconfidence (0) | 161 | 0.0953 | 0.5438 |
| High overconfidence (1) | 139 | 0.1438 | 0.8798 |
| Combined | 300 | 0.1178 | 0.7184 |
Mean difference = −0.0485, t = −0.5825. Not statistically significant. This is a preliminary independent-samples difference test, not the formal hypothesis test. Note the distinction: the CEO Overconfidence variable in Table 1 is defined by the DER/industry-median rule; the t-test below groups observations around the sample mean (0.46). These are two different operations. Full hypothesis testing follows in Section 5 (direct effects) and Section 6 (moderation).
Notable Correlations
• CEO narcissism is positively related to both risk-taking (r = 0.115**) and firm performance (r = 0.146**).
• CEO overconfidence is negatively related to firm performance (r = −0.193***).
• Leverage is strongly positively correlated with CEO overconfidence (r = 0.764***), indicating overconfident CEOs tend to rely more heavily on debt financing.
| Variable | (1) | (2) | (3) | (4) | (5) |
|---|---|---|---|---|---|
| CEO Narcissism | 0.00402* (1.99) | 0.00405* (2.00) | 0.00376 (1.88) | 0.00378 (1.88) | 0.00382 (1.91) |
| CEO Overconfidence | −0.00737 (−1.34) | 0.0107 (1.27) | 0.0112 (1.30) | 0.00966 (1.12) | |
| Leverage | −0.0621** (−2.80) | −0.0635** (−2.79) | −0.0617** (−2.72) | ||
| Firm Age | −0.000231 (−1.94) | ||||
| Constant | 0.0252*** (3.35) | 0.0286*** (3.61) | 0.0366*** (4.39) | 0.0368*** (4.39) | 0.0448*** (4.82) |
| N | 300 | 300 | 300 | 300 | 300 |
H1 CEO Narcissism → Firm Risk-Taking
Supported. Coefficient positive and significant at the 10% level. CEOs with strong narcissistic traits tend to pursue bold strategies and ambitious decisions that increase the firm's exposure to risk.
H3 CEO Overconfidence → Firm Risk-Taking
Not supported. CEO overconfidence does not appear to influence risk-taking, suggesting that confidence alone is insufficient to drive risky strategic actions.
| Variable | (1) | (2) | (3) | (4) | (5) |
|---|---|---|---|---|---|
| CEO Narcissism | 0.0101* (2.55) | 0.0102** (2.63) | 0.00942* (2.49) | 0.00918* (2.44) | 0.00914* (2.42) |
| CEO Overconfidence | −0.0362*** (−3.45) | 0.0114 (0.72) | 0.00434 (0.27) | 0.00581 (0.36) | |
| Leverage | −0.164*** (−3.92) | −0.146*** (−3.41) | −0.147*** (−3.45) | ||
| CEO Gender | 0.0507* (2.05) | 0.0471 (1.88) | |||
| N | 300 | 300 | 300 | 300 | 300 |
H2 CEO Narcissism → Firm Performance
Supported. CEO narcissism exerts a significant positive influence on firm performance across all models (t ≈ 2.42–2.63).
H4 CEO Overconfidence → Firm Performance
The paper reports a significant negative coefficient in Model (2): −0.0362***, significant at the 1% level. After controls are added (Models 3–5), the coefficient becomes positive and statistically insignificant. The paper's discussion and conclusion treat H4 as supported with a negative effect. Note: the abstract's phrasing ("enhances performance") is the outlier — the regression tables and discussion sections are internally consistent.
| Variable | (1) | (2) |
|---|---|---|
| CEO Narcissism | 0.00597** | |
| Female Director | 0.0395 | −0.0361 |
| Narcissism × Female Dir. | −0.0197 | |
| CEO Overconfidence | −0.0119* | |
| Overconfidence × Female Dir. | 0.0346 | |
| N | 300 | 300 |
H5 Gender Diversity moderates Narcissism → Risk-Taking
Not supported. Interaction term = −0.0197, not significant.
H6 Gender Diversity moderates Overconfidence → Risk-Taking
Not supported. Interaction term = 0.0346, not significant.
| Variable | (1) | (2) |
|---|---|---|
| CEO Narcissism | 0.0133*** | |
| Female Director | 0.181** | 0.0208 |
| Narcissism × Female Dir. | −0.0357 | |
| CEO Overconfidence | −0.0419*** | |
| Overconfidence × Female Dir. | 0.103 | |
| N | 300 | 300 |
H7 Gender Diversity moderates Narcissism → Firm Performance
Not supported. Interaction term = −0.0357, not significant.
H8 Gender Diversity moderates Overconfidence → Firm Performance
Not supported. Interaction term = 0.103, not significant.
Null Moderation Result — an important finding
The presence of female directors on the board neither enhances nor diminishes the impact of CEO narcissism or CEO overconfidence on firm risk-taking or firm performance. All four interaction terms (H5–H8) are not statistically significant.
In Indonesia, the proportion of women on corporate boards remains relatively low, and their influence may be limited by entrenched hierarchical and cultural norms that constrain active participation. Female board representation may be more symbolic than substantive in influencing executive behavior. In addition, many Indonesian firms have concentrated ownership structures, where decision-making authority is heavily centralized in the CEO or controlling shareholders, reducing the practical influence of the board.
This contrasts with studies from developed markets (Sbai et al., 2024; Teodósio et al., 2021), which found that gender-diverse boards effectively constrain risk-taking behavior. The divergence can be attributed to differences in institutional maturity, governance enforcement, and the actual role of women in corporate decision-making.
| Variable | (1) Leverage | (2) ROE |
|---|---|---|
| CEO Narcissism | −0.00393 | −0.0000451 |
| CEO Overconfidence | 0.294*** | −0.0243 |
| Female Director | −0.0226 | −0.168 |
| CEO Gender | −0.124*** | 0.301** |
| Firm Age | 0.000236 | 0.00177* |
| N | 300 | 300 |
Risk-taking is replaced with leverage; firm performance is assessed using ROE. This is an alternative-measure robustness check. The paper does not provide a detailed coefficient-by-coefficient interpretation of Table 9; do not claim it proves every main finding is unchanged.
| Hypothesis | Finding | State |
|---|---|---|
| H1 Narcissism → Risk-taking | Positive, significant | Supported |
| H2 Narcissism → Performance | Positive, significant | Supported |
| H3 Overconfidence → Risk-taking | Not significant | Not supported |
| H4 Overconfidence → Performance | Negative, significant in Model 2 only | Supported (negative) |
| H5 Gender mod. Narcissism → Risk | Not significant | Not supported |
| H6 Gender mod. Overconf. → Risk | Not significant | Not supported |
| H7 Gender mod. Narcissism → Perf. | Not significant | Not supported |
| H8 Gender mod. Overconf. → Perf. | Not significant | Not supported |
CEO narcissism positively affects both firm risk-taking and firm performance. CEO overconfidence exhibits a negative relationship with firm performance. Gender diversity does not significantly moderate the relationship between CEO traits, risk-taking, and firm performance.
These findings align with Upper Echelons Theory, emphasizing that CEOs' psychological attributes play a crucial role in shaping strategic decisions and corporate outcomes.
The study reinforces UET's core proposition while extending it to emerging market contexts characterized by hierarchical governance and collectivist norms. It contributes to behavioral finance by emphasizing how cultural and institutional features shape the manifestation of cognitive biases among executives.
In collectivist or high power-distance cultures, CEO traits may exert stronger, less moderated effects, representing a boundary condition for traditional UET assumptions.
• Firms should integrate psychological profiling into leadership selection and governance design to promote balanced risk-taking.
• HR and nomination committees should integrate psychological assessments to identify potential overconfidence or narcissistic tendencies that may affect corporate strategy.
• Gender diversity initiatives should move beyond numerical representation to ensure substantive participation and oversight power.
• Regulators and policymakers may consider revising governance codes to enhance the structural authority and monitoring capacity of female directors, particularly in high-risk industries.
The observation period is relatively short (2019–2023), relying on annual and financial reports from this period. The sample is restricted to the industrial sector on the IDX; generalizability to other sectors and markets is not established.
• Employ a longer observation period to capture long-term effects of CEO traits.
• Broaden the scope of references and data sources to strengthen validity.
• Examine how gender diversity moderates these relationships across different organizational and cultural contexts.
• Broaden the scope across industries and cultural contexts.