Three essays on CEO behavior and corporate performance: evidence from IPO survival, organizational capital, and executive pay

Wei, Xiaoqin (2026) Three essays on CEO behavior and corporate performance: evidence from IPO survival, organizational capital, and executive pay. PhD thesis, University of Glasgow.

Abstract

The thesis consists of three independent empirical chapters that examine the effects of executive characteristics on corporate outcomes, corporate governance, and managerial compensations. Specifically, the first chapter empirically examines the impact of the use of optimistic tone in IPO prospectuses on IPO survival. The results suggest that net positive tone has strong positive predictive power for IPO survival. The second chapter investigates how CEO turnover affects the accumulation of organizational capital and the results suggest that CEO turnover increase the accumulation of organizational capital. The third chapter¹ analyzes the impact of managerial overconfidence on pay-for-luck compensation. The introduction chapter provides an overview and main findings of the thesis.

The first empirical chapter examines how the use of optimistic language in IPO prospectuses (S-1 filings) influences IPO survival. First, to address the issue of negative phrases being wrapped in positive words, I use FinBERT to detect positive and negative sentences and construct a measure of relative optimism. Second, I use Loughran and McDonald (2016)’s financial dictionary to build an alternative measure of managerial optimism based on the relative frequency of positive word counts. The findings show that the use of optimistic language serves as a positive signal for IPO survival in both FinBERTbased and Loughran and McDonald (ibid.)-based measures. This optimistic tone retains strong predictive power even after controlling for other determinants identified in prior studies. Robustness checks confirm that optimistic tone consistently predicts IPO survival across various survival models. The results remain robust across several additional specifications, including entropy balancing and the internal instrumental variable (IV) approach of Lewbel (2012). Specifically, optimistic tone is associated with a 50% to 60% lower IPO failure hazard. Furthermore, the analysis shows that underwriter quality and venture capital (VC) backing strengthen this relationship, amplifying the positive effect of optimistic tone on IPO survival.

This study empirically examines the determinants of organizational capital, with a particular focus on the impact of CEO turnover. I find that the appointment of a new CEO significantly stimulates the accumulation of organizational capital, increasing its stock by approximately 15%. Furthermore, a one standard deviation increase in managerial ability corresponds to an estimated 7%-8% increase in the standard deviation of organizational capital. Our findings further show that the positive association between CEO turnover and organizational capital is stronger when successor CEOs have higher managerial ability and stronger long-horizon incentive alignment. These patterns are consistent with the view that CEO quality and incentive design shape post-turnover organizational capital accumulation. However, because the chapter does not directly observe internal organizational changes such as hiring, training, management-system adoption, or process redesign, these results are interpreted as heterogeneity evidence rather than definitive proof of specific mechanisms. Furthermore, the post-turnover increase in organizational capital is more pronounced when successor CEOs have higher measured managerial ability. By contrast, the estimated post-turnover response is weaker when successor CEOs score lower on ability-related measures. These results describe average differences across turnover events and should not be interpreted as direct evidence of individual CEOs’ intentions or strategic motives. Our results remain robust across subsamples defined by different turnover reasons, including voluntary, involuntary, and forced turnover. These estimates are consistent with a causal interpretation under standard difference-in-differences assumptions, but the chapter does not claim that the DiD designs fully eliminate all threats to identification. Additionally, I observe that the CEO compensation structure including performance based pay and the CEO pay gap-is associated with a stronger post-turnover response to the positive effect of CEO ability on organizational capital. Finally, the presence of outside CEOs and highly capable CEOs further strengthens the positive effects of CEO turnover on organizational capital.

The third chapter examines how CEO overconfidence amplifies the “pay-for-luck” phenomenon in executive compensation. Using a decomposition of firm performance into exogenous “luck” and firm-specific “skill” components, we find that overconfident CEOs receive disproportionately higher rewards for positive market shocks while avoiding equivalent penalties for negative shocks. To address endogeneity concerns, we instrument CEO overconfidence using the industry-level density of overconfident CEOs and Lewbel (2012)’s internal IV approach. Our results remain robust across alternative overconfidence measures, empirical specifications, and governance conditions. Further analysis suggests that overconfident CEOs engage in greater risk-taking behaviors and higher R&D investments which reinforce the effects of CEO overconfidence on pay-for-luck. Additionally, we find that stronger corporate governance and DoDD-Frank Act mitigates the extent of overconfident CEOs’ pay-for-luck. These findings contribute to the literature on executive compensation and behavioral corporate finance, offering implications for incentive design and governance reforms.

1. This chapter is accepted by International Review of Financial Analysis (Wei 2025)

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Subjects: H Social Sciences > HG Finance
• Colleges/Schools: College of Social Sciences > Adam Smith Business School
• Supervisor's Name: Tsoukas, Professor Serafeim and Gucbilmez, Dr. Ufuk
• Date of Award: 2026
• Depositing User: Theses Team
• Unique ID: glathesis:2026-85897
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 30 Apr 2026 13:33
• Last Modified: 01 May 2026 08:29
• Thesis DOI: 10.5525/gla.thesis.85897
• URI: https://theses.gla.ac.uk/id/eprint/85897