A stock's price crashes when bad news accumulated within the firm is released all at once. I suggest two incentives for bad news hoarding-managerial opportunism and overconfidence-and examine their effects on firm-specific stock price crash risk. In a corporate setting in which there is separation of ownership and control, managers have multiple incentives to conceal bad news, such as extracting rent, securing their jobs, maintaining reputations and meeting the conditions for extra compensation. Overconfident managers also have incentives to ignore negative news. They tend to overestimate the cash flows of investment projects and misperceive negative news as temporary or untrue. They are reluctant to disclose negative news because doing so might prompt investors to halt the projects. Thus, both opportunism and overconfidence can motivate managers to hoard bad news, which in turn can lead to stock price crashes. These two incentives are not, however, mutually exclusive. Opportunism assumes that managers are rational expected-utility maximizers, whereas overconfident managers are expected to behave according to irrational psychological traits. This suggests that overconfident managers can conceal bad news even if they are benevolent to shareholders. The difference allows them to empirically decide which incentive is the main source of the stock price crash. I use tax avoidance (long-run effective tax rate) as a proxy for opportunism because it is widely involved in managerial opportunistic decisions such as earnings management and resource diversion. As a proxy for firm-level overconfidence, I use a dummy variable constructed following Schrand and Zechman (2012). Overconfidence is a persistent trait that affects all business decisions, suggesting that the level of overconfidence can be inferred from the firm's financial and investment decisions. Schrand and Zechman (2012) developed an overconfidence score using a list of financial and investment decisions typically conducted by overconfident managers. Following Chen, Chen, Cheng, and Shevlin (2001) and Hutton, Marcus, and Tehranian (2009), I use two measures of firm-specific stock price crash risk: the likelihood of extreme and negative firm-specific weekly returns and the negative conditional skewness of firm-specific weekly returns. Using non-financial firms listed on the Korea Stock Exchange for the 2001~2011 period, I find that while tax avoidance is positively related to stock price crash risk, overconfidence is not. The results are consistent with the hypothesis that opportunism motivates managers to hoard bad news, which leads to stock price crashes. The finding is robust when I control for firm-fixed effects and use alternative tax avoidance measures, such as the book-tax income difference and the long-term effective tax rate adjusted for earnings management. I also find that the relation between tax avoidance and crash risk is more pronounced in firms with greater information asymmetry. In contrast, information asymmetry does not help explain the relation between overconfidence and crash risk. I further examine the effect of investor protection on the relation between tax avoidance and crash risk. Investor protection variables include largest shareholder ownership, foreign ownership, proportion of outside directors, amount of external financing, big audit firm dummy, chaebol affiliation dummy and the disparity between control and cash flow rights. It appears that the effect of tax avoidance on crash risk is less pronounced for firms with greater investor protection. Specifically, firms with higher largest shareholder ownership and greater external financing, firms audited by big audit companies, non-chaebol firms and firms with lower disparity between control and cash flow rights all exhibit relatively weak relations between tax avoidance and crash risk. The results reinforce that opportunism is a more promising source of stock price crashes than overconfidence. Earlier studies have tried to find the determinants of stock price crashes, focusing on market mechanisms. Recently, related research expanded its focus to include firm-side mechanisms, but most studies have explained stock price crashes using the agency framework. This study complements those works by suggesting a managerial behavioral trait as a source of stock price crashes.