We investigate the relationship between auditor conservatism and auditor changes. In specific,we explore whether the mandatory or voluntary auditor changes affect auditor conservatism during the sample period of 2006-2008. Auditor selection mechanism has been switched from the auditor assignment rule to the market competition rule in 1982, in order to improve the quality of financial information. However,the market competition rule is criticized for impairing auditor independence due to increased pressure on auditors who want a long-term audit engagement. In 2006, the Korean regulatory body introduces the mandatory auditor rotation rule which firms have to switch their incumbent auditors to new auditors if they are audited by the same auditors during the last 6 years. The rule intends to protect auditor independence. It is believed that auditors can maintain auditor independence because the mandatory auditor rotation rule provides the lower quasi-rent for new auditors (DeAngelo 1981; Geiger and Raghunandan 2002). Prior literature shows that the voluntary auditor changes under the market competition rule may affect the new auditors' audit decisions. In particular, the newly hired auditors prefer more conservative accounting treatments in order to reduce audit failure risks stemming from information asymmetry between the auditors and the new clients (Shin et al. 2007). Since the mandatory audit rotation rule allows the client firms freely choose the new auditors, the mandatorily changed auditors may face the identical audit risks faced by the voluntarily changed auditors. That is, regardless of the mandatory or voluntary changes, the newly hired auditors may ask more conservative accounting treatments in order to alleviate the audit risks from their lack of experiences with and knowledge of the new clients. On the contrary, the mandatorily changed auditors' audit risks may be different from the voluntarily changed audits' risks. While Dhaliwal et al. (1993) and Jenkins and Velury (2008)show that some firms have an incentive to voluntarily change their incumbent auditors to increase the reported earnings (i.e., earnings management), the firms that have to change their incumbent auditors due to the mandatory auditor rotation rule may not have an earnings management incentive. In other words, voluntarily changed auditors are likely to be more conservative for new client rather than mandatorily changed auditors, because auditors with voluntarily changes may face high audit risk that may cause the negative effect on the investors'decision (Hackenbrack and Hogan 2002). This argument leads to our hypothesis that the level of conservatism is lower when the firms change the incumbent auditors mandatorily than when they change the auditors voluntarily. Our model first uses the Penman and Zhang's (2002) conservative measure. The results show that the C score and the Q score of the mandatorily changed auditors are lower than the voluntarily changed auditors: The coefficient for the dummy variable of the mandatory auditor change (MAN) is significantly negative at least 1% level in Penman and Zhang's (2002) measure. We further use the Beaver and Ryan(2000), Basu's (1997) and Ball and Shivakumar's (2005)models for the conservatism measure. The results confirm that the new mandatorily changed auditors are less conservative than the voluntarily changed auditors. These results support the argument that the mandatorily changed auditors evaluate their audit risks lower than the voluntarily changed auditors' risks because the mandatory auditor changes are not occurred for the earnings management purpose and/or the newly hired auditors can research potential clients' business situations in advance. Additionally, we examine the Penman and Zhang's (2002) conservative measure of the mandatorily and voluntarily changed auditors between KOSPI-listed firms and KOSDAQ-listed firms. We find that the new mandatorily changed auditors are less conservative than the voluntarily changed auditors only for KOSDAQ-listed firms. We also compare the Penman and Zhang's (2002) conservatism measures of the mandatorily changed auditors and the retained auditors. It shows that there is no significant difference in the conservatism measures between two groups. This result ascertains the argument that the mandatory auditor rotation does not have any impacts compared to the auditor retention case (Rho 2009). This paper contributes to the literature of the mandatory auditor rotation by showing the significant differences in conservatism between mandatory and voluntary auditor changes. It also provides a policy implication on the auditors' contract termination rights, which are adopted in 2009, by showing the conservatism differences between the auditor changed by the external factors (i.e., the mandatory auditor rotation rule) and by the managers' voluntary decision makings.