Based on a sample of Chinese listed-A share firms from 1998 to 2001, this study examines associations between income-increasing BLEM (earnings management through below-the-line items), audit opinions and auditor locations. I hypothesize that firms with higher earnings management through BL (below-the-line) items are more likely to receive MAOs (modified audit opinions) and that the association between earnings management and audit opinions will be stronger for firms audited by non-local auditors. It is also expected that firms audited by non-local auditors tend to have less earnings management. Empirical evidences support all the three hypotheses. I use two models to test the three hypotheses and use two methods to measure BLEM: direct method and industry-adjusted method. The sample includes 3364 observations. The results of main test show that there is a significant and positive association between BLEM and the frequency of MAO. And this association is strengthened when firms are audited by non-local auditors. I also find that the clients of non-local auditors report less BL earnings. All these findings are consistent with my three hypotheses. I go further to analyze the three contents of BL earnings: investment income, non-operating net income and subsidy. The results show that all the three hypotheses are applicable for the former two items. The main results of three sensitivity tests support my hypotheses and provide some new findings. Firstly, the significant association between modified opinions and earnings management is only applicable for unqualified opinions with explanatory notes, but not for other modified opinions. Secondly, the association between earnings management and audit opinion and the impact of auditor location on such association disappear in 2001. It is partly attributed to two regulations issued in 2001.