Note(s): Panel A presents the percentage of low accounting quality observations among high and low volatility firms. Low accounting quality is the accounting quality below the median level measured by Accounting Quality1 and Accounting Quality2. High (low) volatility is the volatility above (below) the median level measured by the average of standardized ROA Volatility and standardized Return Volatility

Panel B presents the logit regression results of the sample data partitioned into two subsamples depending on whether volatility level is above or below the sample median. Columns 1 and 3 present regression results of the observations with the volatility level below the sample median. Columns 2 and 4 present regression results of the observations with the volatility level above the sample median. The volatility level is a composite measure, which is the average of standardized ROA Volatility and standardized Return Volatility. In Columns 1 and 2, the independent variable Accounting Quality1 is a composite measure, which is the average of standardized Big4, standardized Earning Predict CF, and the opposite of standardized Abnormal Accruals. In Columns 3 and 4, Accounting Quality2 is also a composite measure, which is the average of the three ranks scaled by the number of observations: the rank in Big4, the rank in Earning Predict CF, and the rank in Abnormal Accruals (in decreasing order)

In Panel C the difference and its significance between coefficients in two regressions of high and low volatility samples are measured by linear probability model. The P-values in the tests of the difference are from two-tailed Z-tests as per Clogg et al. (1995) and Paternoster et al. (1998). All other variables are defined in Appendix. *, **, and *** indicate that the estimated coefficients are statistically significant at the 0.10, 0.05, and 0.01% level, respectively. P-values in brackets are from two-tailed t-tests