This study examines the degree to which Loughran and McDonald (LM) word lists are informative at the item level of SEC filings, such as risk factors (RF) and management’s discussion and analysis (MDA) disclosures in 10-X reports. In this context, we explore if sentiment types are informative when associated with other material events, namely cybersecurity breaches. Our results support the assertion that sentiment types, beyond positive and negative, are informative at the individual disclosure item level, as tested in the RF and MDA sections. We also find that investors respond to different types of sentiment between RF and MDA. We find an economically significant estimated average economic impact of $469 million/firm. We further contribute to the literature by applying novel statistical methods that advance empirical accounting literature.
Measurement of dispersion and variation have been studied and evaluated in many applications. Volatility in the field of finance is an important measure as it directly impacts allocation, risk management, and valuation. Pitman Closeness criterion is used to compare estimators of standard deviation from equity returns in a control charting application. Three estimators are evaluated over the 30 DJIA component stocks in an effort to determine if one method of estimation has better performance within an application of control charting for identifying outliers. The study uses three sample sizes to also determine if the better estimator is sample size dependent.
Market Volatility has been investigated at great lengths, but the measure of historical volatility, referred to as the relative volatility, is inconsistent. Using historical return data to calculate the volatility of a stock return provides a measure of the realized volatility. Realized volatility is often measured using some method of calculating a deviation from the mean of the returns for the stock price, the summation of squared returns, or the summation of absolute returns. We look to the stocks that make up the DJIA, using tick-by-tick data from June 2015 - May 2016. This research helps to address the question of what is the better measure of realized volatility? Several measures of volatility are used as proxies and are compared at four estimation time intervals. We review these measures to determine a closer/better fit estimator to the true realized volatility, using MSE, MAD, Diebold-Mariano test, and Pitman Closeness. We find that when using a standard deviation based on transaction level returns, shorter increments of time, while containing some levels of noise, are better estimates of volatility than longer increments.
The findings of this study are similar to those of Chow, et al (2018) where the Parkinson Estimator outperforms the other range-based estimators when comparing the volatility of the weekly average of the daily returns. When looking at the variances of the weekly averages of the weekly returns the Parkinson Estimator is again superior when considering the MSE method of comparison. However, the Rogers & Satchell Estimator is slightly a better performer than the Parkinson Estimator when considering the comparisons of MAD and PC. As is often the case, the most appropriate method of comparison is often dependent on more than one aspect of your study. However, from these results the time interval is less impactful on the overall performance of the estimators than the method of overall comparison. Future research may look at shorter time horizons at monthly or weekly intervals, and could utilize high frequency data for establishing the calculated measure for variance.
We use the Pitman Closeness Criteria to compare four range-based volatility models. We utilize the daily high-low-open-close data for the S&P500, DJIA and NASDAQ indices, which are readily available, instead of a commonly used, yet more difficult to obtain, high frequency data to calculate the volatility of the daily returns. We then use the 30 equities of the DJIA to make similar comparisons of estimator performance. In both cases of the comparison of estimator performance for indices and equities, the better performing estimator is dependent on the method of measurement of variance, as well as the method of performance comparison.
