Where technology meets regulation: Investigating the impact of the SEC's electronic filing requirements on earnings estimate accuracy

This study examines the impact of the SEC's electronic filing requirements (EDGAR) on the accuracy, dispersion and volatility of earnings estimates. Using a data set of 275 firms that had at least 4.5 investment analysts providing earnings estimates over a 28-year period, I study the impact of EDGAR on forecast accuracy, the dispersion of earnings estimates and their volatility across the year. Analysis is done at both the individual analyst and consensus levels. Existing research evaluating the impact of EDGAR focuses on market returns and trading volume. This study expands that analysis to consider measures of forecast accuracy. The research looking at the SECs Regulation FD does include measures for forecast accuracy and dispersion but has no differentiation in the pre-EDGAR period related to the requirement for EDGAR filings. This study measures forecast accuracy in the pre-EDGAR period.;The analysis shows that when aggregated, the consensus estimate in the post-EDGAR period (beginning 1/1/1994) has reduced forecast errors, reduced dispersion across analysts and reduced volatility across the year. These results are significant at the .01 level. When considering the data for the individual analysts, the results are different. In the post-EDGAR period, both the forecast error percentage and volatility increase. There is no statistically significant impact of EDGAR on dispersion when analyzing individual analyst data. The difference in results indicates that the distribution of data to a broader population with fewer filters creates 'noise' in the system resulting in increased volatility and less accuracy in the individual estimates. In aggregate however, these individual perspectives drive to a more accurate, more consistent and less volatile estimate. With more individual positions considered, the overall results are improved.;In this analysis, though the EDGAR variables are statistically significant, the R-squared values in all models are very low. The presence of EDGAR does not explain a meaningful portion of the differences in the accuracy, dispersion or volatility of earnings estimates. Technology does have an independent effect however to impact the effectiveness of regulation; it must be combined with other factors. A logical continuation of this research is to explore these factors. With the SEC's technology initiative to incorporate XBRL data in the disclosure documents underway, the time is now to leverage advances in technology to improve regulatory effectiveness.