Evaluating XBRL’s Impact on the Timeliness, Accuracy and Coverage of Security Analysts’ Next-Year EPS Forecasts

• In Event: Non-GAAP Earnings and Earnings Forecasts

Abstract

This research uses Detail I/B/E/S data to study the impact of XBRL on the timeliness, accuracy, and coverage of security analysts' forecasts of firms' next-year EPS. We find that analysts’ next-year EPS estimates announced during 0-15 days after the filing of 10-Ks with XBRL during 2011-2013 were announced, on average, 8.0% to 10.1% faster than the estimates announced in the same 0-15 days following the filing of 10-Ks without XBRL during 2008-2010. Also, we find that analysts’ next-year EPS estimates announced during 0-30 days after the filing of first-, second-, and third-quarter 10-Qs with XBRL during 2011-2013 were announced, on average, 21.1% faster than the estimates announced in the same 0-30 days following the filing of 10-Qs without XBRL during 2008-2010. We find that XBRL’s improvements on the timeliness of analysts’ forecasts are restricted to short timeframes of 0-15 and 0-30 days following the filling of 10-Ks and 10-Qs. Also, we find evidence that using 10-Ks filed with XBRL helped analysts to reduce the mean absolute error of their next-year EPS forecasts by -20.9% in the first year when prior-year 10-Ks and current-year 10-Qs were filed with XBRL as compared to the last year when prior-year 10-Ks and current-year 10-Qs were filed without XBRL. We find that the number of analysts’ forecasts increased by 6.6% from 83,077 to 88,531in the first year when prior-year 10-Ks and current-year 10-Qs were filed with XBRL as compared to the last year when prior-year 10-Ks and current-year 10-Qs were filed without XBRL. Also, we find an increase in the clustering of estimates nearer to the 10-Q reports in first year with XBRL as compared to the Last Year without XBRL. The improvements we find in timeliness are perhaps more important for the future than for the present because we demonstrate that the XBRL-based system is now in place for users to electronically input and process larger volume of financial statement data more efficiently and more frequently.

Authors

• Sherwood Lane Lambert, University of West Florida - Pensacola

• Kevin Krieger, University of West Florida