Again, one can see that the TDAR feature set outperforms the others.
Abbreviations: ACE = Acquisition and Control Environment, AER = active error rate, AR = autoregressive, EMG = electromyogram, IMES = Implantable MyoElectric Sensor, LDA = linear discriminant analysis, TD = time-domain, TDAR = timedomain/autoregressive, TER = total error rate, TMR = targeted muscle reinnervation.
Exhibit 2, Panel B presents results for the TDAR for three groups classified according to the percentage of equity held by institutions.(15) The results clearly indicate that the higher the institutional ownership, the less negative the TDAR.
Next, in order to control for other possible effects, we estimated the cross-sectional relation between TDAR and institutional ownership by estimating weighted least squares (WLS) regressions of the form:
The results reported in Exhibits 2 and 3 documenting a positive relation between TDAR and the level of institutional ownership are consistent with the effective-monitoring hypothesis.
In addition to the control variables used in the previous section, we also included the announcement-period price reaction (TDAR) as an additional control variable in the event that some analysts use this information to revise their earnings forecasts.(20) We find no relation between AFR and institutional ownership.
In general, there is a significant and positive relation between abnormal forecast revisions in the five-year earnings growth forecast and TDAR. The positive relation between LTAFR and TDAR indicates that equity issue announcements convey information regarding long-term earnings also.
For each of these two groups separately, we estimated weighted least squares regressions between TDAR, AFR and LTAFR with institutional ownership.
16 While Asquith and Mullins |1^ find a positive relation between TDAR and run-up, Masulis and Korwar |20^ find exactly the opposite relation.