R] is PYEC (the prior-year earnings change, previously defined as [E.
M_UP] is PYEC for cases of momentum upwards (when PYEC and CYEC have the same sign and are positive) and zero otherwise.
This difference may be explained by our scaling of the forecast error and PYEC by price.
To assess the extent of this bias we also ran Equation (4) again but distinguishing between the two types of reversion, PYEC positive/CYEC negative and PYEC negative/CYEC positive.
However, when PYEC is negative, they seem to be especially surprised when earnings decline further.
The qualitative interpretation of our results is similar to that of Easterwood and Nutt, yielding a positive, but insignificant response to prior-year earnings change for normal values of PYEC and a significant negative shift for high values.
This result shows that analysts overreact to high values of PYEC and underreact in other cases.