Small Negative Surprises : Frequency and Consequence

Using a large sample of quarterly observations for the 16 years, 1984-99, I present four types of related temporal evidence: (1) a decrease in the tendency of managers to report quarterly earnings that fall slightly short of analyst estimates [small negative surprises of no more than three cents]; (2) the temporal decrease in the tendency of managers to report small negative surprises pertains more to growth than to value firms; (3) the adverse valuation consequence of reporting small negative surprises has increased in severity in recent years; and (4) the temporal increase in the adverse valuation consequence of reporting small negative surprises pertains more to growth than to value firms. My frequency results are robust to alternative definitions of small negative surprises, and my valuation results are robust to including median surprises as a potential correlated omitted variable and are not due to temporal changes in the frequency of losses