251 /302
251
臺大管理論叢
第
28
卷第
2
期
2.1.1.3 Behavioral Explanations
Based on the psychology literature, Chen, Gores, and Nasev (2013) argue that
overconfident CEO are more likely to overestimate future demand and thus less likely to
cut SG&A costs when sales decrease. Using a sample of S&P 1500 firm between year
1992 and 2011, they find that SG&A cost stickiness increases with the degree of CEO
overconfidence.
2.1.1.4 Consequences of Cost Stickiness
Another strand of cost behavior literature analyzes the properties and consequences
of cost stickiness. Banker and Chen (2006) demonstrate that when a time-series earning
forecast model incorporates cost stickiness, the accuracy of this model increases
substantially over that of other models with only the line items in the financial statement.
In contrast to the conventional view that an increase in the SG&A cost-sales ratio as a
negative signal about future profitability and firm value, Anderson, et al. (2007) argue that
an increase in the ratio of SG&A to sales could be driven by cost stickiness, which may
represent deliberate retention on SG&A resources based on managers’ expectation that
revenue will increase in the future. Anderson et al. (2007) find that future earnings are
positively associated with changes in the SG&A cost-sales ratio in periods in which sales
decline. Anderson et al. (2007) suggest that abnormal returns may be earned on by going
long on firms with high increases in the SG&A cost-sales ratio and going short on firms
with low increases in the SG&A cost-sales ratio in sales-declining periods.
Some studies investigate analyst forecasts. For example, Kim and Prather-Kinsey
(2010) argue that because of difficulty in obtaining internal cost data that contains various
cost components and cost drivers, using proportionate cost model, analysts imperfectly
adjust cost behaviors, resulting in systematic earnings forecast errors which are positively
associated with their’ forecasted sales growth rates. Moreover, Weiss (2010) documents
that firms with stickier cost behavior have less accurate analyst earnings forecasts than
firms with less sticky cost behavior. He argues that stickier costs result in a smaller cost
saving when activity levels decrease. Smaller cost saving leads to a greater decline in
earnings, increasing the variability of the earnings distribution, thus, less accurate earnings
forecast. Weiss (2010) further find that firms with stickier costs have lower analyst
coverage. A weaker market response to earnings surprises for firms with stickier cost
behavior is also documented, suggesting that investors recognize cost stickiness to some
extent and are aware that earnings predictability decreases and reported earnings are less