*This preview shows
page 1. Sign up
to
view the full content.*

**Unformatted text preview: **es usually include large amounts for severance payments.
Employers also lose investments made in ﬁrm-speciﬁc training if employees
are released when demand falls and new employees must be hired when
demand picks up again. In addition, companies experience productivity
losses because morale declines when employees are laid off, and they may
experience more turnover because employee loyalty is eroded.
H4b: The degree of stickiness increases with the employee intensity (ratio of number of employees to sales revenue) of the company. 3. Empirical Tests of Stickiness of SG&A Costs
An empirical model that enables measurement of the SG&A response
to contemporaneous changes in sales revenue and discriminates between
periods when revenue increases and revenue decreases is presented. The
interaction variable, Decrease Dummy, takes the value of 1 when sales revenue
decreases between periods t − 1 and t , and 0 otherwise.
Model (I):
log SG&Ai,t
SG&Ai,t −1 = β0 + β1 log Revenuei,t
Revenuei,t −1 + β2 ∗ Decrease Dummyi,t ∗ log Revenuei,t
+ εi,t .
Revenuei,t −1 This model provides the basis for our test of stickiness of SG&A costs.3
Because the estimation is cross-sectional with a wide variety of industries and
large differences in the size of ﬁrms, the ratio form and log speciﬁcation
improves the comparability of the variables across ﬁrms and alleviates potential heteroskedasticity. Empirically, the Davidson and MacKinnon [1981]
test rejects the linear form in favor of this loglinear model. Results are qualitatively similar to those presented for all our models when we estimate them
with linear speciﬁcations.
The log speciﬁcation also accommodates economic interpretation of
the estimated coefﬁcients. Because the value of Decrease Dummy is 0 when
revenue increases, the coefﬁcient β1 measures the percentage increase in
SG&A costs with a 1% increase in sales revenue. Because the value of
3 If the traditional ﬁxed- and variable-cost model is valid, upward and downward changes
in costs will be equal and consequently β2 = 0. Furthermore, if ﬁxed costs are present, β1 < 1,
signifying economies of scale. STICKY COSTS 53 Decrease Dummy is 1 when revenue decreases, the sum of the coefﬁcients,
β1 + β2 measures the percentage increase in SG&A costs with a 1% decrease
in sales revenue. If SG&A costs are sticky, the variation of SG&A costs with
revenue increases should be greater than the variation for revenue decreases. Thus, the empirical hypothesis for stickiness, conditional on β1 > 0
is β2 < 0.4 3.1 DESCRIPTION OF DATA The primary variables used in our analysis are SG&A costs (annual Compustat #189) and net sales revenue (annual Compustat #12). The data set
includes annual data for industrial ﬁrms covering the 20 years from 1979
to 1998. Data were drawn for all ﬁrms included in the PST (primary, supplementary, and tertiary) and full-coverage ﬁles of Compustat 1999. We
sc...

View
Full
Document