Microsoft Word 2007 ichrie conference Proceedings Final-Final 06-06-07. doc

 
2007 Annual International CHRIE Conference & Exposition 
131
Bublitz and Ettredge (1989), we calculate the dependent variable, CAR, as the sum of residuals from a market model 
over a 12-month period, using a value weighted market index of New York and American Stock Exchange firms. 
The 12-month accumulation period run from nine months prior to the fiscal year-end through three months 
afterwards (Beaver, Clarke, & Wright, 1979).
Earnings are decomposed as follows: 
 
 
EPS = SPS – MPS – OEXPS
 
 
 
 
 
(3)
 
Where 
EPS = COMPUSTAT annual data item 58, primary earnings per share excluding extraordinary items and 
discontinued operations, 
SPS = COMPUSTAT data item 12, annual net sales, divided by data item 54, the number of shares used in 
computing primary EPS, 
MPS = COMPUSTAT data item 45, annual advertising expense, divided by data item 54, 
OEXPS = other expenses per share, a residual amount: SPS – EPS – MPS. 
Albrecht, Lookabill and McKeown (1977) provide evidence that the time-series of earnings deflated by the 
beginning of period stockholders’ equity can be appropriately modeled as a random walk. Therefore, our deflated 
earnings forecast error per share (EFEPS) for firm 
i
in year 
t
is: 
EFEPSi,t = (EPSi,t – EPSi,t-1)/Pi,t-1, 
 
(4) 
Bublitz and Ettredge (1989) argue that the behavior of changes in advertising outlays per share, deflated by 
per-share stock price, appears to be white noise. Therefore, we use the random walk model to compute forecast 
errors for earnings components and then deflate by the share price at the beginning of a year. This practice is 
common in current cost market studies for which short data histories dominate (Bublitz, Frecka, & McKeown, 1985; 
Schaefer, 1984). 
MFPSi,t = (MPSi,t – MPSi,t-1)/Pi,t-1
, 
(5) 
SFEPSi,t = (SPSi,t – SPSi,t-1)/Pi,t-1
, 
(6) 
OEXFEPSi,t = (OEXPSi,t – OEXPSi,t-1)/Pi,t-1
, 
(7) 
Based on discussions in the previous section, we construct our model as follows: 
CARi,t = A0 + B1SFEPSi,t + B2MFEPSi,t + B3OEXFEPSi,t + ei,t
(8) 
The hypothesis of interest using the “zero” benchmark is: 
H1: B
2
≥
0. 
Inability to show significance in a directional test of B
2
less than 0 is consistent with marginal marketing 
outlays constituting long-lived, zero net-present-value investments. On the other hand, if the directional test gives 
significant result, most of marketing outlays expire in the current period.
When OEXFEPS is used as a benchmark, our hypothesis of interest is: 
H2: B2 
≤
B3 
Inability to show significance in a directional test of B2 less than or equal to B4 is consistent with 
marketing outlays being evaluated by the market as similar to conventional expenses. Alternatively, given our 
analysis, significant results from the directional test is consistent with marketing outlays having long-lived benefits, 
thus creating intangible assets.