Accounting choices under ifrs and their effect on over-investment in capital expenditures

3.2 Research Design 
To examine whether there are increases or decreases in over-investment after 
IFRS
mandatory adoption, I estimate the following regression: 
CAPEX
it
 = β
0
 + β
1
IFRS
it
 + β
2
OVER_INV
it
 + β
3
IFRS*OVER_INV
it

+ β
4-14
CONTROLS
it
 +
β
15-25
CONTROLS
it
*OVER_INV
it
 
+ 
ε
it
 
(4)
 
where 
CAPEX
it
and
 OVER_INV
it
are as previously defined. 
CAPEX
it
is obtained entirely 
from the cash flow statement and it represents cash paid for investments in 
PPE
by firm 
i
during year 
t
.
 IFRS
it
is an indicator variable equaling one for firms adopting 
IFRS
after 
January 1, 2005 and zero otherwise.
18
Following prior literature (e.g., Biddle and Hilary 2006; Biddle et al. 2009), I 
introduce a wide set of controls that are related to firms’ capital expenditures and, thus, 
18
All variables in this paper are measured in the pre-
IFRS
period separately from the post-
IFRS
period, 
which means there is no overlap between the two periods. For example, I scale variables in the post-
IFRS
period by lagged total assets in the post-
IFRS
period. 

---

20 
could confound my findings. Specifically, I control for a variety of innate firm 
characteristics that prior research has shown to be related to firms’ investment behavior. 
These controls include firm size (
SIZE
), tangible fixed assets (
TANGIBILITY
), cash flow 
to sales (
CFO_SALE
), length of the operating cycle (
OPERATING_CYCLE
), dividend 
payouts (
DIV
), market-to-book ratio (
MTB
), frequency of losses (
LOSS
), capital structure 
(
CAP_STRUCTURE
), and bankruptcy risk (
Z_SCORE
). I measure 
SIZE
as the natural 
logarithm of end of year market value of equity and 
TANGIBILITY
as the ratio of gross 
value of 
PPE
to total assets. 
CFO_SALE
is the ratio of cash flow from operating activities 
to net sales. 
OPERATING_CYCLE
is the natural logarithm of the sum of receivables to 
net sales and inventory to cost of goods sold multiplied by 360. 
DIV
is an indicator 
variable that takes the value of one if the firm paid a dividend, and zero otherwise. 
MTB
is the ratio of market value of total assets to book value of total assets. 
LOSS
is an 
indicator variable that takes the value of one if the firm’s net income before extraordinary 
items is negative, and zero otherwise. 
CAP_STRUCTURE
is the ratio of long-term debt to 
the sum of long-term debt and market value of equity. 
Z_SCORE
is a measure of 
bankruptcy risk (distress) computed following the methodology in Altman (1968).
19
Consistent with prior literature (e.g., Biddle and Hilary 2006), I predict that firms that are 
larger, more profitable, and have more tangible fixed assets, higher 
MTB
ratio, and lower 
bankruptcy risk, will have higher investment in 
PPE
(capital expenditures).
Further, I include closely-held shares (
CLOSELY_HELD
) in equation (4). Prior 
research (e.g., Berle and Means 1932; Jensen and Meckling 1976; Ang et al. 2000) 
suggests that the separation of ownership and control increases the level of agency 
conflict between insiders and outside shareholders and, hence, could affect managerial 
investment decisions. 
CLOSELY_HELD
is measured as percentage of closely-held shares 
19
See "List of Variable Definitions" in Appendix A for more details.

---

21 
for firm 
i
as reported by WorldScope.
20
I also include 
BIG4_5
, an indicator variable 
equaling one when the firm’s auditor is either one of the big four or five auditors and zero 
otherwise, to control for any potential effect this governance variable has on over-
investment (i.e., the interaction between 
BIG4_5
and 
OVER_INV
in equation (4)).
21
Finally, I include country and industry fixed effects in equation (4) to control for cross-
country differences (e.g., rule of law) and for industry-specific shocks that could affect 
firm’s investment behavior.
In the context of equation (4), 
OVER_INV
is increasing in the likelihood of over-
investment. The estimated 
β
3
coefficient measures the incremental effect that 
OVER_INV
has on investment in 
PPE
(
CAPEX
) after 
IFRS
mandatory adoption. Therefore, if over-
investment in 
PPE
is lower in the post-
IFRS
period relative to the pre-
IFRS
period as 
predicted by hypothesis 1 (H1), I expect a significantly negative 
β
3
coefficient.