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

 
4.2.2 Test of H2 
To test whether UK firms that switched from fair-value accounting under UK 
GAAP to historical cost accounting with strict impairment rules under 
IFRS
(
FV_HS
firms) exhibit greater reductions in over-investment relative to other EU firms that used 
historical cost accounting with impairment testing prior to 
IFRS
adoption (
HL_HS
firms), 
I employ two tests. First, I employ a univariate test using difference-in-differences to 
examine the changes in capital expenditures between the pre-
IFRS
period and the post-
IFRS
period for the 
FV_HS
firms and for the 
HL_HS
firms. Second, I use a multivariate 
test by introducing to equation (4) an indicator variable, 
FV_HS
, that takes the value of 
one for 
FV_HS
firms, and zero otherwise. Specifically, I estimate the following 
regression: 
CAPEX
it
 = β
0
 + β
1
IFRS
it
 + β
2
OVER_INV
it
 + β
3
IFRS
it
*OVER_INV
it
 + β
4
FV_HS
it

+ β
5
IFRS
it
*FV_HS
it
 + β
6
FV_HS
it
*OVER_INV
it
 + β
7
IFRS
it
*FV_HS
it
*OVER_INV
it

+ β
8-18
CONTROLS
it
+
β
19-29
CONTROLS
it
*OVER_INV
it
 
+ 
ε
it
 
(7)
 
Consistent with my second hypothesis (H2), I predict that the estimated
 β
7
coefficient on 
the interaction between 
IFRS
, 
FV_HS
, and 
OVER_INV
to be negative. 
34
The percentage is computed as 
β
3
/mean_
CAPEX
(-0.3870/-3.3539). 

---

30 
Table C6 presents the univariate results, using difference-in-differences, of the 
effect on investment in 
PPE
for 
HL_HS
firms and 
FV_HS
firms that are more likely to 
over-invest based on the highest tercile of 
OVER_INV
. If there is indeed a reduction in 
over-investment in the post-
IFRS
period relative to the pre-
IFRS
period, then among 
these firms I expect lower investment in 
PPE
in the post-
IFRS
period. For 
HL_HS
firms, 
the mean 
CAPEX
is significantly lower (p < 0.01) in the post-
IFRS
period than in the pre-
IFRS
period. This result indicates that there is a reduction in over-investment in the post-
IFRS
period among 
HL_HS
firms and is consistent with my first hypothesis. Further, for 
FV_HS
firms, the mean 
CAPEX
is significantly lower (p < 0.01) in the post-
IFRS
period 
than that in the pre-
IFRS
period. This result indicates that there is a reduction in over-
investment in the post-
IFRS
period among 
FV_HS
firms and is also consistent with my 
first hypothesis. When comparing 
FV_HS
firms to 
HL_HS
firms using difference-in-
differences, the difference in mean 
CAPEX
is significantly negative (p < 0.05), which 
suggests that 
FV_HS
firms exhibit greater reductions in over-investment in the post-
IFRS
period relative to 
HL_HS
firms. 
Table C7 presents the multivariate results based on equation (7) of the effect on 
over-investment in 
PPE
for my total sample that includes both 
HL_HS
firms and 
FV_HS
firms. Consistent with my earlier results in Table C5, the estimated 
β
3
coefficient on the 
interaction between 
IFRS
and 
OVER_INV
is significantly negative (-0.3332; p < 0.01). 
As predicted by hypothesis 2, the estimated 
β
7
coefficient on the interaction between 
IFRS
, 
FV_HS
, and 
OVER_INV
is significantly negative (-0.6852; p < 0.05). The 
magnitude of 
β
7
coefficient suggests that when the likelihood of over-investment is high, 
FV_HS
firms, on average, exhibit an incremental reduction of approximately 20.4% in 
CAPEX
(capital expenditures) after 
IFRS
adoption relative to 
HL_HS
firms.
35
35
The percentage is computed as 
β
7
/mean_
CAPEX
(-0.6852/-3.3539). 

---

31 
Interestingly, I find no significant difference in the levels of investment in 
PPE
(capital expenditures) between the pre-
IFRS
and the post-
IFRS
periods. The multivariate 
results in Tables C5 and C7 reveal that the estimated 
β
1
coefficient on 
IFRS
is not 
significant. This finding suggests that the reduction in over-investment after 
IFRS
adoption is not at the expense of an increase in under-investment.
Taken together, the univariate and the multivariate results in Tables C6 and C7 
provide evidence that supports my second hypothesis (H2). UK firms that switched from 
fair-value accounting under UK GAAP to historical cost accounting with strict 
impairment rules under 
IFRS
exhibit greater reductions in over-investment in 
PPE
after 
IFRS
adoption relative to other EU firms that used historical cost accounting with 
impairment testing prior to 
IFRS
adoption. 
Overall, my findings in this sub-section provide evidence relevant to the debate 
on whether fair-value accounting for non-financial assets, such as 
PPE
, should become 
mandatory or more broadly used (e.g., Ball 2006; Barth 2006; Schipper 2005; Watts 
2006). In particular, my findings indicate that having historical cost accounting with strict 
impairment rules as an accounting alternative for non-financial assets under 
IFRS
promotes more efficient investment decisions. Therefore, my findings suggest that any 
shift towards a mandatory fair-value accounting for non-financial assets under 
IFRS
could have an adverse effect on investment efficiency.