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

 
4.1.2 Asymmetric Timeliness of Loss Recognition 
As indicated earlier, the findings on whether adoption of 
IFRS
leads to less or 
more timely loss recognition for all assets (i.e., 
PPE
, other non-financial assets, and 
financial assets) are mixed and inconclusive (e.g., Barth et al. 2008; Ahmed et al. 2010). 
In this sub-section, I examine whether timely loss recognition for all assets is higher 
under 
IFRS
than under domestic GAAP among EU firms in my total sample. To do so, I 
use Basu (1997) model which indicates whether ‘bad news’ is recognized in earnings in a 
were likely less disciplined in their investment decisions during the pre-
IFRS
period relative to the post-
IFRS
period. 

---

27 
more timely manner than ‘good news.’
32
The Basu (1997) model uses the following 
equation: 
EARN
it
 = α
0
 + α
1
D
it
 + α
2
RET
it
 + α
3
D
it
*RET
it
 + u
it
(5)
 
The dependent variable, 
EARN
it
, is earnings per share before extraordinary items 
scaled by stock price at the fiscal year-end of 
t
-1 and 
RET
it
is the 12-month compound 
return ending three months after the fiscal year-end of 
t
. 
D
it
is an indicator variable 
equaling one if 
RET
it
is negative, and zero otherwise. 
RET
it
is used as a proxy for 
economic gains (good news) when it is positive and for economic losses (bad news) when 
it is negative. 
α
2 
captures the sensitivity of earnings to good news. 
α
3
captures the 
incremental sensitivity of earnings to bad news relative to good news (i.e., timely loss 
recognition or, in particular, the asymmetric timeliness of loss recognition in earnings). I 
am interested in examining the asymmetric timeliness of loss recognition in earnings 
following 
IFRS
adoption. Accordingly, I expand equation (5) as follows:
EARN
it
* = α
0
 + α
1
D
it
+ α
2
RET
it
 + α
3
D
it
*RET
it
 + α
4
IFRS
it
 + α
5
IFRS
it
*D
it

+ α
6
IFRS
it
*RET
it
 + α
7
IFRS
it
*D
it
*RET
it
 + u
it 
(6) 
Following Ahmed et al. (2010), I measure 
EARN
it
*
, the dependent variable in 
equation (6), as the residual from regressing 
EARN
it
(defined above) on industry and 
country fixed effects.
33
In equation (6),
 α
7
measures any change in the asymmetric 
timeliness of loss recognition following 
IFRS
adoption. I interpret a positive 
α
7
coefficient as indicating an increase in asymmetric timeliness of loss recognition 
following 
IFRS
adoption.
32
As indicated earlier, prior literature (e.g., Dietrich et al. 2007; Givoly et al. 2007) has shown that Basu 
(1997) model suffers from several drawbacks. Even though, Basu (1997) model is considered the most 
widely-used model to estimate the asymmetric timeliness of loss recognition.
33
This approach of measuring the dependent variable in equation (6) is also in line with Ball et al.’s (2011) 
suggestion. In their recent paper, Ball et al. (2011) argue that the correlation between the ‘expected’ 
components of earnings and returns biases the estimate of the asymmetric earnings timeliness in Basu 
(1997) model. To address this bias problem, they suggest fixed-effects regression as an example. Therefore, 
by obtaining the residual from regressing earnings on country and industry fixed-effects, I control for the 
‘expected’ earnings component across countries and industries.

---

28 
Table C4 presents the results of a pooled regression based on equation (6). The 
estimated coefficient 
α
7
is positive, although it is not significant. This finding suggests 
that there is some modest evidence of an increase in the asymmetric timeliness of loss 
recognition following 
IFRS
adoption and contrasts with that of Ahmed at al. (2010) who 
document a significant decrease in the asymmetric timeliness of loss recognition 
following 
IFRS
adoption.
Taken together, my preliminary results suggest that there is more timely loss 
recognition for 
PPE
following 
IFRS
adoption. The frequency of impairment losses for 
PPE
is significantly greater in the post-
IFRS
period relative to the pre-
IFRS
period and 
there is some modest evidence of an increase in the asymmetric timeliness of loss 
recognition following 
IFRS
adoption. Therefore, these results are consistent with my 
prediction of having more timely loss recognition for 
PPE
under 
IFRS
strict impairment 
rules.