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

H2. 
UK firms that switched from fair-value accounting under UK GAAP to 
historical cost accounting with strict impairment rules under IFRS will exhibit 
greater reductions in over-investment in PPE relative to other EU firms that 
used historical cost accounting with impairment testing prior to IFRS adoption. 
 
 
17
I assume that each firm normally has multiple assets of 
PPE
. 

---

18 
CHAPTER III 
MEASURES, RESEARCH DESIGN, AND SAMPLE SELECTION
3.1 Measures of Over-investment 
To measure the increase or decrease in over-investment, I first examine whether a 
firm is more likely to over-invest following the approach in Biddle et al. (2009). I use two 
firm-specific characteristics that affect the likelihood a firm will over-invest. Specifically, 
I use the firm’s cash level (
CASH
) and free-cash flow (
FCF
) as two partitioning variables 
based on the argument that firms with low 
CASH
and low 
FCF
are more likely to be 
financially constrained. Alternatively, firms with high 
CASH
and high 
FCF
are more 
likely to face agency problems and to over-invest (e.g., Jensen 1986; Harford 1999; Lie 
2000; Richardson 2006).
FCF
is measured as cash flow from operating activities less predicted capital 
expenditures (
CAPEX
). Following Biddle et al. (2009), I estimate predicted 
CAPEX
as a 
function of sales growth based on the following model: 
 
CAPEX
t
 = γ
0
 + γ
1
SALES_GROWTH
t-1
 + υ
t
 
 

(1) 
where 
CAPEX
t
is the natural logarithm of capital expenditures scaled by lagged total 
assets in year 
t
. 
SALES_GROWTH
t-1
is the percentage change in net sales in year 
t-1
.
I 
estimate equation (1) by year for each 2-digit industry with at least 10 observations in a 
given year. I then estimate the predicted capital expenditures for each firm 
i
using the 
estimated coefficients from equation (1): 
PREDICTED_CAPEX
it
 = γ
0
 + γ
1
SALES_GROWTH
it-1
 

(2) 
where 
PREDICTED_CAPEX
it
is predicted capital expenditures for firm 
i
in year 
t
. 
Hence, the free-cash flow (
FCF
it
) for each firm 
i
in year 
t
is measured as follows:
FCF
it
= 
CFO
it
– 
PREDICTED_CAPEX
it
(3) 
where 
CFO
it
is cash flow from operating activities scaled by lagged total assets.

---

19 
My second proxy for the likelihood of over-investment is cash level (
CASH
). For 
each firm 
i
, I measure 
CASH
it
 
as cash and cash equivalents at end of year 
t
scaled by 
lagged total assets.
After measuring 
FCF
and 
CASH
for each firm-year, I rank firm-years into terciles 
based on 
FCF
and 
CASH
. I re-scale the ranked values to range between zero and one. 
Following the approach in Biddle et al. (2009), I then create a composite score measure, 
OVER_INV
, which is computed as the average of ranked values of the two partitioning 
variables. I do so because each variable is likely to measure the likelihood of over-
investment with error and aggregating these two variables reduces the measurement error 
in the individual variables. Thus, 
OVER_INV
measures the likelihood of over-investment 
based on 
CASH
and 
FCF
. As 
OVER_INV
increases, the likelihood of over-investment 
increases.