Research design
4.1 Sample selection
We begin to select firms included in the S&P 1500 index from the Compustat
Execucomp database for years 2007-2010 as CEO tenure needs to be computed using
the CEO data from this database. The Compustat Execucomp database provides the
data on CEO compensation, shareholding, beginning year of being CEO, etc. for those
firms. This procedure yields a raw sample of 5,037 firm-year observations. Next,
we collect the data from the Compustat North America database and the CRSP database
to compute all dependent and independent variables used in the regression analysis.
After the elimination of observations with missing data, the sample size is reduced to
3,436 firm-year observations. Like Roychowdhury (2006), we delete firms in regulated
industries (SIC codes between 4400 and 5000) and banks and financial institutions
(SIC codes between 6000 and 6500) to further reduce the sample to 3,136 firm-year
observations. As Roychowdhury (2006) finds that firms manipulate real activities to
avoid reporting annual losses, we restrict the sample to firms with small positive
earnings[3]. This yields a final sample of 100 firm-year observations for years 2007-2010.
We manually collect the data on audit committee and board characteristics by reviewing
proxy statements downloaded from the EDGAR database for these 100 firm-year
observations. Panel A of Table I reports the sample breakdown by year. There are 18, 24,
38 and 20 firms in 2007, 2008, 2009 and 2010, respectively. Panel B of Table I reports
that there are 58 firm-year observations in manufacturing industries and 42 firm-year
observations in non-manufacturing industries.
4.2 Measurement of real earnings management
Like Roychowdhury (2006), we consider three real earnings manipulation approaches:
(1) sales manipulation;
(2) reduction of discretionary expenditures; and
(3) overproduction.
Sales manipulation means that managers attempt to temporarily increase sales during
the year by offering price discounts or more lenient credit terms. Reduction of
discretionary expenditures means that managers reduce discretionary expenditures
such as advertising expenses, R&D expenses, and selling, general and administrative
expenses to increase earnings. Overproduction means that managers increase earnings
by producing more units of goods than necessary to lower fixed costs per unit and
thereafter cost of goods sold.
Offering price discounts to increase sales results in lower margins. As margins decline,
the cash inflow per sale becomes lower. Similarly, offering more lenient credit terms like
lower interest rates (zero-percent financing) leads to lower cash inflow. Thus, sales
manipulation activities are associated with lower current-period cash flows from
operations than what is normal given the sales level. Our first measure of real earnings
management is abnormal cash flows from operations (ACFOt), which are estimated using
the following cross-sectional regression for every two-digit SIC industry and year[4]:
Sedang diterjemahkan, harap tunggu..