Page 83 - Azerbaijan State University of Economics

THE JOURNAL OF ECONOMIC SCIENCES: THEORY AND PRACTICE, V.72, # 1, 2015, pp. 61-94

As exhibited in Figure above, the returns (Ʈ) in various time intervals are the

following (Campbell 1997):

- Estimation window

Ʈ incorporates returns from T0+1 through T1

- Event window

Ʈ incorporates returns from T1+1 through T2

- Post-estimation window

Ʈ incorporates returns from T2+1 through T3

The length of estimation and event windows are specified by:

L1 = T1 - T0

L2= T2 - T1

For the purpose of our study, we have applied the estimation window of 321

days, meaning that information on returns on every security and corresponding

control market index in our sample was collected from 300 prior and 20 days after

the announcement day of a transaction. Brown and Warner (1985) stated in their

research that estimation period for daily security prices of 120 days before the day of

event are satisfactory to calculate the abnormal returns.

2. Select the sample of analysis

Before selecting the sample, researcher should establish the criteria for screening

and filtering the firms to be included in study. One of the prerequisites for event study

is to choose only public companies listed on any stock exchange. Other criteria can be

regarding size, industry and geographic coverage of the companies, form of payment

83