Variable Measurement The analysis of Turkish investment incentives i. Gini coefficient

1366 Disclosure index developed by Botosan 1997 is classified into the following five categories: background information, summary of historical financial statistics, key non- financial statistics, projected information and management discussion and analysis. Each category has total score that is measured by using several disclosure items. The score depends on the level of information provided in annual report. In measuring each item, Botosan 1997 provides the list of disclosure items with explanation and guidance anchor to define the score so it is useful to minimize subjectivity in determining the score. In contrast to US and Europe, Indonesia has no institution that provides disclosure index. Furthermore, most studies related to disclosure level in Indonesia employ Botosan Index Tjakradinata 2000; Adhariani 2004; and Budiarti 2007. As mentioned above, Botosan index could minimize subjectivity in measuring company disclosure. In measuring disclosure level, it is necessary to understand the style and format of annual report of listed companies. In Indonesia, Badan Pengawas Pasar Modal Bapepam and Lembaga Keuangan as a regulator to monitor stock exchange in Indonesia has released a decree No: KEP-134BL2006 Regulation No X.K.6 that state there is obligation for publicly listed companies to publish an annual report. The decree also provides a guidance of the style and content of annual report that is mandatory to be informed to the public. According to that decree, companies have to explain significant changes compared to the last annual report. The decree also requires certain information to be disclosed in the annual report, i.e., important financial summary, commissioner ‘s report, 235 Board of Director BOD report, company profile, management discussion and analysis, report on corporate governance, responsibility of BOD on annual report, and audited financial report. Specifically on management discussion and analysis, a company has to explain the following information: review on operation of each segment of business, financial performance analysis, discussion and analysis on business condition, significant changes, impact to the company, prospect and achievement and also explanation about major accounting policies being used. The company can provide additional information 235 Indonesian companies apply a two-tier board, i.e., the supervisory board or the board of commissioner and the management board or the board of directors. 1367 voluntary disclosure as well such as information about company strengths or other ‘good news‘ which can be beneficial to the company. εeasuring disclosure level can be performed by determining and evaluating disclosure items provided in the annual report which has to comply with the regulation standard. In this study, the level of disclosure is calculated based on the list of disclosure items and weighted according to weighted item based on study by Botosan 1997. The formula of total disclosure value is given below:    5 1 i ij i SCORE TSCORE Disclosure total TSCORE is total score for company j in category i to all category category 1 to 5. This score then will be divided by the maximum score of disclosure level:    5 1 max i i ij i TSCORE TSCORE DISC Ownership structure is proxied by the percentage of ownership by large shareholders block holders. Large shareholder is defined as a shareholder who own more than 5 of outstanding shares. Higher percentage of large shareholders shows a more concentrated ownership. The competition level reflects the competition among companies which belong in the same industry classification SIC. To measure the competition level, we use Herfindahl Index HI 236 . The value of competition is 1 minus Herfindahl Index 1-H1 and has a range between 0 until 1. The value close to zero means that the competition is low meanwhile the value close to one means that the competition is high. To calculate HI for each industry based on SIC, the study used the formula: 236 The other proxy for competition is Concentration Ratio CR. It is called CR-4 or CR-8 which reflect the number of four or eight mean the number of biggest companies Leuz 1999 and Harris 1998. This proxy is not used in this study due to not all industries have at least four public listed companies. This study uses Herfindahl Index because this index is commonly used in other research and also by US Department of Justice to investigate fraud of antitrust Birt, Bilson, Smith and Walley 2006. Besides Helfindahl Index is also well known Herfindahl- Hirschman Index HHI that has the same value with Herfindahl Index HI multiply by 10000. HHI has value 0 to 10000. If the value closes to 0, it shown perfect competition meanwhile if the value closes to 10000 means the market or industry is monopoly. 1368 2 1 Rj R HI ij n i J j    R ij is a revenue of company i in industry j, n j is a number of companies in industry j .    j n i j Rij R 1 R j is the revenue total of all companies in industry j.

3.2. Sample

To be included in the final sample, the observation has to meet the following criteria: 1 The companies are manufacturing listed companies; and 2 The company has an annual report in 2006. Furthermore, this study uses Standard Industry Code SIC to determine industry classification. This information is obtained from OSIRIS data base per April 1, 2008. On the other hand, we use one year annual report to measure company disclosure with the assumption that the format of annual report is constant over time and hence the annual report is a good proxy of overall company disclosure as well Botosan, 1997. 3.3. Empirical Model We employs Ordinary Least Squares Regression as shown below to investigate the influence of large shareholders, competition level and total assets to disclosure level. DSCORE = o + 1O+ 2C + 3TA +  ......................................................................1 where o = Intercept O = Large shareholder is defined as shareholder who own more than 5 of outstanding shares C = Competition TA = logTotal Aset  = error 1369

IV. EMPIRICAL RESULT 4.1 The Result of Sample Selection