| Summary: | The main objectives of this study are to investigate the effect of the transaction gains (losses) on the earnings response coefficients and to investigate how investors respond differently on three accounting methods of the transaction gains (losses). The first method, based on benchmark treatment of PSAK No. 10, treats transaction gains (losses) as revenue (expense). The second, based on alternative treatment of PSAK No.10 ( I S M No. 4), treats the gains (losses) as partially capitalized accounts. The third, based on BAPEPAM regulation No. V m G10, treats them as fully-capitalized accounts. Another objective is to provide evidence on factors affecting earnings response coefficients. This study uses the sample of 225 firms listed in the Jakarta Stock Exchange. This study uses a time-series data of 225 firms during 1993-1999, and cross-sectional data of 1999. The hypotheses are tested applying two empirical models. The first, cumulative abnormal returns are regressed on unexpected earnings and annual return. The second, earnings response coefficients are regressed on transaction gains (losses), earnings persistence, earnings growth, earnings predictability, beta risk, capital structure, f r size and industrial effect. Those are called long-windows event study im method. Cumulative abnormal returns are measured using market-adjusted model with fivedays window period. Unexpected earnings are measured using the random walks model. A n u d returns are measured using the ratio between stock price changes and sock price of the previous period. Earnings response coefficients are measured using the magnitude of coefficients on the first regression model. Transaction gains (losses) are measured using a n n d data of total transaction gains (losses). Earnings persistence is measured using the magnitude of coefficients of time-series earnings regression. Earnings growth is measured using the ratio of market to book value of equity. Earnings predictability is measured using the variance of earnings shock of time-series earnings regression. Beta risk is measured using market-adjusted model. Capital structure is measured using leverage. Firm S ize is measured using market value of equity. Industrial effect is measured using dummy variables. The results of this study are as follows: (1) the effect of transaction gains (losses) on earnings response coefficients is statistically si@cant; (2) investors respond inMerently the fkms recognizing different methods of transaction gains (losses); and (3) the effect of earnings persistence and earnings predictability on earnings response coefficients are statistically si@cant. This study, hence, primarily contributes to regulations of BAPEPAM and Financial Accounting Standard Committee of Ikatan Akuntan Indonesia, especially to BAPEPAM regulation No.V m GlO and ISAK No.4 ofFinancial AccoUnting Standard Committee, IAI.
|
|---|