Pengaruh income smoothing terhadap keinformatifan laba
Sri Wahyu Agustiningsih - S4308005 - Fak. Ekonomi dan Bisnis
The objective of this research is to obtain empirical evidence related to the effect of income smoothing on the earnings informativeness of manufacturing firms listed in Indonesia Stock Exchange (IDX) during 2005-2007 period. Earnings informativeness is measured using future earnings response coefficient (FERC) measuring the effect of current stock return on the future earning.
Population used in this research was all manufacturing firms listed in IDX during 2005-2007 periods. Sample was selected using purposive sampling by which it was obtained 65 firms. The research used multiple regressions as analysis tool with the aid of computer software for statistic SPSS version 16.00.
Result of the research shows that data were distributed normally for all variables. In the classic assumption test, there is no autocorrelation, multi co-linearity and heteroskedasticity tendency. Using multiple regressions model, two further supplement tests were done, main and advance test. Supplement test using earning persistence model shows that income smoothing may strengthen the persistence of earnings of a firm and in benchmark CKSS model it shows that the information on the future earning may be described in the current stock return. Result of main test income smoothing interaction in benchmark CKSS model shows that income smoothing may lower informativenes level of firm earnings. This result is supported by advance test using decomposition model.
From the result of study it is concluded that income smoothing significantly affects earnings informativeness. It is also suggested that future research on this topic prolong research period, enlarge research sample and differ sample into certain industry category so that it is found more in-dept and further results.
Keywords: income smoothing, earnings informativeness, future earning response coefficient (FERC), earnings persistence model, benchmark CKSS model, decomposition model.