The balance sheet identity model : testing ordinary least square assumptions using Malaysian market / Merani Che Ali

Models based on a relation between market value and book values employing
balance sheet variables are used continuously in the accounting research literature.
The basic model is well known as The Balance Sheet Identity model as first
mentioned by Landsman in 1986. Among other researchers who have based their
work on this model are Kane and Unal (1990), Shevlin (1991), Gopalakrishnan and
Sugrue (1993), McCarthy and Schneider (1995), Jennings et al (1996) and Pfeiffer
(1998). However, all of them were facing several econometric problems when
estimating the model.
Basically, these problems are related to the procedure for the estimation of the
parameters of a population regression line provided by the ordinary least squares
(OLS). OLS is based on a number of assumptions about the variables and the error
term that must be satisfied in order to ensure the interpretations of the regression
estimates are valid.
This study empirically examines whether the model using Malaysian data will
encounter the same econometric problems. In doing so, we tested five common
assumptions namely, normality, serial correlation, linearity, heteroscedasticity and
multicollinearity.
The empirical results of the test reveal that the models using Malaysian data are
facing linearity, normality and heteroscedasticity problems.