In this study, we find that both the logit
regression and discriminant analysis models
identify the same variables as the most appropriate
for discriminating between bidder
and target firms. However, though logit regression
suggests liquidity as the only
important variable for the stated purpose,
discriminant analysis, besides suggesting
liquidity as the most important variable for
the sample data, also reveals the relevance
of the other variables such as profitability,
size, risk and growth. Given that both techniques
have yielded similar results, one could
rely on the findings as obtained from both.
Liquidity, profitability, size, growth and
risk serve as the appropriate predictive variables
for discriminating between bidders and
targets in the case of Indian mergers and
acquisitions during the period 2002 to 2006.
The discriminant model is shown to have a
moderate degree of classification accuracy
(64.8 per cent) and is supported by the logistic
regression model (67.9 per cent). Beyond
this, when applied to a holdout sample of
bidder and target firms, the logit model produced
slightly better results. The signs and
magnitudes of the coefficients from the discriminant
function indicate that the higher
profit and lower growth bidder firms take
over the lower profit and higher growth
targets.