| dc.description.abstract | Industry-switching occurs as firms exit one industry and then enter another industry.
Due to the combination of exit and entry which are two core issues in traditional industrial
economics, industry-switching shows the multiplicity of firm dynamics. In addition,
a large number of empirical results display that it spreads widely in reality. However, a
formal study on industry-switching is almost absent from the literature. Owing to the
ignorance of industry-switching, i.e. neglecting industry-switching firms’ pre-switching
and post-switching situations, existing researches are incapable of rigorously and completely
capturing firms’ dynamic decision process, which is likely to cause incorrect
inference for industry development. In view of this, this dissertation focuses on the
theoretical and empirical study on industry-switching First, we utilize the interdependence
coefficients in the input-output table to create an industry-switching degree index
(ISDI) to describe how firms switch among industries. ISDI improves disadvantages
in the traditional categorization method to unify and quantify each individual firm’s
heterogeneous industry-switching behavior. Based on ISDI, we then employ Heckman’s
two-step estimation and the quantile regression technique to investigate the industryswitching
behavior among Taiwan’s electronics manufacturing industries during 1986 to
1991. Empirical results exhibit that industry growth rate, concentration rate, firm age,
scale, and productivity affect firms’ switching decision. Moreover, the effect of each of
these determinants vary with firms’ degree of industry-switching. At last, we extend
existing theoretical model to allow industry-switching to also be an endogenous decision
other than staying and exiting decisions. In equilibrium, an interaction among firm
characteristics, industry characteristics, and market conditions shapes the staying, failure,
and industry-switching choices. Our industry-switching model enables us to explain
some stylized facts that are observed by us but can not be interpreted by the literature.
Furthermore, it provides with the correct inference for industry development. | en_US |