1. Introduction
Productivity is an important indicator of manufacturing sector performance. Many
studies have documented the link between productivity growth and other indicators of
success, including employment creation, export status and technology adoption. Tybout
(2000) reviews literature relevant to developing country contexts, and focus has
regularly been on the relationship between firm turnover and productivity. Firm level
data have been used extensively, and many studies suggest that entry firms are more
efficient than enterprises exiting a particular sector. Accordingly, it is widely agreed that
firm turnover is an important source of sector-level productivity growth.1 No attention
has however so far been paid to the potential effect, differences in the delineation of
entrants and exits may have on efficiency outcomes. For example, Aw et al. (2001) pool
firms, which change legal ownership form, location or sector, with respectively firms
that close down production altogether (“real” exits) and newly established firms (“real”
entrants).
Our prior is along the lines of Bernard et al. (2006a) that while pooling is common it is
also problematic. Efficiency levels are likely to vary among the different kinds of
exiting and entering firms, and this suggests that the overall contribution to productivity
growth from firm turnover is a composite. Similarly, there are several reasons to believe
that the factors which affect sector switching (i.e. a firm’s choice to change sector) may
be very different from those, which influence “real” exits and entrants. To illustrate,
efficient rural firms may tend to move towards urban growth centers to benefit from
agglomeration advantages (Henderson, 1986). Pooling these firms with other exits can
be seriously misleading if the analyst is interested in understanding the determinants of
firms that close down altogether or vice-versa. Well established sector switchers may
also have better management experience than new entrants as noted in the overview by
Bartelsman and Doms (2000); and switchers may have better knowledge of the general
business environment and the market conditions facing the firm. All this suggests that
(i) potential aggregation errors should not be ignored, and (ii) sorting out the
characteristics of different groups of firms is likely to add to our understanding of firm
dynamics and the contribution of firm turnover to productivity growth.
1 See for example Aw et al. (2001), Bartelsman and Doms (2000) and Tybout (2000).