that produce products from different sub-sectors, so our sample is restricted to 44,712
observations on 20,521 manufacturing firms.17
[TABLE 1 ABOUT HERE]
For the production function analysis, the output variable is defined as the gross value
added of the firm deflated by the industrial output price index relevant to the two-digit
sub-sector. It is constructed by adding total labour costs to gross profit. Inputs include
(i) labour, measured as the total number of persons employed at the end of the year in
question,18 and (ii) capital, measured as the total assets of the firm at the end of the year
deflated by a capital price series. Descriptive statistics are presented in Table 2.
[TABLE 2 ABOUT HERE]
Table 3 outlines the number of sector switchers, exits and entrants in the sample. Sector
switching “OUT” counts the number of firms, which change main production sector
(two-digit ISIC) in the subsequent year. Similarly, sector switching “IN” documents the
number of firms that have just entered a new sector as compared to the previous year.
We are interested in comparing firms that fall into the sector switching “OUT” and “IN”
categories to “real” exits and new entrants, respectively. Around 4.6 percent (on
average) switch out of a given sector each year. This is somewhat below the average
number of exit firms (8-10 percent on average exit each year, depending on the sample
considered). Even larger differences exist between sector switchers “IN” and new
entrants, mainly due to the nature of the data (registered firms).19
[TABLE 3 ABOUT HERE]
17 As a robustness check we also consider the sub-sample of firms which remain in the sample for the
entire period, thus eliminating the impact of exit/entry decisions.
18 All firms with four employees or less are excluded from the analysis, representing around 20 percent of
the total sample in all years.
19 Given that only registered firms under the enterprise law are covered makes analysing several aspects
of firm entry problematic. Firms entering in 2002, 2003 or 2004 may have existed for several years before
registering, and therefore do not constitute entrants in strict terms. Registration involves several benefits
to firms (easier access to credit etc.). However it also makes firms more visible to government authorities
(and especially tax collectors). It is therefore uncertain during which stage in their life-cycle a firm
decides or is forced to register. Moreover, the post 2001 surveys did not collect information on
establishment year. Given the nature of the data we therefore focus most of our attention on efficiency
differentials between incumbents, exits and sector switching firms in what follows.
13