Appendix: Construction of variables

Firms value added Y is defined as the real value of “industrial output'' minus the real value of “industrial input.'' The nominal values of industrial output and input are reported in the annual reports filed by firms in the Z-Park and are then deflated by industry-specific price deflators for output and input taken from the National Bureau of Statistics of China (2005).

The capital stock K is constructed using the following procedure. First, nominal investment is defined as the book value of current fixed assets minus the book value of fixed assets in the previous year, plus the value of the depreciation. Second, nominal investment is deflated by the price deflator for investment to create the value of real investment. The investment price deflator is again taken from the National Bureau of Statistics of China (2005). Third, the initial stock is defined as either the deflated book value of fixed assets in 1995 for firms that were established before 1995 or the deflated book value of fixed assets in the year of establishment for other firms. Finally, the capital stock is constructed by the perpetual inventory method.

Labor L is the total labor minus labor engaged in R&D activities. We subtract labor engaged in R&D to avoid double counting, since our value-added equation includes the R&D stock as another input.

Firm market share SHARE represents the ratio of a firm's sales to the total sales of the industry into which the firm is classified.

We construct a firm's own R&D stock using the following procedure.¹³ First, we obtain real R&D expenditure by dividing nominal R&D expenditure by the investment price deflator taken from the National Bureau of Statistics of China (2005). Second, we estimate the growth rate of firm i's real R&D expenditure g^r_i by running an OLS regression of the log of real R&D expenditure, r_i, on years. Third, if firm i was established before 2000, we estimate the firm's R&D stock in 2000, R_{i, 2000}, by

where 0.15 denotes the depreciation rate of the R&D stock and T_i is firm i's year of establishment or 1995 when firm i was established before 1995. That is, we assume that R&D activities before 1995 do not affect the R&D stock in 2000 or later. This assumption can be justified, since 65% of all firms in the data set for the year 2000 were established in or after 1995.¹⁴ If firm i was established in 2000 or later, the R&D stock in the year of establishment is equal to the real value of R&D expenditure in that year. Finally, the R&D stock for subsequent years is constructed using the perpetual inventory method assuming a depreciation rate of 15%.

We add one to firm-level R&D stocks and industry-level capital and R&D stocks before we take the log of these variables.

Download this Paper [ PDF 194.5KB| 27 pages ].

Comment(s)

There are [0] comment(s) for this entry. Post a comment.