Economic reforms and state enterprise productivity in China: An application of robust estimation and latent variable measurement methods

The thesis estimates the effect of economic reform on total factor productivity (TFP) in Chinese state enterprises in the 1980s. The analysis is based on econometric estimation of the production function, using a panel dataset of 765 state industrial firms in the years 1980-89. Annual TFP growth is found to range from 2% to 8% in nine sectors of state industry, or 4% on average across industries. Qualitative variables from a survey of firm managers are used to produce reform indexes that measure the extent of reform at the firm level. The reform indexes are derived from estimates of a latent variable measurement model. An overall reform index is estimated to have a positive impact on TFP in eight out of nine sectors. The principal finding of the study is that the most reformed firms achieved TFP growth 12 percentage points greater than the least reformed firms. Since data quality is a serious concern, robust and resistant estimation methods are surveyed and then implemented. Empirical conclusions are based on robust estimates. Of particular methodological interest is the analysis of qualitative reform variables in terms of a statistical latent variable measurement model. The derived reform index provides a quantified measure for the assumed underlying qualitative characteristic of the firm. The latent variable measurement technique may be applied in a wide range of situations where a summary index of an individual characteristic is of interest.