trends. The sample in this step is comprised of about 3,900 firms, which nonetheless cover
almost all industries and locations.
The distribution of estimated time trends is presented in Table 9. While about 10 percent of
the best performing firms improved their efficiency about 1.3 percent per year, the average (or
median) firm experienced a slight efficiency decline. For the 10 percent of the worst performing
firms, efficiency decreased by an annual rate of about 2 percent. This finding serves as an
additional argument in support of Hypothesis 5.
In the last step of the empirical analysis, we explore the determinants for the positive or
negative firm-specific time trends in efficiency. We regress the parameter estimates θi as in the
previous analyses on the same set of explanatory variables. The partial R2s are reported in Ta-
ble 10 and the parameter estimates (selected variables) are displayed in Table 11. The picture
that emerges from this analysis of firm-specific efficiency trends is in line with the former re-
sults: the overwhelming part of the variation in efficiency trends is explained by industry and
location. Other environmental or firm-specific factors have only minor impact.
The estimates presented in Table 11 suggest that, first, a change in efficiency is independent
of the size of the firm. Second, two factors determine the development of efficiency: the indus-
try in which the firm is operating and its location. Third, only two of the outsourcing activities
have a significant impact: quota of external contract work (positive sign) and quota of exter-
nal services (negative sign). However, these effects appear to offset one another. One further
remarkable contrast to the analysis for the level is that R&D has now a positive impact on the
development of efficiency, albeit with extremely low explanatory power. We infer from these
findings that there is an inverse relationship between R&D and the level of efficiency, but that
firms with a higher R&D intensity tend to improve their efficiency over time.
6 Conclusions
This paper analyzed the importance of a variety of factors to the productive efficiency of firms,
with particular emphasis on industry, location, R&D, and size. In a first step, we obtained
estimates from a translog production frontier and then, in a second step, performed analysis
of covariance to investigate the determinants for firm-specific productive efficiency and its dy-
namics. We employed the concept of partial R2 to gauge the relative importance of the various
factors.
The translog production function estimates for firms covering the entire manufacturing sec-
tor are in accordance with predictions from neoclassical theory for competitive product and
factor markets, that is, the average firm operates with constant returns to scale technology. Sec-
ond, industry affiliation is the most important factor, having the largest share in the model’s
explanatory power. This holds both for the level and the development of efficiency. Third, size
22
More intriguing information
1. Second Order Filter Distribution Approximations for Financial Time Series with Extreme Outlier2. The name is absent
3. The name is absent
4. The name is absent
5. Declining Discount Rates: Evidence from the UK
6. The name is absent
7. Exchange Rate Uncertainty and Trade Growth - A Comparison of Linear and Nonlinear (Forecasting) Models
8. The name is absent
9. The name is absent
10. THE ECONOMICS OF COMPETITION IN HEALTH INSURANCE- THE IRISH CASE STUDY.