Technical Efficiency and Technical Change in Canadian Manufacturing Industries

Table 2

Estimation results.

	TFE-RSI	TFE	BC	BC-RSI
Constant^‡
ilgt
Number of observations	903	987	987	903
Log-likelihood	401.41	401.11	412.23	395.05
Wald	252.92	279.46	60731.7	53223.22
Prob. >	0.00	0.00	0.00	0.00

Constant terms are industry specific for the TFE model (not shown) and are not included in the BC models.
Significant at %, not significant otherwise.
In recognition of the existence of one parameter that is not statistically significant in the last two columns (the parameter ilgt is not statistically significant).
(i) TFE and TFE-RSI, respectively, stand for true random effect panel stochastic frontier without and with residual inclusion.
(ii) BC and BC-RSI, respectively, stand for Battese and Coelli time-decay panel stochastic frontier without and with residual inclusion.
(iii) The significance of the inefficiency term in the BC specifications is based on the inverse logit of gamma (ilgt), where is equal to . This is so because the optimization is parametrized in terms of the inverse logit of gamma given that must be between zero and one.