Total factor productivity and firm size in the German energy industry: evidence from a symbolic transfer entropy panel causality test

This paper applied a symbolic transfer entropy panel causality test, to investigate the non-parametric Granger-type causality between total factor productivity (TFP) and firm size. This test overcomes the limitations of previous panel causality tests generated by the presence of outliers and structural breaks. We drew on a dataset of 535 German electricity and gas firms, analysing the period 2007–2015, and used different approaches to compute the TFP. The symbolic transfer entropy test indicated, in general, a bidirectional causality between TFP to firm size, when the first and the second lags in the causal relationship were used. However, the causality was not significant at the level of the third lag. Moreover, the way TFP was computed categorically influenced the results. Further, the non-parametric symbolic transfer entropy test brought clarifications to the opposite results reported by classic panel causality tests. Our findings show that increasing firm’s productivity represents a precondition for firm’s growth. At the same time, large firms in the energy industry are more efficient compared with small firms.