A minha tese corroborada

"Four Patterns of Plant Dynamics
.
The distribution of productivity among plants may arise in four ways.

First, it may be the result of a random draw in the level of productivity in each period or of errors in measurement. Second, it may be the result of a random draw in the growth of productivity rather than in the level. Third, it may be the result of plants of different vintages.
Fourth, it may simply reflect permanent plant heterogeneity. We ask what these possibilities would imply about plant dynamics and the way in which the distribution evolves over time."
...
"We can now look back to the four alternative patterns that we discussed at the beginning of the paper and presented in figure 1. Which one (or more) has the data supported? The answer is fairly clear. The overall pattern of the data is best described as a combination of the random shock/measurement error case (figure IA) and the plant fixed effects case (figure ID). The regression toward the mean and many other signs show the importance of the random shocks. The strong persistence, visible most dramatically in the 10-year transition matrix, supports the plant fixed effects framework.
...
We have spent enough time studying the individual observations to realize that true productivity in these plants is not known. And there will be important plant-specific shocks thatw ill cause even an accurately measured productivity measure to move around. The more interesting finding is that there is strong persistence in relative productivity. The results on wages and productivity suggest that differences in worker quality may not be the main reason for the persistence of relative productivity.
.
What appears to be important is management quality, broadly interpreted to include technology choice and product choice. (Moi ici: Esta é a minha tese, a escolha dos clientes-alvo a servir e dos produtos e serviços a oferecer é, apoiando-me em Marn e Rosiello, o ponto mais importante para garantir uma elevada produtividade assente na produção de mais valor acrescentado, mais valor originado e não em menos custos e mais eficiência) Of course, this conclusion is tentative since we lack direct evidence on management
quality."
.
Trechos retirados de "Productivity Dynamics in Manufacturing Plants" de Martin Neil Baily, Charles Hulten, David Campbell, Timothy Bresnahan e Richard E. Caves, publicado por Brookings Papers on Economic Activity. Microeconomics, Vol. 1992 (1992), pp. 187-267.