Measuring capital services by energy use: an empirical comparative study

From an engineering perspective, the service that a capital good provides is energy conversion – that is, the physical ‘work’ done by a machine. A capital good’s service can thus be measured directly by the energy consumed in production. We show important empirical advantages of this approximation over traditional measures. The empirical application reveals that this approach avoids a number of conceptual problems of the latter. Furthermore, this measure captures the utilization of the capital stock more accurately as it is more sensitive to fluctuations in economic activity. With a growth accounting exercise, it is shown that the differences between the new and the traditional measures are important for empirical work. Using the new measure yields significantly different results. Especially in times of global recession it provides higher and more feasible total factor productivity growth rates.