Improving the estimation of total factor productivity growth: capital operating time in a latent variable approach

Measuring labor and capital services accurately is essential to obtaining reliable estimates of production functions and total factor productivity (TFP) growth. Using data on the operating time of capital, a series that exists for the French business sector, greatly improves the measurement of effective capital services in production. The ensuing estimation results are consistent with Cobb–Douglas technology under constant returns to scale, with the factor elasticities not statistically different from their income shares. In the same framework, TFP growth is estimated as a latent variable and found to be less volatile than accounting residuals, negatively correlated with employment, and free of cyclicality. It is statistically best estimated as a first-order autoregressive process, with an autoregressive coefficient of 0.95. Total factor productivity growth was estimated to have declined steadily between the mid-1970s and mid-1990s, but the rate of decline has diminished since then.