Cointegration Analysis of the Aggregate Production Function through Autoregressive Distributive Lags Models (ARDL)

This paper tests the validity of two-factors and augmented aggregate production function, applying bounds testing (ARDL) approach to cointegration, which is more appropriate for estimation in small sample studies, using different combinations of variables that measure the production result, capital and labor. In the augmented production function, along with “conventional imputs” of labour and capital, are included “unconventional imputs” like real gross foreign investment direct flows and impact of trade openness variable (sum of export and import values to production ratio). Based on quarterly data for Bulgaria, covering the period from first quarter of 1996 to third quarter of 2008, long-run relationships are investigated using the bounds testing cointegaration procedure, developed by Pesaran, and short-run dynamic parameters are obtained by estimating an error correction model (VECM). The results show that production, labor and capital are cointegrated in 4 combinations without trend (three of them describing the augmented production function) and in 6 combinations with linear trend (two of them describing the augmented production function). All the valid combinations indicated that ?lasticity of capital is between 0.001% and 0.286%, while elasticity of labor is between 0.316% and 0.533%.