ABSTRACT
The technological progress of an economy is considered to be the detrimental factor for growth in the medium or long run. However, the contribution of such a factor to growth cannot be estimated easily. The residual growth is usually considered to be the best proxy for total factor productivity growth (TFPG), which represents the level of technological shift of an economy for a given period. Various techniques and methods have evolved during the last four decades for measuring it and they are built on parametric, semi-parametric and non-parametric approaches. Whatever may be the approach, neo-classical production has been the basis of all. According to the parametric and semi-parametric approaches, it conventionally uses either factor share (Solow, 1957) or factor elasticity (Olley and Pakes, 1996; Levinsohn and Petrin, 2003) in order to subtract the factor contribution from the output growth. However, such a simple calculation often ignores market imperfections and, therefore, misleads the true level of productivity growth. For example, the product market power simply inflates the residue without changing the technology frontier. Similarly, when union workers receive wage rent, the residue will essentially be lower simply because of labour market imperfections. Even the non-parametric approach built on the linear and non-linear programming techniques using real values cannot avoid this problem. Therefore, the simple calculation of residual change cannot provide the actual productivity growth unless one eliminates the effects of market imperfections from the residual growth.
