Is There a Real TFP Growth Measure for Malaysia´s Manufacturing Industries?

Abstract

I. Introduction Total factor productivity (TFP) growth is an important measure of potential output growth given the nature of the diminishing returns to input use in the long run. Thus, Malaysia in her drive to enjoy sustainable growth to raise its living standards is set on focusing on TFP growth as stated in Malaysia's Second Industrial Master Plan 1996-2005. In fact, the manufacturing sector which has increased its contribution to gross domestic product (GDP) output from 19.3 per cent in 1979 to 34.2 per cent in 1996 has been identified as a key growth engine in this transformation process. Hence, it is imperative and timely for an analysis on the productivity growth performance of this sector to be undertaken. This study adds to the existing empirical literature in three ways. First, previous studies on Malaysian manufacturing have only considered the nonfrontier measure using the divisia translog index approach. To date, using the nonfrontier approach, Tham (1996, 1997) and the Productivity Report 1999 provide evidence of declining TFP growth for the Malaysian manufacturing sector in the 1990s (see Table 3). (1) How would this result compare with the use of the frontier approach? Will the frontier models also provide low TFP growth measures? This is one of the issues addressed in this article. As for the earlier studies, the nonparametric technique adopted computes TFP growth as a residual since it measures anything and everything of output growth that is not accounted by input growth. More importantly, the translog index TFP growth measure ignores the concept of technical inefficiency (by unrealistically assuming that all industries are technically efficient) and inaccurately interprets technical change as TFP growth. Thus in this study, frontier measures are used to overcome these major drawbacks. In the productivity literature, TFP growth is shown to be composed of both technical change (frontier shift) and technical efficiency (catching up effect). While the frontier effect indicates how far the efficient frontier itself has shifted over time due to the use of better technology and equipment, the catching up effect reflects how far the industry has moved towards the efficient frontier due to the better use of technology and equipment. The second difference in this study is that empirical robustness is ensured by the use of both the parametric and nonparametric frontier approaches to calculate TFP growth. Under the parametric approach, a stochastic production frontier model incorporating non-parallel shifts is estimated. With the nonparametric approach, the data envelope analysis (DEA) technique is used. Using a panel data set of twenty-eight manufacturing industries (see Appendix 1 for a list) from 1981 to 1996, a measure of TFP growth is first obtained and then decomposed to technical change and change in technical efficiency for both models. The results are then compared to previous studies with a focus on the Malaysian manufacturing sector as TFP growth studies on the aggregate economy may have broad implications that are not necessarily reflective of the TFP growth performance of specific sectors in the economy. The third contribution of this article is that the comparative performance of the results from alternative methodologies would add to similar work by Bjurek and Hjalmarsson (1990), Coelli and Perelman (1999), and Kumbhakar, Heshmati, and Hjalmarsson (1999) which provide mixed evidence of similarities in the results from the use of various models. Often, the choice of the method is said to depend on a range of factors. For instance, if the researcher simply wants to know if output growth is TFP or input-driven growth, then either approach would suffice. However, to answer questions on maximum productive or best practice output levels, the stochastic frontier can be used to understand the industries' catching up behaviour with respect to its own maximum potential, while DEA allows for the study of the performance of each industry relative to efficient industries in the sample. …