An empirical analysis of Total Factor Productivity (TFP) in Mexico, 1993-2018: determinants, decomposition and convergence

Abstract

This research estimates TFP at the establishment level in Mexico at 5-year intervals from 1993 to 2018. The production function used to estimate TFP includes a mark-up correction to overcome the omitted price bias, and the Wooldridge (2009) approach overcomes the endogeneity bias. The TFP analysis is divided into three categories: determinants, decomposition and convergence. The first category of the research identified six TFP determinants at the establishment level. The main findings are: (i) firm age positively impacts TFP, indicating a learning-by-doing effect; (ii) managerial and organisational efforts to reduce costs increase TFP; (iii) higher industrial concentration positively impacts TFP, suggesting that higher competition does not necessarily lead to higher TFP (i.e. Schumpeterian models); (iv) MAR externalities positively impact TFP due to localised economies of scale; (v) Jacobian externalities negatively affect TFP as a result of potential urbanisation costs; (vi) population density negatively affects TFP due to potential congestion costs in most sectors. The second category of the research analysed TFP decomposition by geographical locations, economic sectors and the contribution of firm selection to TFP growth. The geographical dimension of TFP indicates three clusters of states with high TFP: (i) some states in the North of Mexico, with potential productivity spillovers due to proximity with the U.S; (ii) states with high agglomeration economies, including Mexico City, Jalisco and contiguous states; and (iii) states in the Southeast, including Campeche and Tabasco, which are states mainly dedicated to oil extraction (i.e., natural advantages). The sectoral dimension of TFP accounts can be classified in three sectors with high TFP: (i) mining, quarrying, and oil and gas extraction, (ii) wholesale and retail trade, (iii) finance and insurance. Overall, services and oil extraction had a high TFP during 2018 while manufacturing activities had a low TFP performance. The Mexican economy had a negligible TFP growth of 0.10% from 1998 to 2018, and there was calculated the TFP growth decomposition regarding the contribution of surviving, entering and exiting establishments. The TFP growth decomposition shows a positive contribution of net entrants but a negative contribution of survivors. The condition of survival has deteriorated its contribution to TFP growth since the financial crisis (2008-2009). Conversely, business creation has been a driver of TFP growth. The third category of the research analysed TFP convergence across Mexican states and municipalities. There is no evidence of TFP convergence across states due to a potential aggregation bias. However, there is evidence of TFP convergence at the municipality level, which is associated with a reduction of TFP disparities across municipalities. Mexican municipalities had an absolute convergence rate of 0.21% per annum (1998-2018), and it would take around 323 years to eliminate 50% of the TFP gap across municipalities (i.e. half-life period). This thesis concludes with recommendations for an industrial strategy to increase TFP in Mexico.