Could Mexico become the new 'China'? Policy drivers of competitiveness and productivity.

• Author: Dougherty, Sean M.
• Position: Report

1. Introduction

   In recent years, shifts in global competitive conditions have caused China to lose competitiveness in some of its dominant export sectors. This has allowed Mexico's unit labour costs to become increasingly competitive with those of China. China has experienced a yearly growth of more than five per cent in unit labour costs, while Mexico's costs have increased at only half that rate. Mexico's catch-up in unit labour costs emerged primarily from a slowdown in China's productivity gains as its workers' wages grew rapidly, while at the same time the RMB appreciated against the USD (OECD, 2013, 2015a).

   This change in costs boosted Mexico's trade competitiveness, particularly in the manufacturing sector, where China's average wage now exceeds Mexico's (Sirkin et al., 2014). In addition, total landed costs for the US market, which include taxes, tariffs and regulatory compliance, as well as transportation and storage, have considerably increased for products made in China since 2005, while they have fallen for products made in Mexico (AlixPartners, 2013; Wang and Hu, 2014). As a consequence, there are increasing incentives for manufacturers to shift parts of their production process from China to Mexico, particularly in light of the proximity to final goods markets in North America. (2)

   Mexico's increasing competitiveness and attractiveness masks, however, one of the countries' fundamental concerns, which is the absence of productivity improvements. Mexico's productivity lags behind that of other major emerging economies, and it has suffered from a negative growth trend. One prominent feature of the Mexican economy as compared with China's is much more extensive employment informality and smaller average firm size (Dougherty, 2015; OECD, 2015a, 2015b). In order to fully take advantage of the increasing cost of production in China, identifying policies to improve productivity is essential for Mexico.

   Firms differ in productivity within even narrowly defined industries in a country. For example, in US manufacturing, the productivity of the 90th percentile plant is almost twice that of the 10th percentile plant (Syverson, 2004). The gap in productivity between high and low productive plants is five to six times larger in Mexico than in the United States (Hsieh and Klenow, 2014). These differences may indicate misallocation of resources across firms with negative effects at the aggregate level (Bartelsman et al., 2013; Hsieh and Klenow, 2009). Differences in productivity across countries can thus be explained by cross-country variation in the distribution of firm productivity.

   Multiple factors influence a firm's productivity, both internal and external to the firm. Among the internal factors, better management practices are associated with productivity gains (Bandiera et al., 2009; Bloom and Van Reenen, 2007). In addition to management quality, the quality of labour and capital influences productivity. Productivity is increasing in workers' education and age (Ilmakunnas et al., 2004), but differences in labour quality across firms only explain a small part of productivity dispersion (Fox and Smeets, 2011). Differences in capital quality are difficult to assess, and therefore, some studies have focused on information technology (IT) capital. IT productivity gains contributed to the acceleration of US productivity growth in the mid-1990s, in particular for IT-intensive industries (Bloom et al., 2012). There is also evidence that product innovation and intangible capital leads to productivity gains (OECD, 2015c). Indeed, the number of products and patent grants are positively correlated with total factor productivity (TFP) (Balasubramanian and Sivadasan, 2011; Bernard et al., 2010).

   In this paper, we focus on the external or contextual factors that influence productivity because they are more related to policy design. Even if contextual factors do not operate directly on productivity, they may influence producers' incentives based on internal factors and, thus, the productivity distribution across firms (Syverson, 2011). External factors influence the productivity of individual firms, but they can also influence aggregate productivity if more efficient firms grow faster than less efficient ones.

   Among the external factors, the literature highlights the importance of geography and foreign direct investment (FDI) since they influence technology and knowledge transfers (Bloom et al., 2013; Keller and Yeaple, 2009; Ciccone and Hall, 1996). Market regulation and competition are other important external factors that influence productivity. Competition increases the market share of more efficient firms, reducing that of less efficient firms and sometimes forcing them to exit (Melitz, 2003). In addition, competition may influence productivity through innovation; however this effect may follow an inverted U-shaped relationship (Aghion et al., 2005). Trade liberalisation is also a source of competition that fosters productivity growth through factor reallocation (Bloom et al., 2011); moreover, trade facilitates access to overseas' knowledge through the imports of intermediate inputs and supply networks (Goldberg et al., 2010). Finally, financial frictions reduce productivity because they hamper firms' investment and technology adoption decisions, as well as generate capital misallocation (Midrigan and Xu, 2014).

   One form of misallocation is informality, which can distort market competition. Informality is a symptom of poor institutional quality such as a burdensome regulatory framework and weak monitoring or enforcement power of the state (La Porta and Schleifer, 2014). Moreover, informal firms avoid taxes and benefit from low hiring and firing costs, allowing them to produce more cheaply than formal firms that face more regulations (Gonzalez and Lamanna, 2007). Second, informality may create labour market distortions: since formal labour is subject to regulatory and tax burdens that generate monetary costs for firms, the marginal cost of a worker increases with a firm's size (Busso et al., 2012; Levy, 2008). Thus, while large firms mostly hire workers legally and are taxed, smaller firms tend to hire less-skilled workers in the informal sector, limiting their productivity. We view informality as an intermediate outcome that may be subject to intervention using a variety of policy tools (Dougherty and Escobar, 2013).

   The aim of this paper is two-fold. First, motivated by the inversion of the unit labour cost differential between China and Mexico, we examine the growth and distribution of total factor productivity at the firm level, to better understand the extent to which inefficiency and misallocation are determining outcomes. Second, the study takes advantage of Mexico's heterogeneity across regions and sectors in terms of productivity, market regulation, and other constraints to identify economic policies that can help to boost productivity in the future. Among various findings, the results imply a strongly negative relationship between informality and productivity, which we investigate further to identify causality. More productive states and industries are found to suffer more from informality than less productive ones, and the negative effects of informality on productivity rise as the level of productivity increases.

2. Productivity patterns in Mexico and China

   2.1 The data

   Chinese and Mexican microdata are used in this study to measure productivity. For Mexico, plant-level data from the Annual Survey of Industries (EIA) and the Annual Survey of Manufacturing Industry (EAIM)--both conducted by Mexico's Institute of Statistics and Geography (INEGI)--were used remotely with INEGI's support. Although the data is plant-level, they can be considered as effectively firm-level because more than 97 per cent of Mexico's firms are single-plant firms (Dougherty, 2014). Since the EIA evolved to become the EAIM, we use EIA data for years 2005-2008 and EAIM data for years 2009-2012. For most industries, the sample is representative of the industry. INEGI selects plants according to their share in an industry's output until they obtain at least 80 per cent of the industry's total. In the cases where a small group of plants covers an industry's output, all of industry's plants are in the sample. In addition, all plants with more than 250 workers are sampled with certainty. Hence, we can expect that the smaller plants are generally excluded from the sample. A second limitation of these data is that we are unable to build a plant-level panel due to lack of plant identifiers. However, the data provide information about plants' location that allows us to match them with state-level policy measures.

   An important difference between EIA and EAIM is the shift in NAICS code classification. EAI uses NAICS 2002 version and EAIM the 2007 version. There are no major changes when considering 4-digit level data. There are 16 minor 6-digit industries for which changes in the NAICS version affect the 4-digit industry, which we then exclude from the sample.

   For China, manufacturing microdata from the industrial firm database of the Chinese National Bureau of Statistics (NBS) is used, from the data provider GTA. These longitudinal data cover the 2000 to 2007 period and include non-state firms with annual sales in current yuan of five million or higher, and all state-owned firms. These data are widely considered to be the best available company data for China during the period (Dougherty et al., 2007; Hseih and Klenow, 2009; Brandt et al., 2014). During the economic census year 2004, about 97% of firms were single-plant, similar to the Mexican data. While the dataset covers only about 20% of firms, these produce over 90% of output. The raw number of firms varies from 160,000 in 2000 to 335,000 in 2007. As a result of exit and entry to the database, about 80% of the firms in a given year can be observed in the previous year. In...