Structural Transformation in Mexico: What is the Role of FDI?

Transcription

1 Structural Transformation in Mexico: What is the Role of FDI? Octavio Escobar Paris School of Business Henning Mühlen University of Hohenheim This version: July 28, 2017 Preliminary version: please do not cite Abstract Foreign direct investment (FDI) flows to Mexico are substantial and play an important role in the Mexican economy in the recent past. This is accompanied by the fact that these investments shape to some extent the economic landscape and structure in this host country. Yet, there is considerable variation in the amounts of FDI within the country. Hence, FDI has different impacts on the economic structure across Mexican states. We provide descriptive figures of FDI and structural change indicators that highlight these issues in many dimensions. Based on these observations, the main aim of this paper is to analyze the impact of FDI on structural transformation. We conduct an empirical analysis on the regional level. The unit of observation is a Mexican State for which we calculate structural change from the reallocation of labor between sectors. The results suggest that there is no effect from aggregated FDI. Considering disaggregated FDI flows, the findings indicate that FDI from the US seem to have negative effects on effective structural change while Spanish FDI has a positive impact. Moreover, there is some evidence that FDI flows in agriculture and services have a positive effect. JEL-Classification: F21, F63, L16, O10, O54 Keywords: FDI, structural transformation, labor reallocation, Mexico, within country analysis, economic development Corresponding author: Henning Mühlen, University of Hohenheim, Stuttgart, Germany, LHenning.Muehlen@uni-hohenheim.de

2 1 Introduction The activities of multinational firms (MFs) have largely increased worldwide in the recent past associated with considerable foreign direct investment (FDI) inflows in many destination countries. The general backgrounds of these investments may be very different and, thus, the resulting picture may look quite heterogeneous across and within FDI host economies. In some countries the FDI projects may be spatially distributed across the country and associated with activities in the manufacturing sector in the form of numerous affiliates that are integrated in global value chains and produce intermediates, while in other countries there may be mainly large single direct investments in the resource sector which are located in a few places. Whatever the differences in the motives and investment decisions are, there is a natural common aspect: FDI shapes to some extent the economic landscape and structure of the host countries. This calls for the question whether FDI inflows have a significant impact on structural transformation the process of a reallocation of economic activity between sectors. 1 The direct relationship between FDI and structural change has been largely ignored in the literature so far, although it is likely to be of high relevance in many economies. Entering an economy and being active in a relatively productive sector, MFs may pull economic activity away from low-productive sectors and, hence, contribute to effective structural transformation. In this context, MFs may induce a labor reallocation between sectors. Considering that MFs are typically highproductive firms that tend to pay higher wages than there local competitors (Mayer and Ottaviano, 2007), there is a large potential that they attract labor to move. Given that this movement occurs between sectors (at least to some extent), FDI contributes to structural transformation. Finally, the potential link between FDI and structural transformation is relevant from the development policy perspective since structural change has been identified as a key factor of economic development (see McMillan, Rodrik, and Verduzco-Gallo, 2014; Timmer, de Vries, and de Vries, 2015). Hence, if FDI affects structural transformation positively, it potentially contributes to economic development. In this paper we conduct a within country analysis and empirically investigate the impact of FDI inflows on structural transformation in Mexico. More specifically, we run regressions based on state-level data between 1998 and We calculate structural transformation from the reallocation of the factor labor across different sectors at the state-level. Additional to our variable of interest (FDI inflows), we 1 We will use the terms structural transformation and structural change interchangeably in this paper. 1

3 include control variables that account for further relevant factors potentially explaining the within-country variation of structural change (e.g. labor market conditions). Mexico appears to be an appropriate country case for the given research objective for a couple of reasons. First, Mexico is a large country comprising 32 federative entities (states). Second, the country received considerable FDI inflows from various source countries during that time. In this respect, we are able to identify the amount of FDI inflows to a Mexican state by each source country. Our following descriptive results show that large amounts of FDI inflows in numerous Mexican states come from the US, Canada, and/or Spain. Third, all Mexican states received FDI. However, there is a high variation in the amounts of FDI across states, sectors, and time. Finally, there is notable structural change with considerable differences between and within Mexican states. The paper s main contribution is characterized as follows. First, we investigate the direct link between FDI and structural change which is widely neglected in the literature so far. There are only a few studies that address this economic relationship (e.g. Amendolagine, Coniglio, and Seric, 2016). Second, the majority of studies examining structural transformation issues focuses on economy-wide structural change in the long-run. The respective investigations conduct mainly cross-country analyses and highlight the role structural change for economy-wide productivity growth (e.g. Duarte and Restuccia, 2010; Timmer, de Vries, and de Vries, 2015). We take a different perspective and focus on within-country variations in the short-run. Moreover, our goal is to examine the effects on structural change. One attempt of investigating the determinants of structural change in a cross-country setting can be found in McMillan, Rodrik, and Verduzco-Gallo (2014). Third, the paper at hand contributes to the understanding of the consequences of FDI considering to what extent it influences the sectoral patterns in a host country. There is a large associated literature that examines the effects of FDI. Yet, the particular focus is different to the one in our work. For example, one strand is interested in the effects of FDI on economic growth (e.g. Borensztein, De Gregorio, and Lee, 1998), while another main strand covers FDI spillovers (frequently investigating the effects on productivity levels using firm-level data (e.g. Görg and Strobl, 2005)). Our empirical results suggest that there is no significant effect of aggregated FDI on structural change within Mexican states. The picture appears to be more diverse. FDI inflows from the US seem to have negative effects on effective structural change while Spanish FDI has a positive impact. Moreover, we find some evidence that FDI flows in agriculture and services have a positive effect. The structure of this paper is as follows. In Section 2 we present a detailed 2

4 portrait of FDI inflows and sectoral patterns in Mexico based on descriptive figures. This is followed by Section 3 where we explain the empirical framework to analyze the relationship between FDI and structural transformation. The related results of the regression analysis are displayed and explained in Section 4. Finally, we provide concluding statements based on the empirical findings in Section 5. 2 FDI and sectoral patterns in Mexico First, we show long-run paths of structural change indicators for the Mexican economy as a whole using data from the GGDC 10-Sector Database (Timmer, de Vries, and de Vries, 2015). Based on these observations, we are able to extract some relevant insights about the development process of the economy. Moreover, we show inward FDI numbers for Mexico using data from the World Bank (2017) and place our research question in the long-run context. Second, we change to a sub-national perspective and illustrate sectoral patterns as well as FDI inflows related to the Mexican states. This view highlights substantial differences within Mexico. In this regard, we make use of data from the Instituto Nacional de Estadítica y Geografía (INEGI; These data are available on a highly disaggregated sectoral level 2 and offer valuable information. Finally, these descriptive figures are a motivation for the following regression analysis. 2.1 Economy-wide structural change and FDI inflows in the long-run In Figure 1 there are undeniable signs for long-run structural change. Most notably, the agricultural employment share decreases significantly over time. Starting in 1950 with almost 60 percent, the share of agricultural employment relative to total employment goes down to about 14 percent in Contrary to this observation, the share of services increases constantly over time and is dominant since the 1970s. The subsector wholesale and retail trade, hotels and restaurants makes up the largest fraction in services and grows from ten percent to 22 percent (not reported in the figure). Interestingly, the share of industry sectors (including manufacturing) first increases until 1980, but then remains roughly constant over time ranging between 25 and 30 percent. Concluding, we note that Mexico appears to be a relatively services oriented economy in terms of sectoral employment shares since the 1980s. 2 We use the 3-digit sector level. 3

5 Figure 1: Long-run path of sectoral employment shares in Mexico Sectoral employment shares, Agriculture Industry Services Note: Data points after 2012 were not available at the time of data extraction. Source: GGDC 10-Sector Database Overall, the long-run picture mirrors to some extent the rationale of dual economy approaches (e.g. Lewis, 1954) where labor moves from a traditional sector (agriculture) to modern sectors (such as manufacturing or services). In these approaches the traditional sector is typically characterized by low productivity levels relative to the modern sectors. Especially, in early stages of a country s development process these productivity gaps between sectors can be observed in the data. Given that in this setting labor moves out of the traditional sector towards modern sectors, it is likely to be employed more effectively on average. Such a reallocation of economic activity reflects positive structural change. However, this story only holds as long as productivity levels remain lower in agriculture than in non-agricultural sectors. Moreover, the marginal effect of positive structural change is supposed to diminish over time. With ongoing labor reallocation towards non-agriculture sectors, the labor productivity in agriculture should increase and, other things equal, productivity gaps should tend to decrease in the long-run. Actually, such productivity gaps between sectors are observable for Mexico. Although employment in agriculture between 1950 and 1970 is considerably high, value added in absolute and relative terms is very low. In the time period the share of agricultural value added ranges between five and ten percent only. Hence, labor productivity levels (in terms of value added per employee) in agriculture are relatively low. Figure 2 illustrates labor productivity levels for the three broad sectors over time. Obviously, there are large productivity gaps between the tra- 4

6 Figure 2: Long-run path of sectoral labor productivity levels Sectoral value added (in pesos) per employee, Agriculture Industry Services Source: GGDC 10-Sector Database ditional sector and the two modern sectors. Consequently, a labor reallocation from agriculture towards industry and/or services should lead to positive structural change in the economy since workers are then more effectively employed on average. Moreover, the figure indicates a slight tendency for convergence of productivity levels visible since the 1980s. Next, we briefly characterize Mexico as a destination country for FDI. Net FDI inflows to Mexico relative to Mexican GDP are depicted in Figure 3 for the period FDI became increasingly important over time. Actually, we can observe three periods with different levels of the FDI-GDP ratio. The first is between 1970 and 1978 when the ratio of FDI to GDP was always lower than one percent. Then, there is a little jump to ratios around one percent in period two ( ). This is followed by a further jump to FDI-GDP ratios of around 2.5 percent (with a peak in 2001 of more than four percent) between 1993 and 2015 (period three). Hence, in the recent time period which is the period of our following analysis FDI inflows play persistently a role in the Mexican economy. It is worth noting that the numbers of the FDI-GDP ratio in period three are comparable with those of the world ratio in that time. Finally, we can draw an important conclusion from this subsection and put our analysis in a long-run perspective. Considering the role of the traditional sector agriculture over time, the Mexican economy has already experienced a typical first step of structural transformation between 1950 and the mid 1980s. As illustrated in 3 Note that there is no data on FDI available before

7 Figure 3: Long-run path of FDI inflows FDI net inflows in % of GDP, Source: World Bank (2017) Figure 1, the employment share of agriculture has substantially decreased indicating that economic activity has been shifted away from (low-productive) agriculture towards the modern sectors. 4 Since the time period of our following analysis is 1998 to 2013, we investigate the relevance of FDI in a setting where structural transformation should be largely characterized by a reallocation of economic activity between and within modern sectors. 2.2 Within country perspective: FDI and structural change at the state-level Mexico is divided into 32 federative entities, 31 States and the capital Mexico City. Since the level of autonomy is similar for the 32 federative entities, for simplicity, we refer to the 32 Mexican states hereafter. Each state is free and sovereign, and has its own congress and constitution (SCJN, 2010). During the 1990s, Mexico s authorities have been working on major structural reforms. First, Mexico undertook important trade liberalization efforts, which transformed the country from a relatively closed economy to one of the most open in the world. The most important structural reforms included a government-wide deregulation program, a new competition law, the elimination of price controls, and the negotiation of free trade agreements (Gurría, 2000). In 1993 Mexico was the first 4 A similar picture can be observed in many developing countries in this period. See Timmer, de Vries, and de Vries (2015) for an overview. 6

8 developing country to sign a free trade agreement with developed countries concluding the North American Free Trade Agreement (NAFTA) with Canada and the United States (Escobar Gamboa, 2013). In addition, the authorities eliminated important restrictions on capital inflows, such as limitations on commercial borrowing from abroad, foreign investment in Mexican securities, and foreign participation in domestic money markets (Gurría, 2000). Second, Mexico s authorities implemented a far-reaching fiscal reform to embrace fiscal discipline. Finally, in 1994, the Central Bank of Mexico (Banxico) became independent, and at the end of that year the exchange rate regime passes from xchange rate bands with managed slippage to free float (Carstens and Werner, 2000). During the legislative periods between 2000 and 2012, the authorities failed to pass any economic structural reform. However, the process of structural reforms accelerated significantly since 2012 with the Pact for Mexico (OECD, 2015). Indeed, an important set of structural reforms including energy, fiscal, financial, telecommunications, competition, education, and labor reforms were approved by the Mexican Congress between 2012 and Hence, in this paper we focus on a period of legislative stagnation in terms structural reforms. This is important to note, since most of the reforms in Mexico are also oriented to improve Mexico s FDI attractiveness (Escobar Gamboa, 2013) FDI inflows FDI data by state are available on an annual basis from 1999 onwards and presented in Table 1. Considering the first column (Mean ), it can be seen that there is a large variation among Mexican states as FDI destinations. The average annual inflows between 1999 and 2013 range from roughly 1.3 billion pesos in Campeche to more than 71 billion pesos in the District Federal or Mexico City. It is quite plausible that the country s capital is the number one destination of foreign investments since it appears to be the economic center with many affiliates of foreign MFs. Overall, the large deviations between the states are of particular interest with respect to the following picture of structural change that can be observed. Columns two to four shed some light on three sub-periods and prove that FDI plays a role in all states throughout the considered time horizon. Each column shows the sum of annual inflows of the respective five-year sub-period. There is a heterogeneous picture among the states. In states like Baja California FDI inflows decrease over time while in others such as Guanajuato the numbers are increasing. However, for many states there is no a clear trend visible. Taking into account the total numbers for the whole economy, it can be seen that total FDI inflows are 7

9 relatively stable over time. Finally, the last three columns refer to the sectoral structure and highlight the differences between agriculture (AGR), industry (IND), and services (SER). The most striking observation is that FDI in agriculture is negligible in all states. Although the AGR share is larger than two percent in Chiapas, Colima, Querétaro de Arteaga, and Sinaloa, the absolute amount of FDI inflows in these states is very low since total inflows are particularly small there. Hence, the other two sectors are dominant in this context. Again, the picture is very heterogeneous. In Guerrero and Hidalgo there is an almost balanced distribution between industry and services while in many other states FDI in one sector is dominant. Across Mexico, FDI in the industry sector is prevalent with 60.7 percent. The information in the data also allows for a differentiation between FDI home countries. This is highly important in the case of Mexico since there are a few major source countries. We illustrate this in Figure 4. More than two thirds of the direct investments between 1999 and 2013 stem from the US, Canada, Spain, Germany. Obviously, the US are the major economic partner as the share makes up percent. Due to various reasons such as the spatial proximity and the NAFTA, Mexico and the US have strong economic ties in terms of financial and trade relations. Baldwin and Lopez-Gonzalez (2015) show that a high share of bilateral trade between the two economies is related to trade in intermediates, or in other words, to supply-chain trade. In this regard, Mexico sources a significant share of intermediates from the US (to produce either final goods or processed intermediates for exporting). 5 These trade patterns are evidence for US MFs activity in Mexico since a considerable share of such trade flows is linked to FDI with horizontal or vertical motives. Spain with its historical relations to Mexico is also a major home country of FDI (12.6 percent) whereas Germany and Canada are of minor importance in comparison. Tracing these bilateral connections, we are able to investigate in the regression analysis whether there are structural change effects in response to FDI stemming from different source countries. 5 The flows of supply-chain trade between the US and Mexico are even significant from the global perspective (Baldwin and Lopez-Gonzalez, 2015, pp ). 8

10 Table 1: Annual FDI inflows by state, Mean Sum Sum Sum AGR IND SER State share share share (m.mxn) (m.mxn) (m.mxn) (m.mxn) (%) (%) (%) Aguascalientes Baja California Baja California Sur Campeche Chiapas Chihuahua Coahuila de Zaragoza Colima Federal District (Mexico City) Durango Guanajuato Guerrero Hidalgo Jalisco Michoacán de Ocampo Morelos México Nayarit Nuevo León Oaxaca Puebla Querétaro de Arteaga Quintana Roo San Luis Potosí Sinaloa Sonora Tabasco Tamaulipas Tlaxcala Veracruz de Ignacio de la Llave Yucatán Zacatecas Mexico: Total Notes: All absolute numbers are given in million pesos (constant values). The last three columns refer to sectoral FDI inflows relative to total FDI inflows ( ). These shares are given in percent and refer to agriculture (AGR), industry (IND), and services (SER), respectively. 9

11 Figure 4: FDI inflows by origin (mean ) 5.469% 47.87% 31.8% 12.6% Canada Germany Other Spain United States Source: INEGI Structural change We follow McMillan, Rodrik, and Verduzco-Gallo (2014) and calculate structural transformation from the reallocation of labor between sectors: ST st = n y jst (Θ jst Θ jst 1 ) (1) j=1 In Equation 1 y jst represents the sectoral labor productivity level in sector j in state s in period t. Θ jst is the respective sectoral employment share. The difference between the employment share of t and t 1 indicates a change over time and, thus, a labor reallocation. A positive change signals a shift of economic activity towards the respective sector while a negative sign indicates the opposite. In this general expression, there are n sectors in the observed economic unit s. The number of n depends on the sectoral aggregation level. To compute equ. 1, we consider the North American Industry Classification System (NAICS) and use data at the 3-digit sector level for all Mexican states, covering the period More precisely, for each state and sector, we use value added and labor units data from the Economic Censuses (Censos Económicos) conducted every five years by the INEGI from

12 to Figure 5: Structural transformation by State Country AGS BCN BCS CAM COA COL CHS CHI CMX DGO GTO GRO HGO JAL MEX MIC MOR NAY NLN OAX PUE QRO QTR SLP SIN SON TAB TAM TLA VER YUC ZAC Structural transformation by state, Structural transformation Figure 5 illustrates the estimates of structural transformation for three different periods. At the country-level, Mexico experienced two consecutive periods of negative structural transformation, but this dynamic has reversed for the last period. That is, between 1998 and 2008 labor moved from more productive to less productive sectors, while between 2008 and 2013 labor moved towards more productive sectors on average. Figure 5 also illustrates the heterogeneity in structural change across 6 Note that Economic Censuses report data from the year preceding the year when it is conducted. 11

13 Mexico s states. Differences among states are related to both amount and dynamics. For instance, Baja California South (BCS), Colima (COL), and Chiapas (CHS) experienced a negative structural transformation over the whole period, but this was more important in COL. On the other hand, only San Luis Potosi (SLP) benefited of positive structural change over the period of study. Finally, Figure 5 illustrates that Campeche (CAM), Tabasco (TAB), and Tamaulipas (TAM) are those states with the largest structural change from one period to another. These three states have the particularity of hosting oil and gas deposits. Since the activities of extracting these resources are characterized by a labor productivity that is significantly larger compared to the other sectors, an increase or a decrease in the employment share in oil and gas extraction industries generates an important structural change. Based on the disaggregated data, we are able to calculate a sense of intrasectoral patterns. Figure 6 illustrates structural change within sectors for the country as a whole and for one individual state (Campeche). At the country-level (Figure 6-a), it is clear that the sector with the highest amount of structural change is the manufacturing, followed by financial services. Compared to the country picture, dynamics in Campeche are quite different. Figure 6-b shows that structural change in this state is dominated exclusively by the mining sector (for which oil and gas extraction industries are part). Moreover, the dynamics of this sector are different than those at the country-level. While in Campeche it experienced two periods of positive change followed by a downward trend over the last period, the country dynamics are the opposite. We may suspect oil and gas industries to be a source of potential bias, in particular for a regression analysis. 3 Methodology: Assessing the role of FDI for structural change Our main objective is to investigate the relationship between FDI and structural transformation in Mexico. In particular, we are interested in the issue whether FDI has an impact on structural change. To analyze this we run a regression analysis where the unit of observation is a Mexican state. In terms of the econometric framework, we apply Equation (2) for cross-state panel estimations: ST st = α + βf DI st 1 + γ X st 1 + λ t + µ s + ε st (2) where ST st represents structural transformation in (a Mexican) state s and period 12

14 Figure 6: Structural transformation within sectors Agriculture, hunting, forestry, and fishing Mining and quarrying Public utilities (electricity, gas, and water) Construction Manufacturing Wholesale and retail trade, hotels, and restaurants Transport, storage, and communications Finance, insurance, real estate, and business serv Community, social, personal, and gov services Other services (a) Country Country-level structural transformation within sectors -15,000-10,000-5, ,000 Structural transformation Agriculture, hunting, forestry, and fishing Mining and quarrying Public utilities (electricity, gas, and water) Construction Manufacturing Wholesale and retail trade, hotels, and restaurants Transport, storage, and communications Finance, insurance, real estate, and business serv Community, social, personal, and gov services Other services (b) Campeche Campeche Structural transformation t. It is calculated based upon the indicator introduced in Equation (1) and can be interpreted as a measure for effective structural transformation due to labor reallocation across sectors. In this context, our preferred indicator is computed from the most disaggregated data in terms of the sector level. This measurement strategy appears to be as accurate as possible. Due to data limitations, we can only consider four points of time for the calculation: 1998, 2003, 2008 and Hence, we are able to calculate three time periods of structural change: , and In Equation 2 the explanatory variable of interest is F DI st 1. It captures the presence of MFs aggregated activity in a state s at time t 1. We use different 13

15 measurements of F DI st 1 explained below. First, it is worth discussing the time structure of this variable. We argue that for the majority of FDI projects it takes some time until the direct investments actually lead to structural change in terms of a labor reallocation. As an illustration think of greenfield investments where it usually takes a considerable amount of time until labor is employed since potential places of work such as production facilities have to be build first. Moreover, additional investments in existing FDI projects providing further job opportunities may also attract labor to move. Again, it is plausible that there is a time gap between the point of time when the investment has been done and the point of time when workers are actually employed. Especially, if labor moves across sectors such time gaps are plausible as adjustment costs of a job change should be higher on average. Given that our dependent variable is calculated from changes between t 1 and t, the FDI indicator should be constructed of lagged values, that is, from years equal to and/or before t 1. We test two types of measurement in this regard. First, an indicator labeled fdi5yrs that is based on the sum of FDI inflows from the year t 1 and the four years before. 7 Second, we consider FDI inflows of the single year t 1 (which appears to be the initial year of the structural change process in our setting). The variable is labeled fdi. Associated with these two indicators, we take into account different subdivisions of FDI inflows. This is explained in detail in Section 4. In principle, FDI stocks rather than FDI flows should be applied as a proxy for the activity of MFs (Wacker, 2016). However, in the given case there is a striking argument to consider FDI inflows as an indicator in our analysis. Since the dependent variable (structural transformation) describes a change over time, it is reasonable to assume that its variation is driven by changes in the activity of MFs (which is measured by FDI flows) rather than by the level (which is associated with the FDI stock). X st 1 in Equation 2 refers to a vector of additional (time-variant and statespecific) control variables. All controls are included as lagged values taking into account the initial year of the structural transformation process (t 1). We consider an indicator that addresses the development level measured by GDP per employee. Additionally, we include net national migration (immigrants minus emmigrants per 100 inhabitants from/to other Mexican states) in order to control for the potential movement of labor between Mexican states. Moreover, we employ a variable of labor market conflicts that addresses labor market frictions and, hence, resulting rigidities of labor reallocation. And, we include the agricultural employment share 7 E.g., (t 1) = 2008: fdi5yrs s2008 is the sum of FDI inflows of the years

16 which represents the degree of structural change. States with a high share are supposed to be relatively backward in terms of structural transformation since a relatively high fraction of labor is bound in the traditional sector which is the case for some states. 8 Moreover, we control for period-specific effects λ t and state fixed-effects µ s. Finally, α is the constant and ε st is the error term. In terms of the estimation method, we employ the fixed effects estimator. Thus, we exploit the panel structure of the data and the within variation in particular. As FDI data is only available since 1999, our econometric strategy limits the T dimension to two periods. That is, we are able to consider structural change between and This time horizon appears to be relatively short when looking at the related literature investigating structural change. Yet, our perspective is quite different from many studies. One the one hand, considerable FDI inflows are a relatively young phenomenon in Mexico and, thus, potential effects are limited to the recent past. On the other hand, the potential effects are supposed to occur with a relatively short time lag. Moreover, we calculate structural change on the basis of quite disaggregated sectoral data. Hence, we take into account not only a shift of economic activity between agriculture, industry and services (as considered in many studies) but also within these sectors. These intra-sectoral reallocations are expected to play a role in the given context as FDI in a subsector (in a Mexican state) may be of great economic importance. Taking these aspects into account, the time dimension is appropriate from our point of view. The results of the regression analysis are presented in the following section. 4 Empirical results Table 2 reports our empirical results employing the five year FDI indicator (FDI5yrs). Hence, we analyze the effect of accumulated FDI which includes activities that are supposed to affect ST with a considerable time lag. Each column refers to a separate estimation including the same set of control variables but employing alternative measurements of FDI5yrs. Among the explanatory variables, only GDP per employee (gdppe) and the agricultural share of employment (AGRemp) appear to have a significant effect. The associated coefficients are significant at the 1 and 5 percent level, respectively, over the different specifications. Both variables have a positive effect on structural transformation. The effects of GDP per employee on structural change suggest that more developed states are more likely to experience positive structural 8 Definitions of all variables can be found in the Appendix A1. Moreover, descriptive statistics of all variables used in the regressions are displayed in the Appendix A2. 15

17 change. However, the result for the agricultural employment share variable suggests that those states with a larger share of labor in this sector in the initial year experience higher structural change. Since this variable is a proxy for initial structural gaps (McMillan, Rodrik, and Verduzco-Gallo, 2014), we can deduce that there is a process of convergence among Mexican states. Table 2: FE-Regressions: 5 year FDI indicator (1) (2) (3) (4) (5) (6) (7) gdppe 0.395*** 0.395*** 0.395*** 0.395*** 0.394*** 0.395*** 0.395*** (35.59) (34.98) (36.67) (36.99) (34.69) (36.28) (35.19) migration 24,669 23,123 24,730 24,927 22,596 24,314 21,853 (1.133) (0.967) (1.105) (1.048) (1.004) (1.033) (0.914) lab.conflicts * (-1.316) (-1.453) (-1.448) (-1.797) (-1.499) (-1.561) (-1.551) AGRemp 5,873** 5,393** 6,034** 6,004** 5,331** 5,819** 5,304** (2.683) (2.202) (2.520) (2.572) (2.319) (2.542) (2.290) fdi 5yrs (-1.084) fdi 5yrs ser (-0.843) fdi 5yrs ind (-0.524) fdi 5yrs agr (-0.335) fdi US 5yrs ** (-2.065) fdi CAN 5yrs (0.0861) fdi ES 5yrs 0.545* (1.888) Observations R-squared Number of state Periods Notes: ***, ** * indicate significance at the 1, 5 and 10 percent level, respectively. t-statistics obtained from robust standard errors are displayed in parentheses. The dependent variable is always structural transformation calculated from Equation 1. All regressions include year dummies not reported. Looking at the variable of interest in Column 1 of Table 2, we consider total FDI inflows aggregated over all sectors. We find a negative coefficient which is insignificant. However, it may be the case that there are positive effects from FDI in one sector (e.g. services) on structural transformation which are compensated by negative effects of FDI in another sector (e.g. agriculture) and, thus, the overall effect is not significant. To test this hypothesis, we employ alternative indicators of our FDI variable considering the sector of destination. That is, we include separately FDI inflows in services, industry, and agriculture. The coefficients presented in columns 2 to 4 of Table 2 are all negative, but not significant confirming the overall finding from column 1. Even if the coefficients of FDI per sector are insignificant, the differences in their size are remarkable, suggesting a sort of heterogeneity among FDI according to the sector of destination. 16

18 We extend our analysis to the high importance of single FDI home countries. In Section 2 we illustrate that a large fraction of FDI stems from the US, Canada and Spain. Hence, we investigate the effects on structural change with respect to these FDI source countries and include indicators measured based on FDI inflows from these economies. Columns 5 to 7 of Table 2 present the related estimates. Interestingly, FDI from the US has a negative and significant effect on structural change, while FDI from Spain has a positive and significant effect. We replicate our estimations in Table 3, but replacing the five-year indicator (fdi 5yrs) by FDI inflows in the initial year of the structural change process (fdi). Using this variable, we limit the analysis to effects with a shorter time gap only and test whether this makes a difference compared to the findings associated with fdi 5yrs. In column 1 the coefficient now is positive, but still insignificant approving the finding that aggregated FDI inflows do not have an impact on structural change. However, when analyzing differences according to the destination sector of FDI, the results show a significant heterogeneity among sectors. FDI in services and agriculture have a positive and significant effect on structural change, while FDI in industrial sector is found to be insignificant. Finally, the results of different origin countries illustrate that Spanish FDI has again a significant and positive effect on structural transformation, while the coefficients of FDI from USA and Canada are insignificant. 5 Concluding remarks We observe considerable FDI inflows in Mexico in the last two decades. Yet, there is a high variation across Mexican states and sectors indicating that the role of FDI is quite heterogeneous within the country. Moreover, we find that the US are the major source country of FDI. At the same time, there is significant structural transformation in the economy. Again, the picture is heterogeneous since structural change differs among the Mexican states. Taking into account these two phenomena, we analyze whether the aggregated activity of MFs in terms of FDI has an impact on structural transformation running fixed effects regressions at the state-level. The results suggest that there is no significant effect from aggregated FDI on structural change. However, a more disaggregated view on FDI reveals important findings. On the one hand, FDI inflows from the US seem to have negative effects on effective structural change. Since this result is related to the five year indicator, the negative effect is based on accumulated FDI including projects that affect ST with a considerable time lag. Considering that the US are the major FDI source country, 17

19 Table 3: FE-Regressions: Initial year FDI indicator (1) (2) (3) (4) (5) (6) (7) gdppe 0.395*** 0.395*** 0.394*** 0.394*** 0.395*** 0.395*** 0.395*** (35.63) (36.04) (36.46) (36.12) (36.70) (36.20) (35.33) migration 21,969 21,355 25,570 26,976 24,088 23,352 21,326 (0.887) (0.929) (1.197) (1.271) (1.020) (1.044) (0.900) lab.conflicts * (-1.542) (-1.693) (-1.364) (-1.819) (-1.196) (-1.467) (-1.404) AGRemp 5,419** 5,332** 6,156*** 5,291** 5,726** 5,823** 5,268** (2.317) (2.379) (2.771) (2.401) (2.599) (2.640) (2.286) fdi (0.909) fdi ser 0.858*** (5.364) fdi ind (-1.525) fdi agr 145.4*** (3.333) fdi US (1.076) fdi CAN (-0.184) fdi ES 0.886*** (3.525) Observations R-squared Number of state Periods Notes: ***, ** * indicate significance at the 1, 5 and 10 percent level, respectively. t-statistics obtained from robust standard errors are displayed in parentheses. The dependent variable is always structural transformation calculated from Equation 1. All regressions include year dummies not reported. this outcome is quite important. On the other hand, Spanish FDI has a positive impact. In this respect, both types of indicators are significant. Moreover, we find some evidence (based on the initial year indicator) that FDI flows in agriculture and services have a positive impact. Especially, the result for agriculture is interesting as the amount of FDI in this sector is very small and, additionally, average productivity levels appear to be relatively low. Yet, the large coefficient suggests that the positive impact on structural change is comparably high. Finally, our within-country analysis shows that there are effects from FDI on structural change. However, the impact from particular FDI inflows may be working in opposite directions and a general conclusion is not possible. 18

20 References Amendolagine, V., N. D. Coniglio, and A. Seric (2016): FDI and structural change in Africa: Does the origin of investors matter?,. Baldwin, R., and J. Lopez-Gonzalez (2015): Supply-chain Trade: A Portrait of Global Patterns and Several Testable Hypotheses, The World Economy, 38(11), Borensztein, E., J. De Gregorio, and J.-W. Lee (1998): How does foreign direct investment affect economic growth?, Journal of International Economics, 45(1), Carstens, A. G., and A. M. Werner (2000): Mexico s monetary policy framework under a floating exchange rate regimepp , Inflation Targeting in Practice: Strategic and Operational Issues and Application to Emerging Market Economies. International Monetary Fund, Washington. Duarte, M., and D. Restuccia (2010): The Role of the Structural Transformation in Aggregate Productivity, Quarterly Journal of Economics, 125(1), Escobar Gamboa, O. R. (2013): Foreign direct investment (FDI) determinants and spatial spillovers across Mexico s states, Journal of International Trade & Economic Development, 22(7), Görg, H., and E. Strobl (2005): Spillovers from Foreign Firms through Worker Mobility: An Empirical Investigation, Scandinavian Journal of Economics, 107(4), Gurría, J. A. (2000): Mexico: Recent developments, structural reforms, and future challenges, Finance and Development, 37(1), 23. Lewis, W. A. (1954): Economic Development with Unlimited Supplies of Labour, The Manchester School, 22(2), Mayer, T., and G. Ottaviano (2007): The happy few: the internationalisation of European firms, Bruegel Blueprint Series. Bruegel, Brussels. McMillan, M., D. Rodrik, and Í. Verduzco-Gallo (2014): Globalization, Structural Change, and Productivity Growth, with an Update on Africa, World Development, 63,

21 OECD (2015): OECD Economic Surveys: Mexico. OECD Publishing, Paris. SCJN (2010): Political Constitution of the United Mexican States. Mexican Supreme Court SCJN, Mexico City, 4 edn. Timmer, M., G. de Vries, and K. de Vries (2015): Patterns of Structural Change in Developing Countriespp , Routledge Handbook of Industry and Development. Routledge, Oxford. Wacker, K. M. (2016): (When) Should We Use Foreign Direct Investment Data to Measure the Activities of Multinational Corporations? Theory and Evidence, Review of International Economics, 24(5), World Bank (2017): World Development Indicators, assessed January 31, 2017, 20

22 Appendix A1 Definition of variables and data sources Variable ST gdppe migration lab.conflicts AGRemp fdi Definition Structural transformation in pesos per employee Gross domestic product in pesos per employee Net national migration (immigrants minus emigrants) per 100 inhabitants Number of labor conflicts Employment in agriculture as a share of total employment in percent FDI inflows in million pesos A2 Descriptive statistics Variable Obs Mean Std. dev. Minimum Maximum ST gdppe migration lab.conflicts AGRemp fdi5yrs fdi Note: All descriptive statistics are calculated on the basis of the sample included in the estimations presented in Tables 2 and 3. Std. dev. refers to the within standard deviation. 21