Made in Mexico: The Energy Reform and Manufacturing Output1
1. Manufacturing activity in Mexico surged after the signing of the North America Free Trade Agreement, NAFTA. Since then, Mexico has attracted or created world-class performers in the manufacturing sector. Greater integration and lower trade barriers with its largest trading partner brought about new investment, new technology, and thus higher output. Today, manufacturing exports account for about 80 percent of total exports, of which about a third corresponds to automobiles.
2. Nevertheless, the cost and quality of electricity, and more recently temporary shortages of natural gas, are drags to competitiveness of the sector. Mexico ranks consistently low on indicators of global competitiveness regarding the costs and quality of electricity. And limited investment in infrastructure has led natural gas pipelines to operate close to maximum capacity and unable to accommodate rising demand, including from the manufacturing sector (Bank of Mexico, 2013).
Quality and Easiness to Get Electriciy in Selected Countries
Source: World Economic Forum Competitiveness Report (2014) for quality of electricity and Doing Business Report (2014) for easiness to get electricity, which looks at procedures, time, and cost required for a business to obtain a permanent electricity connection for a newly constructed warehouse.
3. The recent energy reform has the potential to alleviate these bottlenecks and provide a new impetus to manufacturing activity in Mexico. The reform opens the hydrocarbons and electricity sectors to private participation in different modalities. Increased investment will help improve existing infrastructure, efficiency in the sector, and the availability of natural gas. As these changes take place, Mexican manufacturing firms will benefit from lower electricity prices.
4. This paper aims at quantifying the impact of the energy reform on Mexican manufacturing activity through its impact on the cost of energy inputs. It is part of a broad agenda aiming at quantifying the effects of the recently enacted package of structural reforms on the Mexican economy. The next section presents stylized facts about the industry and its evolution since NAFTA, followed by an assessment of the economic impact of the reform, starting with estimates of elasticities of manufacturing output to energy prices and the scope for the reform to affect energy prices in Mexico. We then assess spillovers within the manufacturing industry, to the rest of the economy, and within the different regions in Mexico, followed by concluding thoughts.
B. The Mexican Manufacturing Sector Since NAFTA
5. Manufacturing activity in Mexico surged after NAFTA, particularly in the production of transportation equipment. Mexico’s car production tripled since NAFTA and today Mexican cars account for about a third of U.S. auto imports, recently surpassing Japan to become the second biggest car exporter to the United States. And Mexico also supplies one third of all U.S. imports of auto-parts. The most recent announcements include Daimler Benz and BMW, with combined estimated investments surpassing US$2 billion.
6. Growth in manufacturing was, however, uneven across sectors and regions. The expansion was notoriously more pronounced in manufacturing than in other sectors, at least initially. But even within manufacturing, growth was highly uneven with transportation equipment being a clear outlier. Moreover, when looking at contributions to growth, transportation equipment is the only sector with meaningful contributions from total factor productivity, with almost a fifth of overall growth. Heterogeneity in performance across sectors translated into geographical disparities in performance, as regional manufacturing clusters implied much stronger performance in the north than in the south.
Mexican Manufacturing Output since Nafta
Source: Instituto Nacional de Estadisticas y Geografia (INEGI), and staff calculations.
7. Mexico’s manufacturing sector faced also important headwinds with the rising importance of China in the global market. The rise of Mexico’s manufacturing sector was hard hit by China’s rise on the global stage when it joined the WTO in 2001 (Kamil and Zook, 2012). China was able to crowd out Mexican exports in the U.S. market because Mexico had lost its advantage in several labor-intensive manufacturing sectors in which it specialized. But almost as quickly as it stumbled, Mexico regained its footing and began to claw its way back. Rising oil prices may have contributed to an inflection point in Mexico’s U.S. market share around 2005 by raising the importance of proximity to the U.S. market as a competitiveness factor as transportation costs increased.
C. The Energy Reform: How Much of a Boost for Mexican Manufacturing?
8. While the energy reform can affect manufacturing production through several channels, we focus on its effect through lower energy costs. The reform can lead to higher capital accumulation as new investment arrives. And with this new investment, technology transfers may also open the door to increases in overall productivity. However, at this juncture, there is substantial uncertainty regarding these broader channels, but it is possible to infer how manufacturing output would respond to lower energy prices from past data. Complemented with estimates of the potential reduction in energy prices, these estimates can help us measure the economic effects of the reform through its impact on energy prices and manufacturing activity.
How Sensitive is Manufacturing Output to Changes in Energy Prices?
9. We estimate the response of manufacturing output to changes in energy prices using a simple panel regression analysis. The left-hand side corresponds to the real gross domestic product for manufacturing industry i, in state j, in year t. The right hand-side includes a lag of the dependent variable, the variables of interest: the lagged change in electricity prices, EL, in natural gas prices, NG, and oil derivatives prices, OD. The focus on these particular energy inputs arises from their importance in industrial production (Table 1). The change is computed after deflating energy prices with the consumer’s price index to reflect changes in real terms. The regression includes controls, X, in the equation below, that have been found to be important in explaining manufacturing activity, including unit labor costs, the real effective exchange rate, the cost of capital, industrial production in the United States, and other variables, all in first difference form. The regressions include also fixed effects at the sector-state pair level, δij.
|Industrial sector total||1,381.1||1,492.3||1,530.6|
10. Among energy inputs, electricity prices have the largest impact on manufacturing output, with an elasticity of up to -0.28. Tables 2 and 3 show the estimated elasticities under various specifications for the most important energy sources. The elasticities range from -0.11 to -0.28. For natural gas, the elasticities range from -0.04 to -0.07. The independent effect of natural gas, aside from its impact through electricity prices, comes from the fact that about 18 percent of the national demand for gas comes from the industrial sector, including manufacturing. The impact of a one percent change in electricity prices far exceeds the one from natural gas prices. Interestingly, oil derivatives come up with a positive sign. While in principle this positive coefficient could be picking up the increased importance of proximity to U.S. as competitiveness factor in a world with rising oil prices and transportation costs (Section I), it remains positive even after controlling for international oil prices. Nevertheless, its statistical significance weakens as additional controls are included (Table 2, columns 5–7; Table 3, column 3)
|Dependent variable: Δyijt||(1)||(2)||(3)||(4)||(5)||(6)||(7)|
|Sample: 1998–2012, annual|
|Dependent variable: Δyijt||(1)||(2)||(3)||(4)|
|Sample: 1996–2012, annual|
11. Changes in electricity prices have a larger effect on the sector of metals and transportation equipment, which includes the auto industry. Naturally, not all subsectors within manufacturing are affected equally, as shown in the figure and in more detail in Table 4. The figure shows the elasticity for each subsector (red dot) and its confidence interval (vertical line across the dot). Clearly, the elasticities vary widely, with the chemical sector exhibiting a -0.1 value and not statistically significant, to the metals and transportation equipment with a highly statistically significant elasticity of -0.51. This is almost twice as large as the -0.28 average elasticity for the whole manufacturing industry discussed above. But this is not entirely surprising, given the importance of electricity as an energy input for this section shown in Table 1.
|Dependent variable: Δyijt, Sample period: 1996–2012||(1)||(2)||(3)|
|ΔNGt−1 *Food, beverages, and tobacco||0.124||0.029|
|ΔNGt−1 *Wood and wood products||0.065||−0.000|
|ΔNGt−1 *Paper and paper products||0.084||0.026|
|ΔNGt−1 *Chemicals and oil derivatives||0.065||−0.059|
|ΔNGt−1 *Minerals, non-metals||0.128||0.061|
|ΔELt−1 *Food, beverages, and tobacco||0.370||0.317|
|ΔELt−1 * Textiles||0.155||0.097|
|ΔELt−1 *Wood and wood products||0.225||0.228|
|ΔELt−1 *Paper and paper products||0.249||0.202|
|ΔELt−1 *Chemicals and oil derivatives||0.302||0.421|
|ΔELt−1 *Minerals, non-metals||0.345||0.230|
Elasticities of manufacturing output to elasticity prices by
Source: INEGI and staff calculations.
By How Much Can the Reform Reduce Energy Prices in Mexico?
12. There is scope to reduce generation costs of electricity in Mexico by substituting fuel oil for natural gas in electricity generation. Electricity tariffs for commercial and industrial users in Mexico are roughly twice as high as those in the United States. But both rely in similar magnitudes—at about 75 percent of total generation capacity—on fossil fuels. However, one dimension in which these countries differ significantly is in the importance of oil derivatives within fossil fuels. Mexico relies much more on oil derivatives, mostly fuel oil, while the U.S. on coal, which is substantially less expensive.
13. Fuel substitution can potentially lower electricity prices on average by about 13 percent, with competition and efficiency gains bringing in further reductions. The shale gas boom in the U.S. has driven prices of natural gas down significantly, and for the last few years a sizable gap between the price of crude oil—and its derivatives, including fuel oil—and natural gas has persisted. Under the current electricity pricing mechanism in Mexico, if fuel oil were to be substituted for natural gas, industrial and commercial tariffs would go down by about 13 percent.2 Given existing plans for pipeline investments, and conversion of fuel oil plants to operate with gas, it seems feasible to obtain these reductions within five years.3 Strictly speaking, this substitution can be done without the energy reform. However, by attracting private investment in low-cost electricity generation, the reform makes it more likely that the gains from this substitution become permanent, and thus reversals—as the one seen in 2010/2011 (Chapter I)—less likely. Additional reductions could arise from increased efficiency. Energy losses in Mexico are high for international standards, which in part come from an aged transmission and distribution infrastructure. With the reform, new investment and the ability for transmission and distribution lines to be operated and maintained—although ownership will still be public—by private companies is possible. This has the potential to lead to efficiency gains which could help reduce electricity prices further. Assuming convergence to U.S. prices as a limiting case, these gains could lead to a reduction of about 50 percent in commercial and industrial electricity prices in total.
Electricity Prices and Generation in Mexico and United States
Source: U.S. Energy Information Administration, Secretaria de Energia de Mexico, and staff calculations.
Energy Prices and Energy Losses.
Source: Statistical Review of World Energy - BP (2014), Secretaria de Energia de Mexico, Instituto Nacional de Estadisticas y Geografia de Mexico, and U.S. Energy Information Administration (2011).
14. Increasing the supply of natural gas to be used in electricity generation would come first from additional imports from the United States. Since 2012, shortfalls of natural gas, explained by lack of spare capacity in gas pipelines to accommodate shocks and increasing demand, has affected the manufacturing sector importantly. Estimates by the Bank of Mexico suggest that without these shortfalls, GDP growth in 2013Q2 in annual terms would have been -0.4 and not -0.7 percent (Bank of Mexico, 2013). For the more immediate future, the potential of the energy reform to lower electricity prices and to avoid shortfalls in supply lies in increased investment in infrastructure to increase the capacity of the pipeline system to import more gas from the U.S. Several private projects are already underway (footnote 3 and Chapter I), and under the reform, all are required to provide open access, which will reduce the likelihood of bottlenecks in supply.
|Elasticities (from Table 1)||−0.11||−0.28|
|Scenario 1: Substitution of fuel for natural gas|
|Increase in manufacturing output||1.4||3.6|
|Increase in overall GDP||0.2||0.6|
|Scenario 2: Convergence to U.S. Levels|
|Increase in manufacturing output||5.5||14.0|
|Increase in overall GDP||0.9||2.2|
Real Effects of Lower Electricity Prices
15. Should a reduction in electricity prices in the ranges shown above materialize, manufacturing output could increase up to 14 percent and overall GDP up to 2.2 percent. The estimated elasticities imply that manufacturing output could increase between 1.4 and 3.6 percent if a reduction in electricity prices of 13 percent materializes. Furthermore, the increase in manufacturing output could reach 14 percent if prices were to converge to U.S. levels. For the economy as a whole, these impulses for manufacturing activity imply an increase in real GDP of up to 0.6 percent and up to 2.2 percent for each scenario respectively, given today’s share of the manufacturing sector in the economy. These level effects could materialize over the horizon that takes electricity prices to exhibit these reductions. As mentioned before, it is reasonable to expect the estimated gains under scenario 1 to materialize over 2016–2019. Timing of convergence of prices to U.S. levels is more uncertain because it requires reducing energy losses in transmission and distribution which is more challenging. Nevertheless, the scenario offers a benchmark of how large the effects over the long-run can be.
16. Increased supply of natural gas could allow substituting LPG and reduce imports of LNG; however, the effect of the reform on other energy prices is unclear. The industrial sector in Mexico consumes about 10 percent of total LPG demanded in the country, which if substituted for natural gas, it could provide an additional impetus to growth in manufacturing. Table 2 shows an elasticity of about -0.09 for LPG prices, but this number may be capturing demand effects as well, and thus it should be interpreted with caution.4 And more broadly, increased availability of natural gas imported through pipelines can reduce and possibly even eliminate, in the long run, the need to import it in liquid form (LNG). Estimating the impact of the energy reform on oil and oil derivatives prices is more complex because those are commodities which are influenced by global demand and supply.
D. Are There Additional Indirect Effects Through Spillovers?
17. To allow for endogenous responses of unit labor costs and assess spillover effects we turn to a panel vector autoregression framework (VAR).5 The results shown in the previous section highlight the direct impact of changes in energy prices on manufacturing output. However, there may be indirect effects that could imply lower or higher total effects on manufacturing output from changes in energy prices. For instance, unit labor costs may react to changes in energy prices, analogously to how they react to oil-price shocks (Blanchard and Gali, 2007), because of a potential substitution effect that induces an increase in labor demand or to labor supply channels associated with higher costs of living if changes in energy prices are passed to prices of good and services. Similarly, a subsector within manufacturing may respond directly to energy prices and indirectly through its dependence on other subsectors. Using the same logic, we explore regional spillovers. We focus our attention on electricity prices which as shown before are the ones with the largest impact on output.
Impulse response functions to a rise in electricity prices
Note: Impulse response functions (solid blue) are calculated for a period of 5 years with a shock of 1% to electricity prices. 90% confidence interval bands (dashed red) are based on parametric bootstrapped standard errors.
18. Allowing unit labor costs to respond endogenously to changes in electricity prices amplifies the response of output. The figure above shows the cumulative response to a 1 percent change in electricity prices both on manufacturing output and unit labor costs in the manufacturing sector. Manufacturing output responds directly to changes in electricity prices. This is the impact effect measured in period 2 in the figure, which would be comparable to the panel regressions in the previous section. However, manufacturing output also responds to the change in unit labor costs, which as the figure shows it is induced by changes in electricity prices. As a result, the direct effect of changes in electricity prices on manufacturing output gets amplified by the indirect effect electricity prices has through unit labor costs. By the fifth year, the cumulative effect of an electricity price shock on manufacturing output is about 50 percent larger than the impact effect.
19. Nevertheless, under the employed methodology, we do not find statistically significant evidence of regional and sectoral direct spillovers. Appendix Figure 1 shows the impulse responses of manufacturing output in each manufacturing subsector as well as in other sectors in the economy to changes in electricity prices. The figure shows two impulse responses in each chart, one corresponding to a model with (dotted purple) and without direct spillovers (solid blue). Spillovers refer to direct dependence of output in one sector to output in a difference sector, with regional spillovers defined analogously. This is done by restricting, in the no-spillovers case, the relevant coefficients to zero. Evidence of spillover effects would show up through a divergence of the two lines. While the figure shows that the lines do differ in several cases, they are always within the 90-percent confidence bands (dashed red lines), suggesting that the difference is not statistically significant. A similar conclusion arises in the analysis looking at regional spillovers.
20. Several reasons can explain the lack of statistical evidence of direct spillovers. First, there could be substantial crossectional and time variation in the spillover coefficients. Second, some interdependence can happen simultaneously, which implies that the impact effect already captures any additional amplification from spillovers. And third, spillovers through direct interdependence may appear weak given the uneven performance in past data across manufacturing subsectors and regions. In sum, direct spillovers of the kind explored in this paper and under the chosen methodology do not appear statistically strong. These results are however not conclusive given the caveats noted above.
E. Concluding Remarks and Policy Implications
21. The energy reform is likely to have important real effects through its impact on manufacturing activity through lower electricity prices. These effects would come on top of the direct effects on growth that would arise from increased investment and production in the energy sector, for instance, from oil and gas exploration and extraction. Other factors, such as technology spillovers and increased foreign direct investment in manufacturing as the sector becomes more competitive could amplify the effects estimated in this paper.
22. In terms of policy priorities, increasing gas pipeline capacity to allow larger natural gas imports from the U.S. will yield the most immediate gains. In addition, existing fuel oil-based plants would need to be adapted to operate with natural gas. This would allow collecting the low-hanging fruit associated with substitution of fuel oil for natural gas in electricity generation. As the reform starts attracting private investment in low-cost electricity generation, the gains from fuel substitution are more likely to become permanent. Further gains will follow from increasing availability of gas throughout the country by an expanded network of pipelines.
23. A strong regulator is critical to ensuring competition in electricity generation and thus making any reduction in electricity prices long-lasting. To this end, synergy with the antitrust reform will play an important role. The new antitrust framework should help ensure an efficient opening of the sector to private investment to ensure healthy competition.
24. For gains in efficiency, it is critical that the operation of transmission and distribution lines encompass the right incentives to improve existing infrastructure. As distribution and transmission lines will remain property of the state, gains in efficiency will arise from having in place the right incentives for the new administrator of the infrastructure to invest and lower the high technical losses in the system. A word of caution is needed. Reducing distribution losses will be challenging and thus the gains from increased efficiency, while in theory are important, are not guaranteed.
Spillovers are assessed using Panel VAR’s with output, real interest rates, U.S. industrial production, unit labor costs, and electricity prices. Manufacturing output (yi,j,t) is measured at each state (i), subsector (j), and year (t), real interest rates (rt) correspond to the ex-post real yield on Moody’s U.S. corporate bonds, U.S. industrial production (USIPt), unit labor costs measured at the subsector (ULCj,t) and electricity prices (Elect). Identification is achieved through Choleski ordering, with the baseline specification assuming the following order, including fixed effects at the state-subsector level:
where J is an upper triangular matrix, ci,j are state-subsector fixed effects, ui,j,t is the error term, Σ is a constant covariance matrix and A is a 5×5 coefficient matrix.
We also estimate a variant of the above system, allowing spillovers within subsectors of the manufacturing industry and non-manufacturing sectors. To this end, we estimate the following variation of the previous equation:
where Yi,t is a 13×1 vector containing the manufacturing output for each subsector, as well as the output growth in non-manufacturing sectors of the economy, taken from the national accounts. ULCt are the total manufacturing unit labor costs at the national level, and ci are state fixed effects. Impulse response functions without spillovers are calculated by restricting the coefficients of output in each sector to the output in other sectors to zero.
Regional spillovers are estimated similarly. Model 4 studies the effect of potential spillovers between regions by estimating:
where Yi,t is a 5×1 vector containing the manufacturing output from each economic region of Mexico, and fixed effects are included at the subsector level.
In all models, the real interest rate and U.S. industrial production are assumed to be exogenous from other variables. Impulse response functions to a 1 percent shock to electricity prices are reported along with bootstrapped confidence intervals.
Appendix Figure 1.Impulse response functions to a rise in electricity prices with subsector spillovers
Note: Impulse response functions are calculated for a period of 5 years with a shock of 1% to electricity prices for both the model without spillovers (solid blue) and with spillovers (dotted purple). 90% confidence interval bands (dashed red) are based on parametric bootstrapped standard errors on the model without spillovers. All variables are expressed as yearly differences.
Appendix Figure 2.Impulse Response Functions to a Rise in Electricity Prices with Regional Spillovers
Note: Manufacturing output for the five economic regions of Mexico. Impulse response functions are calculated for a period of 5 years with a shock of 1% to electricity prices for both the model without spillovers (solid blue) and with spillovers (dotted purple). 90% confidence interval bands (dashed red) are based on parametric bootstrapped standard errors in the model without spillovers. All variables are expressed as yearly differences. The regions are defined as follows: Center (Distrito Federal, Hidalgo, Estado de Mexico, Morelos, Puebla, Tlaxcala), Center-West (Aguascalientes, Colima, Guanajuato, Jalisco, Michoacan, Nayarit, Queretaro, San Luis Potosi, Zacatecas), Northeast (Coahuila, Chihuahua, Durango, Nuevo Leon, Tamaulipas), South-Southeast (Campeche, Chiapas, Guerrero, Oaxaca, Quinatana Roo, Tabasco, Veracruz, Yucatan), and Northwest (Baja California, Baja California Sur, Sinaloa, Sonora). All variables are expressed as yearly differences.
Alvarez and Valenciaforthcoming. “Made in Mexico: Energy Reform and Manufacturing Growth.” IMF Working paper.
Bank of Mexico2013. “Efectos del Desabasto de Gas Natural sobre la Actividad Económica.” Informe sobre la InflaciónJulio–Septiembre2013 pp. 31–34.
BlanchardO. J. & GaliJ.(2007). The Macroeconomic Effects of Oil Shocks: Why are the 2000s so different from the 1970s? (No. w13368). National Bureau of Economic Research.