Introduction

For a long time, the price of maize (Zea mays), expressed in real terms, was characterized by its stability and even by its secular decline, with occasional oscillations. However, between 2005 and 2012 the price per ton of this product in the global market increased from 112.5 (constant 2010 dollars) to 277.4, recovering a value that was not observed since 1975 (^{Banco Mundial, 2015}). In Mexico, the average rural price of the ton of grain maize had a similar evolution when going from 1964 to 3737 (constant 2010 pesos)³. In nominal terms, the increase was 154 % in those seven years, period during which the general price index in the economy barely increased 33 %.

The increase in this grain’s price, which began in 2005 and prolonged until 2012, has already been analyzed from different perspectives, being an object of strong controversy because maize is not just any product, but rather the basis of the diet and the essence itself of the Mexican population⁴. In this occasion we resort to the suggestions of the dependentist and neostructuralist schools, with the aim of analyzing the factors that explain the increase in maize price, its consequences, the problems and the opportunities that arose with the rupture of the historical pattern of price evolution. This analytical framework allows interpreting the phenomenon in question from a broad perspective, when considering the economic rationality of the national agents implicated, but which are immersed in a system of political-economic relationships characterized by international asymmetries and inequalities. With this the objective of this research will be reached: showing to what degree the unilateral decision of the United States of promoting ethanol production contributed to increasing the price of this cereal, resulting in welfare increases for the agents of this country, in detriment of the Mexican agents that, nevertheless, could also benefit from the US strategy.

To reach this objective, in the following section the motives that drive the ethanol production from maize are explained, as well as the peculiarities of the maize itself as product of very inelastic offer and demand, and the theoretical suggestions that help to understand the case. Later the analysis techniques and the data sources used are described. In the fourth section of the text, the results obtained through the estimation of a model of simultaneous equations for the maize price in the United States are described, as well as the calculation of elasticities and of variations of the producer and consumer surpluses, and the comparison of the changes in the volume produced, the surface sown and the productivity of both countries. Finally, a section of conclusions weaves together the most relevant contributions of this research.

Materials and methods

With the objective of understanding to what extent ethanol production by the US contributed to increasing the price of this cereal, to verify how Mexican agents were affected and whether these adopted a strategy to take advantage of the change in prices, in the first place it is suggested to calculate the crossed elasticity that reflects the impact of the variation of the amount of ethanol produced in the US on the international maize price, for which the functions of offer and demand of this product are estimated. Then, the producer and consumer surpluses are calculated and their variation is analyzed.

To estimate the functions of offer and demand of maize, a system of simultaneous equations is presented. A set of five equations that allows studying the effects of various factors on the price of maize is described, modifying the original specification by ^{Park and Fortenbery (2007) 6}. This time, the research is centered on the incidence of ethanol production.

The first equation of the model corresponds to the maize offer:

It includes as explicative variables: price of maize (PM _t ), interest rate (I_t) and three dichotomous variables (D₁ to D₃) that are used to correct the seasonality when dealing with data from trimester periods⁷.

The maize demand is disaggregated into three equations that correspond to the categories proposed by the ^{USDA (2016a)}: livestock diet (AG_t ), exports (XT_t), and foods, alcohol and industrial uses (AAI_t). The demand equation of maize for livestock diet,

is made up of the variables: price of maize (PM_t), price of soy (PS_t), poultry livestock (GA_t), bovine livestock (GV_t) and porcine livestock (GP_t), measured in millions of pounds, and the three dummies mentioned before.

The demand for maize exports

is expressed in function of the price of maize (PM_t), price of wheat (PT_t), dollar index (DX_t)⁸, the weighted mean of the gross domestic product per capita in China and India (PIB_t)⁹, and the dichotomous variables. The demand for foods, alcohol and other industrial products,

is explained by the price of maize (PM_t), etanol production (ET_t), population of the United States (Pob_t), the prices of oil (PP_t), soy (PS_t) and wheat (PT_t), a dummy (D₀₇) that takes the value of 1 in the period prior to the promulgation of the Energy Independence and Security Act from 2007 and 0 in the rest, and the three seasonal dummies.

The model includes a fifth equation, of equilibrium, corresponding to the price of maize,

which is determined by the offer (QS_t), the demands (AG_t, XT_t and AAI_t) and the seasonal dummies.

To solve the system we resort to the technique of least squares in three stages (MC3E), which, being a method of complete information, allows estimating the equations in their structural form as a whole, which increases the asymptotic efficiency of the estimations.

The data used come mostly from the ^{USDA (2016a} and ^b), from which information was obtained for the following variables: prices of maize, wheat and soy; amount of ethanol produced in the United States; maize offer; maize demand disaggregated into agricultural uses, exports and food, alcohol and industrial uses; and bovine, poultry and porcine livestock in millions of pounds. The interest rate is the one published by the ^{OECD (2016)}. The data of the dollar index are from the ^{Federal Reserve (2016)}, those of the per capita gross domestic product in China and India from the World Bank, the population of the US from the ^{U.S. Census Bureau (2015)}, and the oil prices from the ^{U.S. Energy Information Administration (2016)}. It should be emphasized that the variables of the model are expressed in logarithmic terms, with the exception of the dichotomous ones; therefore, their coefficients can be interpreted directly as elasticities.

As has been mentioned, the periodicity of the data is per trimester, covering since the first trimester of 2000 to the fourth of 2014, when ethanol production becomes a relevant factor. In total there are 60 observations for each series that allow estimating the functions of maize offer and demand in the United States.

Based on these functions, the impact that ethanol production has had on the price of maize can be calculated through the concept of cross price elasticity. Concretely, for the previous system, the elasticity of the price of maize with regard to the amount of ethanol, there is the following expression:

for which a positive sign should be expected, since higher amounts of ethanol must be related to higher prices of maize, once the effect of the other variables included in the model is controlled.

The variations that these changes imply in the welfare of US agents can be measured with the producer (EP) and consumer (EC) surpluses that correspond to the functions of offer and demand of the system of simultaneous equations:

For the case of Mexico, we find that the country is price-accepting in the maize market, which is why the shocks from international offer and demand are transmitted through prices to the national producers and consumers. It is relevant to test how these changes have affected the welfare of Mexican agents. With this aim, the offer and the demand can be approached with the elasticities that ^{FAPRI calculates (2015) 10}, ε =0.22 for the offer and |η|=0.12 for the demand, the amounts produced, imported and exported that ^{FAO (2016)} presents, the prices by ^{SAGARPA (2016)}; specifically, the mean rural price expressed in pesos per ton, and assuming the functional form of constant elasticity, as has been done for the United States in the model of simultaneous equations¹¹:

The offer indicates the amount produced in Mexico, that is, the offer by national producers, while the demand is expressed in terms of apparent national consumption (amount produced, minus exports, plus imports). From these functions of offer and demand it is possible to calculate the producer and consumer surpluses, and to analyze their evolution in time.

Finally, to study whether producers from Mexico and the United States take advantage of the business opportunities that arise with the rise in prices, the evolution of production in both countries is analyzed resorting to the identity:

where Y represents the production (measured in tons), T is the surface sown (expressed in hectares), and the quotient Y/T is the production per hectare or productivity of land. Due to the mathematical properties of the logarithms, it is fulfilled that the variation rate of the production is approximately equal to the sum of the variation rates of its two components: surface sown and land productivity; that is:

The comparison of what happened in Mexico and in the United States contributes relevant information about the relationships that producers from each country have had when faced with the change in prices, allowing the analysis of how much production has increased and whether this increase is due to the expansion of the surface sown (extensive margin) or to the more productive use of the surface already available (intensive margin).

Results and discussion

The estimation of the model of simultaneous equations through the technique of least squares in three stages gives the following results (p-values are shown in parenthesis):

The offer equation, QS_t, presents the signs expected. The amount offered depends positively on the price and negatively on the interest rate, which implies higher production costs.

The equation of livestock diet, AG_t, also shows the expected sign in the maize price, negative in this case. The price of soy has positive sign, indicating that it functions as a substitute of maize. This substitution-effect is also reflected in the variable of porcine livestock, GP_t, since its negative sign indicates that the maize demand from this sector decreased in face of percentage increases in the amount of the livestock. The parameters of the variables that correspond to poultry and bovine livestock present the signs expected.

In the third equation, the price of maize does not maintain a statistically significant relation with the amount demanded of this cereal, in this case, for export, XT_t. In addition, none of the variables of the equation helps to explain the maize exports that seem to have a more residual behavior than a regular one, something that is also seen in the estimations by ^{Park and Fortenbery (2007)}. In fact, stability in seasonal variations is not even observed.

In the equation of maize demand for foods, alcohol and industrial uses, AAI_t, the non-significance of the price of maize stands out, as well as the positive and fully significant sign of the amount of ethanol. Concretely, the coefficient estimated for ethanol shows that, for each increase of one percentage point in the production of ethanol, the maize demand of this class increased 0.446 %. The variable D₀₇, which indicates the implementation of the 2007 Energy Bill, was found to be significant.

The fifth equation represents the equilibrium of the system. In it the price is defined through the interaction of the offer and the three partial demands.

Once the estimations of all the coefficients of the model of simultaneous equations are available, the impact of ethanol production, not in the amount demanded but on the price of maize, is understood based on the cross price elasticity:

This result indicates that when facing an increase of 1 % in ethanol production, the price of maize has been increasing 0.609 %, that is, ethanol production has had a positive and important effect in the increase of the price of maize.

When it comes to the effect that the increase in the price of maize had on the welfare of society in the United States, we find that when going from 2.12 dollars per bushel in 2005 to the highest in the series, 7.04 in 2012, the producer surplus increased 13 973 million dollars per trimester. In contrast, the consumer surplus, estimated from the demands of the model of simultaneous equations, was reduced in 12 837 million (4942 from the demand for livestock diet and 7895 from the maize demand for foods, alcohol and industrial uses). As a whole, because it is a country with surplus production in maize, the increase of the producer surplus compensates for the loss suffered by the consumers, resulting in a net positive variation of 1136 million, reflecting an improvement in welfare.

The coefficient of the variable PM_t in the offer equation is the price elasticity of maize offer: ε = 0.149. The price elasticity of the total demand for maize in the United States can be recalculated from the aggregation of the agriculture and livestock demand and the demand for foods, alcohol and industrial uses, obtaining a value of η = -0.173. With these elasticities the variations of surpluses were calculated. For the case of Mexico, as has been argued, it is understood that the country is price-accepting, so the analysis of the determination of the price loses relevance, since it is fixed in the international markets. However, the study of the variations of surpluses maintains its interest.

The ^{FAPRI database (2015)} facilitates elasticities of ε = 0.22 for the offer and |η| =0.12 for the demand of maize in Mexico, quite similar to those that have been obtained through the model of simultaneous equations for the United States. ^{Barkley et al. (2011)}, who also analyze the impact of the price shocks in the maize market in Mexico, resort to the figures published by ^{Rosegrant et al. (1995)}, in turn taken from ^{Roningen et al. (1991)}: ε = 0.17 and |η| =0.46. Although there are barely differences in the offer elasticity, the demand by ^{Barkley et al. (2011)} turns out to be quite more elastic than the one built from the ^{FAPRI data (2015)}. This gives rise to discrepancies between the results and the conclusions of these authors and those that are presented next. In any case, the inelasticity of offer and demand is a predominant characteristic, as corresponds to the maize market.

In Mexico, between 2005 and 2011, the price of grain maize went from 1578 pesos per ton to 4078, reaching its historical maximum (^{SAGARPA, 2016}). Discounting the general inflation of the economy (^{INEGI, 2016}), the increase went from 1964 to 3929 constant pesos 2010. The effects of this change in the welfare of producers and consumers can be quantified measuring the variation of their surpluses. For this purpose the annual offer and demand functions of maize in Mexico are estimated based on the elasticities by ^{FAPRI (2015)}, of the amounts produced, imported and exported published by ^{FAO (2016)}, and of the ^{SAGARPA (2016)} prices:

Stemming from these specifications, the variations of surpluses provoked by changes in the prices are calculated. Specifically, there is that in face of an increase in the price from 1964 to 3929 pesos, the consumer surplus is reduced in 55 078 million pesos, at the time that the producer surplus increases in 32 419 million. Due to the situation of the Mexican economy, with a negative balance in maize, the increase in the producer surplus does not compensate the loss suffered by the consumers. As a whole, the rise in price generates an annual loss of welfare valued in 22 659 million pesos.

Although consumers can do little in face of the increase in the price of maize, given that it is a good of first necessity for which they barely find substitutes (the strong demand inelasticity, |η|=0.12, is a reflection of this), the producers can react in different ways to take better advantage of this favorable circumstances to their interests. The increase in price, caused exogenously, in part by policies destined to benefit the US producers, is presented as an opportunity for Mexican producers in the sense suggested by neostructuralism. Have they used this opportunity to their advantage?

The figures indicate that only in part. In recent years, in Mexico, the increase in production has been attained exclusively through the expansion of the intensive margin (Table 1). Far from growing, the number of hectares sown with grain maize has reduced noticeably since the maximum from 1994 (9 196 478 ha) to the minimum from 2012 (7 372 218 ha). After the enforcement of NAFTA, nearly two million hectares devoted to the cultivation of this cereal have been lost. In this same period, the surface of the total cultivation of the country was reduced in “just” one million hectares. This gives an idea of the severe restructuring that commercial openness has entailed for the maize sector specifically.

In more recent years, with the increase in prices, the trend has not reverted. Since 2005, the cultivation surface has continued to be reduced until, although it is true, at a lower rhythm. The increase in productivity of the land has been constituted into the single factor of momentum for production. The use of the intensive margin, at the expense of the extensive, is the fundamental characteristic of recent years.

In the United States, however, the growth of production has been based on the expansion of both margins: both the intensive, and the extensive. In fact, the increase of the surface sown of maize, which has been seen since 1995 (coinciding with the beginning of NAFTA), has not prevented the improvement of productivity of the land, giving place to a sustained growth of production in time, more consistent and robust than in Mexico.

Conclusions

The previous results show that ethanol production by the United States had a direct impact, and relatively important, in the rise in the price of maize. This change implied an increase of welfare for the whole of the US society (with surplus in maize), but a loss for the Mexican society (with negative balance). When examining the neostructuralist approach from the evidence found, it can be said that although it is true that with the rise in price opportunities emerged, these were better taken advantage of by United States producers than by those in Mexico, where only the intensive margin was exploited.

Cases like the one analyzed question whether agents of the periphery have sufficient capacity to detect and take advantage of the opportunities generated by the policies that belong to the center of the system. The same peripheral location makes the understanding of these decisions difficult, at the same time that the lower level of development limits the options for response. The private sector is left practically dismissed under these conditions. Not in vain, neostructuralism itself admits the defects of laissez-faire and concedes the State an active role in the promotion of development. What perhaps is not entirely clear in its proposal is that the peripheral State, just from being, also suffers limitations and, without an objective and certain planning, has difficulty taking advantage of the opportunities present sometimes.