Statistical analysis

The five equations making R² up the model in a structural manner showed from 62.03 to 98.43 %. Therefore, most of the changes in endogenous variables are explained by exogenous variables, which hypothetically were considered the most influential in the former (Table 1).

Regarding the individual t tests, in the relationship of the supply of crude oil in its structural form, the estimated coefficients of the international lagged supply of crude oil, the lagged supply of national crude oil, the amount of proven oil reserves in Mexico, as well as the ratio of the real international price of crude oil to the real price of a barrel of crude oil were statistically significant, at a 5.0 % level of significance. In contrast, the estimated coefficient of crude oil extraction cost in Mexico was not statistically significant. This variable was not discarded because it was considered important in shaping the model.

The relationship of the demand of gasoline, the estimated coefficients of the national price of gasoline, the GDP and the number of registered vehicles in circulation presented the expected signs and were statistically significant, according to the t statistic. Within this relationship, as suggested by ^{Sánchez et al. (2015)}, the increase in the demand of gasoline in Mexico is determined by increases in its price and consumer income.

For the relationship between the national price of gasoline and the international price of gasoline, the expected coefficients for their variables, international price of gasoline and international price of crude oil, respectively, were statistically significant, and their signs were as expected since they agree with the theory, as well as with ^{Ibarra and Sotres (2008)}.

Finally, for the relationship of the barrel price of crude oil, the expected coefficients of the international price of crude oil and the barrel price of crude oil lagged one period were in statistical terms significantly different from zero, and their resulting signs matched with predictions, as well as those obtained by ^{Barrañón (2008)}.

Economic analysis

We calculated the elasticities of the structural form with the partial derivatives of each equation and the average values from 2011 to 2015 of the variables used, in order to quantify the effects established in the functional ratios. The analysis of the elasticities for the whole model implies the concept of ceteris paribus, that is to say, when establishing a change in some independent variable on the dependent, we assumed that the rest of the independent ones remain unchanged.

The response of the supply of crude oil was reversed and elastic in the face of changes in the international supply of crude oil lagged one period, given its elasticity (-1.26) (Table 2). This allows to explain why Mexico takes as reference primarily the supply and prices of the USA because its production could be substituted; thus, by increasing the lagged production of crude oil abroad, the national production would be reduced.

Given the previous effect, during the study period, the international supply of lagged crude oil increased by 24.50 %, for which the national crude supply decreased by 30.88 % (331 771.16 thousand barrels). Regarding the cost of extraction, it resulted negative and inelastic (-0.17), which is opposed to IMF studies (2017), which through an integration model points to this value as positive and elastic in the world. The increase of such variable would mean a reduction in Mexico’s supply because it would bring about higher production costs, and therefore, lower profits. However, for countries that have the technology to produce derivatives and refined petroleum products, like the USA, Japan, Canada or Germany, this cost would translate into investment in technology, which in the long term would lead to an increase in crude production.

The supply of crude oil presented a positive ratio (0.66) compared to the lagged supply of crude oil. During the study period the supply of lagged crude oil decreased by 8.95 %, which caused a reduction of 5.92 % (63,560.97 thousand barrels) in the supply.

According to the amount of proven reserves of oil in the country, supply presents an almost null and negative elasticity (-0.04), which indicates that the current trend is downward for the national reserves. This in turn affects inversely the supply of crude in the country, while in studies carried out worldwide (^{Alsahlawi, 2009}) this elasticity has a value ranging between 0.03 and 0.70, which would explain the growing consumption of oil in the world that leads to ensuring energy needs, despite the level of production.

The relationship of the international price of oil and the price of the national barrel of crude oil presents a positive and very inelastic response in the supply of oil in the country, given its elasticity of 0.0005. Regarding the national supply price elasticity, through an analysis of consensual estimators and sensitivity tests, ^{Payet (2005)} found a value for the USA of 0.2, while the ^{IMF (2005)} calculated a world value of -0.05 for the period from 1973 to 2004. As a price ratio, the effects of the two variables converge. According to economic theory, it would be expected that in view of the increase in the price of domestic oil, the supply would expand, which is almost neutralized by changes in the international price of oil. Therefore, the one-period lagged amounts of national and international supplies would be the variables with greater influence on the supply of crude oil in the country.

The relationship of gasoline demand showed a negative and inelastic response to changes in the domestic price of the product (-0.67). ^{Mendoza (2005)} found through a cointegrated vector method an elasticity of -0.47, result that agrees with that of our study. Regarding the GDP variable, the demand for gasoline presents a positive and inelastic response (0.22), thus signaling gasoline as a normal good in Mexico. In studies conducted by ^{Mendoza (2005)} and ^{Dahl and Sterner (1991)}, the latter for the USA, similar elasticities of 0.07 and 0.48 were observed, respectively, indicating a positive ratio between income and demand, as indicated by economic theory. The demand of gasoline experienced a positive and inelastic response (0.71) to the changes in the registered number of vehicles in circulation. Studies in Denmark (^{Bentzen, 1994}), using cointegration techniques, found an elasticity of 0.89, which agrees with our study; this would indicate that an increase in the number of vehicles would be followed immediately by a similar increase in the demand of gasoline. The price of gasoline and the registered number of vehicles in circulation are the most significant variables in the demand for gasoline, because of the theory (law of demand) and the huge growth of the automotive industry in the country during the last decades (^{INEGI, 2015}).

In the relationship of the national price of gasoline based on its international price, a positive and inelastic response was obtained (0.46); this explains the fact that Mexico imports more than half of the gasoline consumed in the country (^{Presidency of the Republic, 2017}). On this account, when prices abroad increase, they also increase in Mexico, though in a smaller proportion. In the face of a 10 % increase in the international price of gasoline, the domestic price responds directly to 4.6 %; this means that the effect of the international price of gasoline is not transferred in full to the national price. The relationship of the international price of gasoline also had a positive and inelastic response (0.54) to changes in the international price of crude oil during the same period, which is due to the fact that, faced with a price increase of crude oil, the basic material for the production of gasoline, it is expected that this effect will be transmitted to the national price of gasoline (^{Presidency of the Republic, 2017}). Within the study period, the international price of crude oil increased by 45.8 %, and represented an increase of 24.94 % (0.13 real dollars per liter) in the international price of gasoline; thus, as in the previous report, the effect of the international price of crude oil on the international price of gasoline is not completely transmitted.

In the relationship of the price of a barrel of crude oil, a positive and inelastic response (0.87) was obtained in response to changes in the international price of crude oil. The same elasticity was calculated by ^{Mota and Mata (2015)} obtaining a value of 0.59, which indicates the same effect with a similar magnitude, and is due to the fact that Mexico takes as reference the price of oil in the USA (West Texas Intermediate, WTI) within its price determination. This price increased by 45.8%, which represented an increase of 39.84 % (10.49 real dollars per barrel) in the national price of crude oil. Finally, the price of the barrel of crude oil presented a positive and inelastic response (0.20) to the value of the same product lagged for one period. This is due to the expectations of the national market in the face of previous price trends. Thus, the increase of 250.46 %, as occurred during the present study period, led to a rise of 49.24 % (12.96 real dollars (2010) per liter) in the price of a barrel of Mexican crude oil.

The elasticities of the reduced form were calculated from the estimated model and indicate the final effects of the exogenous variables on the gasoline foreign trade balance (Table 3). The foreign trade balance of gasolines was defined as the demand minus the supply of this good.

In the relationship of the foreign trade balance of gasoline, a negative and elastic response was obtained as a result of increases in the international supply of crude oil lagged one period (-3.69); thus, by increasing 10 %, this variable would cause a reduction of 36.88 % in the foreign trade balance (imports), which means that this is the factor with the highest degree of incidence in this balance. This suggests that the trade balance of gasoline in Mexico is based largely on the past behavior of oil production. The foreign trade balance of gasolines due to changes in the cost of oil extraction in Mexico was negative and inelastic (-0.50), which would mean that as a result of a 10 % increase in the cost mentioned, the trade balance would be reduced by 5 %. This effect would be the result of investments in new extraction projects in deeper waters, suggesting that in the face of greater investment in extraction processes, yields would increase, bringing about a reduced trade balance of gasoline in the country.

The effect of the crude oil supply in Mexico lagged one period represents positive and more than proportional changes in the trade balance (1.93). Thus, for a reduction of 3.65 %, as in the last study period, it would result in a decrease of 7.06 % (9274.45 thousand barrels) in the foreign trade balance of gasoline. This production has a more than direct impact on the trade balance, therefore investments and actions destined to increase oil production in the country acquire greater importance. Meanwhile, the effects caused by fluctuations in the amount of proven oil reserves would be negative and inelastic within the trade balance (-0.11). If a 10 % increase in the variable in question occurred, the result would be a reduction of 1.1% in the foreign trade balance. Both the meaning and discrete magnitude of the effect are due to Mexico’s current need to import more than half of the gasoline consumed in its territory, a situation that in terms of the trade balance diminishes the relevance of the amount of oil available in the country.

As the lagged oil supply, increases in the international price of oil would represent an increase in the balance of foreign trade (0.32) by transmitting its effect through the oil supply and the demand of gasoline (via the national price of gasoline through its price in the international market), although its effect is not reflected in full. In this way, for a reduction of 47.50 %, as occurred in the period from 2014 to 2015, changes in the balance would mean a fall of 15.36 % (20,179.70 thousand barrels) in the trade balance. This situation is due to the considerable fall in world oil prices in recent years.

Because of changes in GDP, the resulting effects on the trade balance would be negative and inelastic (-0.42). This indicates that with a 10 % increase in GDP, changes within the foreign trade balance would be only 4.2 %. This situation obeys, as already mentioned, to the character of gasoline as a normal good in Mexico’s consumption.

Another important factor in the hydrocarbon trade balance is related to the movements in the registered number of vehicles in circulation, which meant negative and more than proportional (-1.36) effects. Thus, in the face of a 10 % increase in the number of vehicles in the country there would be a 13.6 % decrease in the trade balance of gasoline. Finally, the foreign trade balance represented positive and very inelastic changes (0.0015) in the face of changes in the international price ratio to the price of a barrel of oil in Mexico. Thus, for an increase of 10 % in this ratio there would be an effect of barely 0.015 % in the foreign trade balance of gasoline in Mexico.