The 2013 Mexican Energy Reform in the Context of Sustainable Development Goal 7

Abstract

In 2013, the Mexican Constitution was amended to allow private firms to participate in the energy sector market. Consequently, the energy reform opened the energy market to private investors, ending the state monopoly of PEMEX and CFE. This article aims to assess the impact of the 2013 Mexican Energy Reform on energy household consumption and, if proven effective, explore its potential to help achieve SDG 7. This longitudinal study gathered data before and after the energy bill reform, from 2012 to 2018, with a non-experimental design. Data analysis to determine the effect of the price variance was estimated through price elasticities of demand, and a logarithmic model was used to determine the relationship between the price and cost of electricity, gas, and fuel. Findings suggest that the 2013 Mexican Energy Reform led to an increase in energy prices that, on the one hand, reduced the consumption of energy generated using fossil hydrocarbons but, on the other hand, affected the portion of the population with less income. Consequently, it is possible to conclude that the 2013 Mexican Energy Reform is irreconcilable with SDG 7 unless substantial additional efforts are made to leave no one behind.

1. Introduction

Aspiring to achieve the requirements for a world where populations worldwide have access to affordable and clean energy is at the core of Sustainable Development Goal 7 (SDG 7). Considering the above, UN member states must focus on assessing their progress toward achieving five targets [1]. As time passed and the deadline approached, the UN General Assembly established a decade of actions that urged countries to accelerate efforts to overcome obstacles hindering the promotion and implementation of diverse SDG initiatives [2]. Yet, this call to intensify efforts to keep the promise to leave no one behind has generally had moderate results. Progress towards accessible and sustainable energy is acceptable compared to other SDGs; there is still a question of whether all of SDG 7’s targets can be met by 2030 [3]. Enhancing effective compliance with SDG 7’s targets is influenced by factors deeply intertwined with the targets in SDG 9. Resilient infrastructure, information and communications, innovation, and finance services are critical not only for SDG 9 but also for SDG 7. Building resilient infrastructure capable of producing energy in extreme weather conditions, natural disasters, or accidents is relevant to a country’s development because it avoids severe disruption to the competitiveness of the industry [4]. Under extreme weather conditions, resilient energy infrastructures become a matter of life and death, particularly for those living in vulnerable zones or countries in the Global South [5]. Yet, progress is falling behind as critical infrastructure energy systems, mainly in the Global South, are not evolving as they should be [6]. Indirectly, SDG 7 demands that countries strengthen their national innovation capacity because innovation drives developments in new climate-resilient infrastructure [7]. Innovation has also favored the design and production of information and communications technology, which is fundamental in the transition toward sustainable energy systems, such as intelligent grids that monitor, control, and diagnose critical operational functions [8].

Above and beyond technical solutions, governments must make a real commitment expressed in government policies to leverage the transition to renewable energy without undermining energy security, which is highly complex and leads to controversy and disagreements about funding mechanisms [9]. The International Energy Agency (IEA) [10] warned policymakers that it is necessary to undertake more efforts to get on track for net zero emissions by 2050. The nexus between renewable energy and energy security is generally at the top of the list of priorities in many countries, which is tackled according to the global and national geopolitical context [11]. Energy security has been conceptualized in different and controversial ways to the extent that there does not appear to be any consensus between scholars and practitioners [12]. Moreover, economic considerations are listed among the main factors taken into consideration by policymakers in determining security policies [13]. In particular, funding sources are shaped, in one way or another, by the political-ideological alignment of the governing party. In this sense, governance of the energy infrastructure plays a critical role in fostering the privatization of infrastructure projects [14]. Nevertheless, the crowding-in of private investment in the energy sector is a strategy many governments have considered feasible to finance works in this sector. According to the World Bank [15], private participation in infrastructure in 2022 was USD 25.9 billion, a 21% increase from 2021.

Mexico is among those countries that have allowed private corporations to do business in its energy sector. The energy sector has been a critical driver of the Mexican economy. Before 2013, foreign investment in the energy sector was banned by the Mexican Constitution. Consequently, the Mexican energy market was dominated by two state-owned firms: Petroleos Mexicanos (PEMEX), an oil company, and Comision Federal de Electricidad (CFE), an electricity company. Then, in 2013, Article 27 of the Mexican Constitution was amended to allow private and foreign firms to do business in its national energy sector, in both the hydrocarbon and electricity sectors, which had previously been prohibited for them [16].

The Mexican government’s arguments to commodify its water, land, and natural resources were multiple. First of all, Mexico could not guarantee the supply of energy for much longer because Pemex’s capacity to explore and produce oil proved to be deficient to the extent that Mexican oil production sharply dropped by over one million barrels from 2004 to 2014 and by that time Mexico was unable to meet the internal demand for natural gas without resorting to import [17], which affected its energy security. According to an analysis prepared by the Senate of the Republic [18], 50% of the estimated resources for oil extraction were located in the depths of the Gulf of Mexico basin. For this reason, Mexico would have to explore waters up to 3000 m deep in the coming years. Second, energy subsidies for energy, gas, and gasoline reached over USD 24,453 million, generating high financial costs, so the Mexican government could not sustain the high levels of existing support [18]. By 2012, subsidies for gasoline consumption exceeded the budget of the Mexican Social Security Institute and the main public poverty reduction program, Oportunidades [19]. A third reason was to improve the technological capacity of deep water operations, transmission, and distribution networks [20], a generic reason expressed in Latin American countries [21].

In the case of Mexico, there were two aggravating circumstances; First, the subsidy was granted to energy generated from fossil fuels that produce carbon dioxide, polluting the environment and contributing to global warming [22]. Second, financial experts have warned that subsidy policies do little to benefit the most vulnerable because they consume less than wealthy consumers, even though their rationale fosters social development, a complex paradox [23]. The latter is a common problem in many developing countries where subsidizing electricity, fuel, and natural gas has led to financial difficulties [24]. Consequently, it is increasingly common for governments worldwide to support energy reforms that abolish subsidies [25].

Consequently, determining the price elasticity of energy demand in detail for electricity, gas, gasoline, and other basic goods before and after the energy reform is of great interest since this indicator is an efficient parameter to measure the effect of an energy policy [26]. The price elasticity of demand determines how much the quantity demanded of a good or service varies depending on changes in its price. The demand for a good or product is elastic if the quantity demanded responds significantly to a variation in price, and the demand is inelastic if the amount required reacts slightly to the variation in price [27]. In the energy market, price changes affect the well-being of companies and households directly and indirectly.

For this reason, scholars keep studying the elasticities of energy demand in different settings [28,29]. Elasticities for energy goods are also increasingly essential to estimate the socioeconomic and environmental effects of the energy policies implemented in countries, which influence the price of energy goods. In this sense, the enactment of Mexican energy reform has drawn attention to price changes, improving production, and long-term incomes; however, their impact on household energy consumption is an issue that has been neglected despite its importance to SDG 7. In this context, this article aims to assess the impact of the 2013 Mexican Energy Reform on household energy consumption and, if proven effective, explore its potential to help achieve SDG 7.

2. Literature Review

Within microeconomics, the consumer theory states that rational consumers seek a way to maximize utility given their budgetary limitations, based on their preferences, and make choices from a set of goods and services [30]. Based on the consumer theory, the demand for a good depends on the price of the good and related goods, household income, and household members. In addition, the number of electrical appliances and temperature determine the electricity demand. Households with more electrical appliances and facing high temperatures will report higher consumption of electricity [31]. Different studies have found that electricity consumption in Mexico is mainly affected by income, and prices are weak because the electricity demand is very inelastic [19,32].

While different studies have found that electricity is a necessity, in the case of gasoline, it is a luxury. The demand for gasoline depends on the number of automobiles and income. Between 1991 and 2012, the number of automobiles rose from 6.6 to 22.4 million, representing an increase of 239% [33]. However, in Mexico, low-income households represent only 16% of the gasoline demand. Evidence finds that an increase in the gasoline price does not significantly affect its demand among high-income households. A policy of increases in the price of gasoline reduces only the consumption of low-income households [34]. Other estimations have found that the price elasticity of gasoline is more inelastic in the short run than in the long run [35].

Analyzing how food consumption is affected by changes in the price of other goods is essential for understanding the impact on households’ well-being. Although the energy reform is expected to reduce the prices of electricity, gas, and gasoline in the long run, negative impacts on households’ consumption might arise in the short run. Ref. [19] finds that in the face of increases in energy prices, Mexican households will reduce their food consumption to afford electricity, which mainly affects the poorest households. Gas and food are also complementary goods. When the price of gas increases, poor households are forced to look for substitute goods for gas. These households use alternatives to gas, like firewood or fuel oil, that could negatively affect health [36].

In sum, the demand for electricity, gas, and gasoline depends on the direct prices, prices of related goods, income, household characteristics, such as temperature, number of automobiles and electrical appliances, and household members. Also, it is expected that changes in the prices of electricity, gas, and gasoline affect food consumption.

3. Materials and Methods

This longitudinal, non-experimental, quantitative research assesses the impact of the 2013 Mexican Energy Reform on household energy consumption before and after its announcement. Household energy consumption data was analyzed from the National Survey of Income and Expenses (ENIGH by its name in Spanish) from 2012, 2014, 2016, and 2018 [37,38,39,40]. ENIGHs are conducted by the National Institute of Statistics and Geography (INEGI by its name in Spanish).

ENIGH experts create the sampling design and follow the probabilistic, stratified, two-stage, and cluster methods, which are defined as:

(a)

Probabilistic: the sampling units have a known and non-zero probability of being selected.

(b)

Stratified: the sampling units are classified according to socioeconomic characteristics.

(c)

Two-stage: the ultimate sampling unit (household) is selected in two stages.

(d)

Clusters: the sampling units are sets of elementary units with heterogeneous characteristics inside and homogeneous outside.

The sampling size per year is shown in

Table 1. Sampling size per year.

Year	Households
2012	64,246
2014	41,427
2016	82,718
2018	87,826

Source: (ENIGH, 2012, 2014, 2016, and 2018).

.

Energy, gas, and gasoline prices were taken from the Energy Regulatory Commission, CFE, and Pemex. The null and alternative hypotheses were formulated as follows:

H0: 

Mexican households did not increase their electricity, gas, and gasoline consumption after the price increase due to the 2013 energy reform.

H1: 

Mexican households increased their electricity, gas, and gasoline consumption after the price increase due to the 2013 energy reform.

Quantile regression was used to estimate the relationship between the following independent variables and the specific percentiles of the dependent variable (see Equation (1)):

• Dependent variable:
  • y1= Electricity, gas, and gasoline consumption
• Independent variables:
  • x1 = Price of electricity, gas, and gasoline;
  • x2 = Price of other basic goods;
  • x3 = Number of household members;
  • x4 = Household income;
  • x5 = Number of electrical appliances in the home;
  • x6 = Number of automobiles per household;
  • x7 = Average monthly temperature by state of the republic.

yi = xi β + u i

(1)

where:

• yi = the dependent variable;
• xi = the i-th element of the matrix of independent variables;
• β = the parameter to be estimated and the slope of the regression line that relates (yi) and (xi);
• u i = the random disturbance that includes all those factors other than the variables xi that influence n yi.

Data analysis was elaborated using the statistical software Stata. The quantile regression considered the following models to examine household electricity, gas, gasoline, and food consumption. The quantile regression method is an alternative to ordinary least squares regression (OLS) since it is less robust due to the sensitivity of its estimates to the presence of heteroskedasticity, structural changes, and atypical and extreme data, which are frequent with survey data, and more related to electrical energy consumption. Unlike the OLS estimator, quantile regression uses the estimate of least absolute deviations (LAD), also known as least absolute errors, establishing a relationship between the regressors and the endogenous variable without restricting random disturbance. It is considered a semiparametric technique. The LAD estimator has the advantage of not being restricted to the normality of the residuals [41]. The quantile regression method allows for estimating regression lines via quantiles of the dependent variable without the inconvenience of being considerably affected by the presence of heteroskedasticity, atypical data, or structural changes as with the OLS, LPS, 2SLS, and GMM estimator since it penalizes the errors linearly and not in a quadratic manner, in this way it is possible to analyze different segments of the distribution of the dependent variable [42]. Likewise, the selection bias due to sample segmentation indicated by [43] is avoided since it uses the deviations concerning the quantiles to weight the regression line in each quantile, thus avoiding sample segmentation. Although not knowing the distribution of the error limits the inference process, for large samples the estimated coefficients are distributed asymptotically like a normal distribution [41]. Furthermore, quantile regression has comparable efficiency to OLS for Gaussian linear models, while it outperforms OLS estimates in non-Gaussian linear models [44].

• Model of the Consumption of Electrical Energy:

ln (electricity consumption expenditure) i = β0+ β1ln (energy price) i + β2ln (L.P. gas price) i + β3ln (gasoline price) i + β4ln (food price) i + β5ln (household income) i + β6ln (number of household members) i + β7ln (number of electrical appliances) i + β8ln (temperature) i + є i

• Model of the Consumption of Liquefied Petroleum Gas:

ln (L.P. gas consumption expenditure) i = β0+ β1ln (energy price) i + β2ln (L.P. gas price) i + β3ln (gasoline price) i + β4ln (food price) i + β5ln (household income)) i + β6ln (number of household members) i + β7ln (number of electrical appliances) i + β8ln (temperature) i + є i

• Model of the Consumption of Gasoline:

ln (gasoline consumption expenditure) i = β0+ β1ln (energy price) i + β2ln (L.P. gas price) i + β3ln (gasoline price) i + β4ln (food price) i + β5ln (household income) i + β6ln (no. of household members) i + β7ln (no. of cars) i + є i

• Model of the Consumption of Food:

ln (food consumption expenditure) i = β0+ β1ln (energy price) i + β2ln (L.P. gas price) i + β3ln (gasoline price) i + β4ln (food price) i + β5ln (household income) i + β6ln (no. of household members) i + β7ln (number of electrical appliances) i + є i

Figure 1 shows a schematic representation of the research algorithms.

4. Results

Table 2. Descriptive statistics.

Variable	2012				2014				2016				2018
	Mean	Standard Deviation	Pr (Skewness)	Pr (Kurtosis)	Mean	Standard Deviation	Pr (Skewness)	Pr (Kurtosis)	Mean	Standard Deviation	Pr (Skewness)	Pr (Kurtosis)	Mean	Standard Deviation	Pr (Skewness)	Pr (Kurtosis)
ln (gas consumption)	0.0586	0.0067	0.0000	0.0000	0.0365	0.0375	0.0000	0.0000	6.1160	1.6153	0.0000	0.0000	0.0388	0.0386	0.0865	0.000
ln (food Price)	3.2447	0.6574	0.0000	0.0000	3.1144	0.6267	0.0000	0.0000	3.0304	0.6873	0.0000	0.0000	3.0298	0.6687	0.0000	0.0000
ln (energy consumption)	0.0626	0.0099	0.0000	0.0000	0.0558	0.0250	0.0000	0.0000	0.0518	0.0210	0.0000	0.0000	0.0555	0.0240	0.0000	0.0000
ln (gasoline consumption)	0.0296	0.0024	0.0000	0.0000	0.0180	0.0143	0.0000	0.0000	0.0265	0.0436	0.0000	0.0000	0.0285	0.0449	0.0000	0.0000
ln (food consumption)	6.5728	0.4348	0.0000	0.0000	5.4652	0.7168	0.0000	0.0000	5.5593	0.7604	0.0000	0.0000	5.6877	0.7440	0.0000	0.0000
ln (gasoline price)	0.0115	0.0761	0.0000	0.0000	0.0175	0.0634	0.0000	0.0000	0.0147	0.0602	0.0000	0.0000	0.0108	0.0580	0.0000	0.0000
ln (energy price)	0.1477	0.0042	0.0000	0.0000	0.1976	0.0587	0.0000	0.0000	0.1808	0.0548	0.0000	0.0000	0.3892	0.0630	0.0000	0.0000
ln (gas price)	0.0074	0.0006	0.0000	0.0000	0.0033	0.0162	0.0000	0.0000	2.5260	2.4322	0.0000	0.0000	0.0025	0.0271	0.0000	0.0000
ln (household members)	1.2307	0.5579	0.0000	0.0000	1.1908	0.5543	0.0000	0.3850	1.1582	0.5592	0.0000	0.0000	1.1379	0.5651	0.0000	0.0000
ln (household income)	9.9310	1.0031	0.0000	0.0000	10.2361	0.7803	0.0000	0.0000	10.3052	0.7797	0.0000	0.0000	10.4153	0.7817	0.4085	0.0000
ln (electrical appliances)	2.0373	0.5884	0.0000	0.0000	2.0102	0.6159	0.0000	0.0000	2.0473	0.5936	0.0000	0.0000	2.0146	0.5959	0.0000	0.0000
ln (temperature)	3.1262	0.1789	0.0000	0.0000	3.1268	0.1854	0.0000	0.0000	3.1220	0.1924	0.0000	0.0000	3.1152	0.1760	0.0000	0.0000
ln (automobiles)	0.1028	0.2845	0.0000	0.0000	0.1036	0.2828	0.0000	0.0000	0.1148	0.2986	0.0000	0.0000	0.1248	0.3098	0.0000	0.0000

Raw Data from ENIGH (2012, 2014, 2016, and 2018).

presents the descriptive statistics of the variables.

As can be seen, electricity and gasoline prices increased after the reform in 2013, accompanied by reduced consumption. The increase in the price of gas was observed in 2016. In turn, food prices have been stable from 2012 to 2018. From 2012 to 2018, real household income and the number of automobiles have increased. Appendix A.1 shows the correlation matrix of the variables in the model. As expected, the consumption is negatively related to the direct prices during the four years. Also, a strong and positive relationship exists between the consumption of all goods and household income. Household income positively correlates with household members, automobiles, and electrical appliances. The variance inflation factor (VIF) is presented in Appendix A.2. The value of VIF for all variables is close to 1, indicating no heteroskedasticity problem.

Table 3. Quantile regression of the consumption of electricity.

Year/Variable	2012			2014			2016			2018
	Q25	Q50	Q75	Q25	Q50	Q75	Q25	Q50	Q75	Q25	Q50	Q75
ln energy price	−0.002	−0.004 *	0.000	0.001	−0.002	0.001	−0.628 **	−0.225 **	−0.636 **	−0.027 **	−0.026 **	−0.028 **
	(−0.31)	(−2.29)	(0.07)	(0.39)	(−1.17)	(0.33)	(−4.59)	(−3.25)	(−7.92)	(−18.25)	(−33.72)	(−30.13)
ln liquefied petroleum gas price	0.003	−0.004 *	−0.007 **	0.063 **	0.035 **	0.033 **	−0.023 **	−0.010 **	−0.009 **	0.025 **	0.019 **	0.022 **
	(0.59)	(−2.08)	(−3.03)	(7.70)	(7.24)	(5.73)	(−12.63)	(−8.30)	(−5.59)	(9.69)	(11.75)	(19.99)
ln gasoline price	−0.013 **	0.000	−0.000	−0.006 **	−0.003 **	−0.001	−0.157 **	−0.111 *	−0.083	−0.008 **	−0.006 **	−0.005 **
	(−3.17)	(0.12)	(−0.05)	(−3.29)	(−4.15)	(−0.66)	(−3.05)	(−2.43)	(−0.81)	(−5.25)	(−10.99)	(−6.95)
ln food price	0.003 **	0.001 **	0.000	0.001 **	0.000	−0.000	0.048 **	−0.003	−0.040 **	0.001 **	0.000 **	−0.000
	(4.08)	(3.47)	(0.72)	(4.25)	(1.05)	(−0.53)	(3.96)	(−0.54)	(−5.64)	(8.21)	(3.94)	(−0.39)
ln household income	−0.001	0.003 **	0.004 **	0.002 **	0.004 **	0.005 **	0.177 **	0.297 **	0.398 **	0.002 **	0.003 **	0.004 **
	(−1.20)	(16.48)	(16.13)	(13.80)	(23.75)	(26.08)	(15.03)	(48.32)	(51.78)	(16.09)	(48.32)	(51.14)
ln household members	0.003 **	0.002 **	0.001 **	0.002 **	0.002 **	0.001 **	0.214 **	0.135 **	0.094 **	0.002 **	0.001 **	0.001 **
	(4.15)	(8.10)	(3.51)	(10.12)	(11.06)	(6.18)	(18.77)	(14.26)	(7.47)	(15.29)	(16.72)	(10.36)
ln electrical appliances	0.022 **	0.008 **	0.012 **	0.011 **	0.007 **	0.006 **	1.113 **	0.657 **	0.555 **	0.012 **	0.007 **	0.006 **
	(16.59)	(26.97)	(29.45)	(23.42)	(45.46)	(26.47)	(35.64)	(61.17)	(82.00)	(32.07)	(51.28)	(56.59)
ln temperature	−0.001	0.003 **	0.006 **	−0.003 **	0.003 **	0.005 **	0.514 **	0.518 **	0.592 **	0.005 **	0.006 **	0.006 **
	(−0.76)	(4.23)	(7.43)	(−4.30)	(6.18)	(9.13)	(19.47)	(28.94)	(19.55)	(9.77)	(22.48)	(25.51)
Constant	0.005	0.002	−0.008 **	0.004	−0.002	−0.009 **	−1.166 **	−0.525 **	−0.953 **	−0.003	0.002 *	−0.001
	(0.86)	(0.67)	(−2.62)	(1.04)	(−1.02)	(−3.70)	(−8.70)	(−6.90)	(−9.67)	(−1.49)	(2.41)	(−0.90)
Pseudo R2	0.133	0.123	0.141	0.086	0.096	0.118	0.097	0.098	0.116	0.094	0.098	0.122

Note: for t-test * p-value < 0.05, ** p-value < 0.01. Raw Data from ENIGH (2012, 2014, 2016, and 2018).

shows the quantile regression of consumption spending according to electricity, L.P. gas, gasoline, and food. The four years 2012, 2014, 2016, and 2018 are divided into 25, 50, and 75% quantiles. In 2012, before the enactment of the 2013 energy reform, the energy price presented a price elasticity of demand (EPD) with a significant negative effect in Q50 (−0.004), which explains that households with an average energy consumption show an inelastic demand, and since its elasticity is less than one, this suggests that consumers are less sensitive to changes in the price of energy due to the quantity demanded. While no statistically significant results were obtained in 2014, the price elasticity of demand for 2016 and 2018 in Q25, Q50, and Q75 presents a highly inelastic elasticity based on the statistically significant estimates in the consumption of electrical energy, which indicates that when energy prices increase, consumption decreases, but to a lesser extent.

The income elasticity of demand for 2012 in Q50 and Q75, coefficients 0.003 and 0.004, respectively, are significant and positive, indicating that electricity is a normal good that is necessary for these households. Concerning 2014, 2016, and 2018, the three quantiles (Q25, Q50, and Q75) also show a positive elasticity, which indicates that electricity is a necessary normal good and not a luxury one since the values are less than one. Therefore, an inelastic income elasticity means that a change in household income does not increase energy consumption.

Regarding the number of household members, this variable shows a significant (<0.01) and positive relationship with electricity consumption in the three quantiles during the four years. Similarly, a significant and positive effect is observed for the number of electrical appliances in the home. That is, an increase in the number of household members and appliances is associated with increased electricity consumption in all quantiles. For 2016, the estimated coefficients and t statistics are high, indicating a strong relationship.

The temperature variable has a significant effect and a positive relationship in the three quantiles, which suggests that it influences the consumption of electrical energy in a positive way; when the temperature increases, it will be reflected in an increase in energy demand, except for Q25 in 2014, which was negative. This finding can be explained because the price of electricity in Mexico increases for ranges of consumption, so the higher the consumption, the greater the monthly balance, regardless of the socioeconomic level of the home. Therefore, despite inclement temperatures, low-income households may be forced to decide between maintaining their basic energy needs and ensuring the well-being of their families. This argument augments evidence reported in previous studies [45,46,47] that indicate welfare losses due to an increase in energy prices tend to be more severe in low-income households than in wealthier ones.

The L.P. gas price variable during 2012 and 2016 in Q25, Q50, and Q75 presented negative estimates; therefore, this variable gives a negative cross-price elasticity of demand (EPCD) and behaves as a luxury good. During 2014 and 2018, in the three quantiles it can be observed that, on the contrary, it presented a positive relationship, which shows a positive cross-elasticity and behaves as a substitute good; that is, L.P. gas or electricity can be consumed indistinctly. The results of L.P. gas for 2014, 2016, and 2018 indicate an inelastic elasticity, showing that a price change does not practically affect the demand for L.P. gas consumption. It is essential for households with few substitutes or no viable options to switch to other products.

Table 4. Quantile regression of liquefied petroleum gas.

Year/Variable	2012			2014			2016			2018
	Q25	Q50	Q75	Q25	Q50	Q75	Q25	Q50	Q75	Q25	Q50	Q75
ln energy price	−2.29 e−17	−0.010	−0.002	−1.49 e−15	−9.09 e−06	0.010 **	−5.11 e−13 **	−3.62 e−13 *	−2.53 e−13	2.64 e−15	−0.001	0.005 **
	(−0.05)	(−1.08)	(−0.53)	(−1.00)	(−0.01)	(4.21)	(−3.94)	(−2.33)	(−0.00)	(0.46)	(−0.34)	(4.41)
ln liquefied petroleum gas price	−0.766 **	−0.023	−0.007	−2.408 **	−2.468 **	−0.087 **	−0.307 **	−0.383 **	−0.456 **	−1.134 **	−0.255 **	−0.022 **
	(−32.29)	(−0.45)	(−1.49)	(−2660.64)	(−16.44)	(−14.02)	(−3.8 e+12)	(−1158.52)	(−2872.75)	(−171.10)	(−9.19)	(−10.35)
ln gasoline price	6.90 e−16 **	−0.012	−0.009 *	−1.87 e−16	−0.001	−0.001	−1.86 e−13 **	−1.38 e−13	−2.17 e−13	−3.33 e−15	−0.048 **	0.000
	(2.85)	(−0.59)	(−1.97)	(−0.51)	(−0.26)	(−0.52)	(−2.88)	(−1.63)	(−0.00)	(−1.14)	(−14.51)	(0.24)
ln food price	−3.91 e−17	0.009 **	0.006 **	1.97 e−16	0.001	0.004 **	1.33 e−13 **	1.24 e−13 **	9.91 e−14	6.96 e−16	0.009 **	0.003 **
	(−1.55)	(8.34)	(6.13)	(1.87)	(1.61)	(14.03)	(3.97)	(2.85)	(0.00)	(1.92)	(20.36)	(26.82)
ln household income	−8.48 e−17 **	0.011 **	0.002 **	4.55 e−16 **	0.001	0.002 **	6.44 e−14 **	8.88 e−14 **	1.72 e−14	−1.88 e−16	0.006 **	0.003 **
	(−3.19)	(14.39)	(4.28)	(5.31)	(1.32)	(9.63)	(3.92)	(5.86)	(0.00)	(−0.57)	(13.47)	(25.13)
ln household members	1.56 e−17	0.003 *	−0.000	1.40 e−17	0.001	0.001 **	2.47 e−14	9.46 e−14 *	4.43 e−14	−5.27 e−16	0.010 **	0.001 **
	(0.77)	(2.13)	(−0.03)	(0.06)	(1.55)	(3.34)	(0.75)	(2.15)	(0.00)	(−0.66)	(15.58)	(9.23)
ln electrical appliances	1.30 e−16 **	0.010 **	0.017 **	−1.14 e−15 **	0.001	0.008 **	1.75 e−15	−5.14 e−14	1.06 e−13	3.21 e−16	0.011 **	0.008 **
	(4.68)	(8.51)	(8.46)	(−6.49)	(1.66)	(18.03)	(0.03)	(−0.66)	(0.00)	(0.47)	(22.30)	(34.63)
ln temperature	−1.62 e−16	−0.078 **	−0.022 **	−1.63 e−15	−0.019	−0.013 **	−5.76 e−13 **	−5.68 e−13 *	−4.67 e−13	−8.64 e−15	−0.078 **	−0.014 **
	(−1.52)	(−12.97)	(−9.38)	(−1.47)	(−1.84)	(−17.27)	(−2.86)	(−2.01)	(−0.00)	(−1.60)	(−30.00)	(−46.26)
Constant	0.006 **	0.122 **	0.066 **	0.016 **	0.068 **	0.060 **	6.364 **	6.733 **	7.090 **	0.007 **	0.165 **	0.061 **
	(31.82)	(5.52)	(18.34)	(2648.37)	(2.58)	(39.32)	(2.4 e +13)	(4181.47)	(3393.76)	(169.42)	(15.74)	(50.25)
Pseudo R2	0.025	0.159	0.068	0.208	0.224	0.038	0.406	0.502	0.287	0.089	0.144	0.040

Note: for t-test * p-value < 0.05, ** p-value < 0.01. Raw Data from ENIGH (2012, 2014, 2016, and 2018).

shows the case of the price of L.P. gas for the following years: 2012 in Q25; 2014 in Q75; 2016 in the three quantiles (Q25, Q50, and Q75); and 2018 in Q50 and Q75. All show a negative coefficient of less than one, which indicates an inelastic demand elasticity, showing that a price change does not practically affect the demand for L.P. gas consumption. On the contrary, for the years 2014 in Q25 and Q50 and 2018 in Q25, the coefficients are statistically significant, negative, and greater than one, which shows an elastic demand, which means that for these years in these households, a change in the price of L.P. gas causes a difference in the quantity demanded.

The gasoline price variable in 2012 for Q25 displays an estimate that is statistically significant and positive. In this case, gasoline is considered a substitute good; as the price of gasoline increases, L.P. gas could be consumed. On the other hand, the price of gasoline for the year 2012 in Q75, for 2016 in Q25, and 2018 in Q50 shows a negative and significant effect; that is, it behaves as a luxury good. It should also be mentioned that no statistically significant results were obtained during 2014.

The income elasticity of demand with L.P. gas consumption is statistically significant, positive, and less than one for the following years: 2012 in Q50 and Q75; 2014 in Q25 and Q75; 2016 in Q25 and Q50; and finally in 2018 in Q50 and Q75, which indicates an inelastic elasticity, therefore L.P. gas is a necessary normal good. On the contrary, in 2012, in Q25, it presented a negative significant estimate of less than one. In this case, there is an inelastic income elasticity, and L.P. gas is considered an inferior good.

Statistically significant results were obtained for household members in 2012 in Q50, 2014 in Q75, 2016 in Q50, and 2018 in Q50 and Q75. This finding suggests that the more members inhabit a household, the greater L.P. gas consumption. Concerning the variable of the number of household appliances for the years 2012 in the three quantiles, for 2014 only in Q75 and for 2018 in Q50 and Q75, they present significant positive coefficients, which indicates that an increase in the use of household appliances is associated with the consumption of L.P. gas, except Q25 in 2014, which presents a negative coefficient, that is, in homes with lower consumption of L.P. gas, it does not affect them if they increase the number of electrical appliances. During 2016, no statistically significant results were obtained.

The results of L.P. gas for 2014, 2016, and 2018 indicate an inelastic elasticity, showing that a price change practically does not affect the demand for L.P. gas consumption. It is essential for households with few substitutes or no viable options to switch to other products. For 2014 in Q25 and Q50 [−2.408 and −2.468], the coefficients indicate that households with low and medium consumption of L.P. gas are more sensitive to price changes, which implies that an increase in the price of L.P. gas would cause a reduction in the amount consumed by these households. It is worth mentioning that for this year, 2014, after the reform, L.P. gas prices had been allowed to be freely established.

It should be noted that in Mexico, many homes (79%) use L.P. gas as an energy source [48]. The inelasticity of energy products will not significantly affect their consumption; however, to maintain this consumption, households may be forced to sacrifice other categories of expenses or their basic needs, such as food or health services. This situation allows us to notice how this could affect the well-being of households, as they would have to adjust their budgets and priorities to deal with increases in the price of L.P. gas. For the latter, public policies and regulatory measures must consider the differential impact these price increases can have on households, especially those with lower incomes, to guarantee the well-being of the entire population.

The price of energy for 2012 and 2014 significantly positively affected gasoline consumption in Q25, Q75, and Q75, respectively, showing a positive cross-price elasticity of demand; that is, energy is a substitute for households with the lowest and highest gasoline consumption. On the contrary, a negative effect is shown in 2012 for Q50; in this case, for households with average energy consumption, it represents a luxury good, which means that they consume both products. This same negative effect occurred in 2016 for Q75 and 2018 for Q50 and Q75.

The L.P. gas price during 2014 in Q75 and 2018 in Q50 and Q75 presented positive estimates; therefore, this variable gives a positive cross-price elasticity of demand (EPCD) and behaves as a substitute good, that is, L.P. gas or gasoline can be consumed interchangeably. During 2016 in Q75 and 2018 in Q25, it can be observed that, on the contrary, the estimates present a negative relationship, which shows a negative cross-elasticity and L.P. gas behaves as a luxury good. However, for 2012, no statistically significant results were obtained.

The results of the gasoline price variable are statistically significant in the three quantiles (Q25, Q50, and Q75) during the four years, indicating an inelastic price elasticity of demand, on the contrary in Q25 and Q50 in the years 2014 and 2016, the coefficients are greater than one, therefore when gasoline prices increase, its consumption is negatively affected. The analysis for the food price indicates that during 2012, 2016, and 2018, only in Q75 did the results show a positive cross-elasticity; in this case, food behaves as a substitute for households with higher gasoline consumption. However, in 2014, no statistically significant results were obtained.

The income elasticity of gasoline demand for 2012 and 2014 in Q25 and Q75, and in 2016 only Q75, presented significant positive coefficients, which indicates that gasoline is a necessary normal good for these households. Concerning 2018, the three quantiles show positive significant coefficients only in Q50 and Q75, which suggests that gasoline is an essential normal good and not a luxury one since the values are less than one; on the contrary, in Q25, the coefficient is negative, which represents an inverse relationship between income and the demand for gasoline in households with the lowest consumption; in this case, it would be considered an inferior good.

The variable number of household members in Q75 in the four years are statistically significantly positive; that is, an increase in household members increases gasoline consumption. For 2016 and 2018, in Q50, the coefficients are negative, indicating that an increase in members does not affect or decrease gasoline consumption. The variable of the number of cars in the four years for the three quantiles show that they are statistically significantly positive, that is to say that, in households, when the number of vehicles in the home increases, they demand a greater consumption of gasoline, except for in Q50 in the year 2012, as this coefficient is not significant.

Table 5. Quantile regression of gasoline.

Year/ Variable	2012			2014			2016			2018
	Q25	Q50	Q75	Q25	Q50	Q75	Q25	Q50	Q75	Q25	Q50	Q75
ln energy price	6.18 e−16 *	−6.08 e−16 *	0.012 **	−3.29 e−16	2.18 e−15	0.018 **	−2.47 e−14	2.15 e−14	−0.083 **	2.30 e−14	−0.029 **	−0.034 **
	(2.27)	(−2.22)	(11.56)	(−0.25)	(1.50)	(9.39)	(−1.35)	(1.31)	(−20.33)	(1.92)	(−6.76)	(−16.99)
ln liquefied petroleum gasprice	−1.19 e−16	1.13 e−16	−0.000	5.30 e−15	−2.91 e−15	0.018 **	6.31 e−16	−5.15 e−16	−0.000 **	−4.17 e−14 **	0.034 **	0.042 **
	(−0.98)	(0.52)	(−0.33)	(1.41)	(−0.88)	(5.17)	(1.62)	(−1.86)	(−15.89)	(−2.77)	(5.20)	(13.77)
ln gasoline price	−0.478 **	−0.480 **	−0.006 **	−1.052 **	−1.100 **	−0.011 **	−3.966 **	−4.051 **	−0.029 **	−0.209 **	−0.120 **	−0.030 **
	(−8.31)	(−2.67)	(−2.84)	(−27.35)	(−16.87)	(−13.15)	(−25.42)	(−20.81)	(−18.30)	(−6.17)	(−6.54)	(−16.42)
ln food price	8.95 e−18	−1.74 e−17	0.001 **	−3.14 e−17	−8.44 e−17	0.000	8.34 e−16	1.12 e−16	0.001 **	2.69 e−17	0.000	0.001 **
	(0.80)	(−1.02)	(6.57)	(−0.37)	(−1.00)	(0.19)	(1.00)	(0.16)	(5.54)	(0.07)	(1.50)	(9.49)
ln household income	9.06 e−17 *	5.96 e−17	0.003 **	1.14 e−15 **	−4.81 e−16	0.004 **	2.52 e−15	9.52 e−16	0.017 **	−8.01 e−15 **	0.005 **	0.018 **
	(2.00)	(1.10)	(32.63)	(3.46)	(−1.84)	(21.27)	(0.98)	(0.29)	(31.96)	(−2.70)	(5.77)	(50.39)
ln household members	9.24 e−18	1.99 e−17	0.004 **	2.49 e−17	2.20 e−16 *	0.001 **	−9.94 e−16	−3.20 e−15 *	0.002 **	−1.39 e−15	−0.000 **	0.002 **
	(0.94)	(1.76)	(3.61)	(0.19)	(2.06)	(2.90)	(−0.95)	(−2.48)	(6.61)	(−1.50)	(−4.05)	(8.19)
ln automobiles	0.002 **	0.004	0.002 **	0.008 **	0.010 **	0.004 **	0.017 **	0.025 **	0.012 **	0.079 **	0.106 **	0.011 **
	(2.74)	(1.00)	(13.03)	(9.70)	(6.42)	(12.91)	(5.94)	(5.91)	(44.46)	(92.25)	(44.56)	(43.68)
Constant	0.007 **	0.007 **	−0.026 **	0.017 **	0.018 **	−0.047 **	0.055 **	0.057 **	−0.134 **	−0.009 **	−0.045 **	−0.110 **
	(8.29)	(2.66)	(−29.13)	(22.08)	(13.77)	(−22.79)	(20.83)	(17.13)	(−20.01)	(−21.77)	(−7.12)	(−28.57)
Pseudo R2	0.250	0.278	0.164	0.253	0.279	0.174	0.255	0.282	0.111	0.153	0.230	0.112

Note: for t-test * p-value < 0.05, ** p-value < 0.01. Raw Data from ENIGH (2012, 2014, 2016, and 2018).

shows the results for the price of food in Q25 and Q50 of the years 2012 and 2018, indicating a positive cross-elasticity, which is why food behaves as a substitute good. In 2014 and 2016, the results presented a positive elasticity but only for Q25; on the contrary, in 2016, for Q75, the elasticity was negative; therefore, it would be considered a luxury good.

Gasoline consumption expenditure presents the same behavior as L.P. gas, with a negative coefficient for households in the three quantiles for the three periods analyzed after the reform. Therefore, this means that households continue buying gasoline despite the price changes; it should be noted that for Q25 and Q50 in the years 2014 and 2016, the coefficients are greater than one; therefore, the price increases in those years had a more significant impact on gasoline consumption; lower-income households are more sensitive to price increases as they allocate a greater proportion of their budget to spending on fuel. The price elasticity of energy demand can vary depending on the availability of options, energy infrastructure, and consumer preferences. In many situations, especially in the short term, energy demand tends to be inelastic due to its essential nature in the lives of households in Mexico.

Table 6. Quantile regression of food.

Year/ Variable	2012			2014			2016			2018
	Q25	Q50	Q75	Q25	Q50	Q75	Q25	Q50	Q75	Q25	Q50	Q75
ln energy price	−0.012	−0.038	−0.026	0.008	−0.034	−0.048	0.151 **	0.073 *	0.100 *	0.058	0.044	0.053
	(−0.19)	(−0.48)	(−0.36)	(0.16)	(−0.58)	(−0.69)	(3.57)	(2.44)	(2.23)	(1.94)	(1.13)	(1.47)
ln liquefied petroleum gas price	−0.289	−0.256 *	−0.132	−1.893 **	−1.671 **	−1.273 **	−0.009 **	−0.006 **	−0.003 **	−0.525 **	−0.240 **	0.021
	(−1.37)	(−2.10)	(−0.99)	(−10.16)	(−9.80)	(−7.48)	(−11.81)	(−8.19)	(−5.24)	(−11.11)	(−3.74)	(0.38)
ln gasoline price	−0.001	0.075	0.124 **	0.069	0.046	0.005	0.066 *	0.044	0.066 *	0.070 **	0.064 **	0.028
	(−0.01)	(0.83)	(2.58)	(1.84)	(0.85)	(0.10)	(2.23)	(1.50)	(2.36)	(2.76)	(3.34)	(1.05)
ln food price	−0.113 **	−0.174 **	−0.258 **	−0.103 **	−0.169 **	−0.250 **	−0.118 **	−0.189 **	−0.280 **	−0.114 **	−0.182 **	−0.271 **
	(−15.18)	(−15.95)	(−24.23)	(−12.83)	(−20.27)	(−34.24)	(−33.15)	(−50.53)	(−74.03)	(−29.05)	(−40.67)	(−50.92)
ln household income	0.119 **	0.160 **	0.181 **	0.148 **	0.184 **	0.201 **	0.139 **	0.170 **	0.187 **	0.136 **	0.167 **	0.190 **
	(12.82)	(20.84)	(31.92)	(30.12)	(42.42)	(60.10)	(39.69)	(47.91)	(74.60)	(72.92)	(79.66)	(54.29)
ln household members	0.244 **	0.193 **	0.181 **	0.205 **	0.172 **	0.162 **	0.206 **	0.174 **	0.154 **	0.206 **	0.181 **	0.165 **
	(23.72)	(25.28)	(20.95)	(36.28)	(22.26)	(25.62)	(56.80)	(50.09)	(39.07)	(53.67)	(66.74)	(61.70)
ln electrical appliances	0.124 **	0.087 **	0.072 **	0.145 **	0.116 **	0.099 **	0.139 **	0.118 **	0.111 **	0.129 **	0.113 **	0.097 **
	(8.06)	(7.38)	(6.48)	(17.08)	(17.59)	(14.20)	(27.11)	(29.67)	(24.52)	(33.78)	(33.57)	(27.25)
Constant	4.082 **	4.166 **	4.426 **	3.570 **	3.736 **	4.094 **	3.833 **	4.023 **	4.400 **	3.915 **	4.096 **	4.413 **
	(39.32)	(43.87)	(46.16)	(85.58)	(80.22)	(92.23)	(121.67)	(148.79)	(146.01)	(165.41)	(186.91)	(157.69)
Pseudo R2	0.099	0.115	0.134	0.115	0.128	0.149	0.104	0.121	0.150	0.105	0.120	0.144

Note: for t-test * p-value < 0.05, ** p-value < 0.01. Raw Data from ENIGH (2012, 2014, 2016, and 2018).

indicates that the income elasticity of demand for 2012 in Q50 and Q75, coefficients 0.003 and 0.004, respectively, are significant and positive, indicating that electricity is a normal good that is necessary for these households. Concerning 2014, 2016, and 2018, the three quantiles (Q25, Q50, and Q75) also show a positive elasticity, which indicates that electricity is a necessary normal good and not a luxury one since the values are less than one; therefore, an inelastic income elasticity means that a change in household income does not increase energy consumption.

The results obtained for the energy price variable present a statistically significant positive effect for 2016 in the three quantiles (Q25, Q50, and Q75); therefore, this variable indicates a positive cross-elasticity interpreted as a substitute good. No statistically significant results were obtained for 2012, 2014, and 2018. With the L.P. gas price variable in the three quantiles for the years 2016 and 2018, significant negative estimates of less than one are observed, which indicate an inelastic cross-elasticity and behave as a poorly complementary good. Likewise, in 2014, the estimates are also significant and negative but greater than one in the three quantiles. Therefore, the elasticity is elastic. The goods are highly complementary; that is to say, food and L.P. gas are used together to satisfy consumer demand.

For the price of gasoline, for 2012 in Q75, the results show a significant positive relationship, which indicates a positive cross-price elasticity and behaves as a substitute good, which explains that if gasoline prices increase, households could allocate more income to gasoline consumption, which would limit their ability to purchase food. This same result is shown in 2016 in Q25 and Q75, and in 2018 in Q25 and Q75. Meanwhile, during 2014, no statistically significant results were obtained.

For the variable of food prices in the four years and all quantiles, they show a negative and significant effect of less than one concerning food consumption, which indicates an inelastic demand. This finding represents that households are less sensitive to changes in food prices and food consumption. The income elasticity of demand for food consumption is statistically significant, positive, and less than one for the four years and the three quantiles. This result indicates that food is a normal good that is necessary for all households. For the variable of household members and household appliances, the results show significant positive coefficients in all quantiles for the four years, suggesting that households with more members and greater use of household appliances lead to increased food consumption.

Concerning food consumption and energy price, the coefficients were positive and statistically significant for 2016, indicating that the cross-price elasticity is positive for households in the three quantiles [0.151, 0.073, and 0.100]. This outcome means that food and energy consumption behave as substitute goods. Changes in consumption behavior indicate a surrogate relationship between food consumption and electricity prices. An increase in energy prices can lead to a decrease in demand for foods that require intensive cooking and electricity consumption as consumers look for options to save on energy costs. The increase in electricity prices can significantly impact household expenses, especially for those with lower incomes, as they may be forced to decide between maintaining their basic energy needs and ensuring their families’ well-being and access to adequate nutrition.

5. Discussion

The analysis of the elasticities concerning the consumption of energy, gas, gasoline, and food presents the price elasticities of demand by quantile from 2012 to 2018. The price elasticity of demand is an indicator that evaluates how the demanded quantity of a good changes when the price of this good exhibits a proportional variation. According to microeconomic theory, this elasticity is usually negative since the law of demand predicts that an increase in the price of a good generally results in a change in the quantity of that good or service that consumers could buy [49].

Table 7. Price elasticity of demand by quantile.

		Electricity	L.P. GAS	Gasoline	Foods
2012	Q25	−0.002	−0.766 **	−0.478 **	−0.113 **
	Q50	−0.004 *	−0.023	−0.480 **	−0.174 **
	Q75	0.000	−0.007	−0.006 **	−0.258 **
2014	Q25	0.001	−2.408 **	−1.052 **	−0.103 **
	Q50	−0.002	−2.468 **	−1.100 **	−0.169 **
	Q75	0.001	−0.087 **	−0.011 **	−0.250 **
2016	Q25	−0.626 **	−0.307 **	−3.966 **	−0.118 **
	Q50	−0.225 **	−0.383 **	−4.051 **	−0.189 **
	Q75	−0.636 **	−0.456 **	−0.029 **	−0.280 **
2018	Q25	−0.027 **	−1.134 **	−0.209 **	−0.114 **
	Q50	−0.026 **	−0.255 **	−0.120 **	−0.182 **
	Q75	−0.028 **	−0.022 **	−0.030 **	−0.271 **

Note: for t-test * p-value < 0.05, ** p-value < 0.01. Raw data from ENIGH (2012, 2014, 2016, and 2018).

allows us to observe that the consumption of electricity in the years 2016 and 2018, after the energy reform, show price elasticities between −0.026 and −0.636 for households in the three quantiles, which indicates that after the implementation of the energy reform, there was a notable increase in electricity prices, which reduced energy demand. These findings are consistent with previous studies carried out in Spain and Chile, which also found that changes in electricity prices significantly affect household demand for electricity [50,51]. Therefore, this research suggests that the 2013 energy reform implementation caused an increase in electricity prices, negatively impacting households by reducing their ability to pay and demand energy.

Findings also showed the negative elasticities of L.P. gas in households and gasoline consumption in the three quantiles for the years 2014, 2016, and 2018, which indicates that an increase in the price of those energy sources led to a decrease in their consumption.

The analysis of the trend in the prices of energy products before and after the energy reform determines some patterns. First, when examining electricity rates, a significant increase is observed between 2012 and 2018, ranging from 8% to 113% [52,53]. This finding indicates that consumers experienced a considerable increase in electricity costs during that period. This poses a dilemma for Mexicans: whether to reduce their energy consumption or allocate a significant portion of their income to pay electricity bills. Neither of these options has the potential to help achieve SDG 7. Although the first contributes towards reducing fossil fuels, it disproportionately affects low-income households, widening social inequalities. These families could have a more challenging time coping with increases in electricity rates; therefore, it does not ensure access to affordable energy for all. The second alternative is not encouraging because it lowers their ability to allocate resources to other essential needs, such as food, education, health, or housing. This situation leads to a more significant gap between those who can afford to pay for electricity and those who cannot. It is possible to conclude that both options are irreconcilable with SDG 7; at least as Mexico speeds up its transition to renewable and clean energy. With regard to this, distinguished scholars have published an array of sustainability solutions to increase the reliability and variability of renewable energy that Mexican policymakers could adopt to strengthen the current energy laws [54,55,56]. Still, this topic is beyond the scope of this article. Worse still, the 2013 Mexican Energy Reform has become politicized, and the media are is battlefield where politicians try to impose their political ideologies rather than not leave anyone behind [57].

Mexico is not the only country that has modified its energy laws. Many countries have found themselves in this situation, needing to increase their productivity or reduce their subsidies’ expenses. Several scholars have remarked that energy reforms worldwide are at a stage where a review of the energy sector’s effectiveness, efficiency, and performance is necessary [25,58,59,60,61]. Therefore, insights from this study might have relevant implications for policymakers worldwide and, hopefully, encourage them to rethink the direction of their energy policies toward ensuring affordable access to reliable and sustainable energy.

6. Limitation of the Study and Future Research

Some limitations in this study should be considered for understanding the results and designing future research. Firstly, a significant challenge was the lack of updated data related to energy consumption and rates on the Energy Information System page of the Ministry of Energy. Furthermore, there were no data at the regional or local level, complicating the identification of consumption patterns in specific geographic areas; this lack of detailed information can limit the understanding of energy trends in different regions. Another challenge concerns the bias in the selection of the National Survey of Household Income and Expenditures (ENIGH) sample since this sample may not adequately reflect the diversity of energy regions in Mexico.

During the research design we considered incorporating monthly average temperature records [62] to facilitate the model’s accurate representation of seasonality. The rationale behind this approach is that cities with extreme temperatures would be grouped into categories with a greater prevalence of warm climates and in which temperature affects energy consumption. Otherwise, extreme temperatures could have an impact on the consistency of the estimators. Future research could consider the effects of extreme temperature variations on the demand for energy in some Mexican regions. Finally, the role of the 2013 Mexican Energy Reform regarding the development of renewable energies and energy efficiency in Mexican homes needs to be analyzed.

7. Conclusions

In Mexico, the energy sector has undergone transformations over the years; it was unquestionable that the industry needed changes in the division of hydrocarbons, electricity, and renewable energies from the extraction, transmission, and supply of energy products such as electricity, gas, and gasoline.

In light of our findings, the practical significance of this study is in the insights provided to scholars and policymakers to meet SDG 7. The analysis of the elasticities concerning the consumption of energy, gas, gasoline, and food presents the price elasticities of demand by quantile from 2012 to 2018, suggesting that the 2013 Mexican Energy Reform led to an increase in energy prices that, on the one hand, reduced the consumption of energy generated with fossil hydrocarbons but, on the other hand, affected the portion of the population with less income.

The results obtained by estimating the price elasticity of energy products, electricity, gas, and gasoline, show that the demand for energy in households is generally inelastic, which means that price changes have a relatively small impact on the quantity demanded and that households are less sensitive to variations in the price of energy, so demand does not vary significantly in response to changes in price; however, price increases after the energy reform had an impact on their consumption in low-income households since it allocated a more significant proportion of their budget to their energy expenses. The findings in the case of food prices for all years before and after the reform show that low-income households are more susceptible to changes in energy prices; the variations in energy costs impact these households more, and they find themselves in a difficult situation when having to decide between meeting their basic needs such as food and health or meeting their electricity costs.

This research suggests that this situation leads to a more significant gap between those who can afford to pay for electricity and those who cannot. Hence, it does not ensure access to affordable energy for all. Consequently, it is possible to conclude that the 2013 Mexican Energy Reform is irreconcilable with SDG 7 unless substantial additional efforts to leave no one behind are made. Still, its benefits can improve by speeding up the country’s transition to renewable and clean energy. Finally, the 2013 Mexican Energy Reform must be depoliticized; a successful and lasting reform requires a depoliticized mechanism that enhances meeting SDG 7. Regrettably, it does not seem feasible for the short term because the 2013 Mexican Energy Reform remains a critical area of debate among politicians.