Introduction

Maize (Zea mays) is the most widely produced agricultural commodity in the world; its production is divided into white maize, used for human consumption, and yellow, used for the production of animal protein and the obtaining of biofuels; particularly, yellow maize is one of the most important products in international markets (^{Simón & Golik, 2018}). According to the Agri-food and Fisheries Information Service (SIAP) and the Ministry of Agriculture, Livestock, Rural Development, Fisheries and Food (SAGARPA) now Secretary of Agriculture and Rural Development (SADER), in 2017, Mexico imported almost one million tons of yellow maize from the United States (^{SIAP-SAGARPA, 2017}). Meanwhile, the national offer is made up of 3.2 million farmers who grow maize on 7.5 million hectares and produce 27.8 million tons (^{INEGI, 2017}), these figures make maize the crop that occupies the largest area in the country. In addition, its importance extends due to the amount of labor employed, especially in areas where the population with the greatest vulnerability and poverty is found (^{Cruz-Delgado, 2009}).

With the disappearance in 1999 of the National Company of Popular Subsistence (CONASUPO), a parastatal that guaranteed the purchase and regulation of maize prices in Mexico, there was a restructuring of the maize market. Now, instead of selling to the parastatal company, 51.5 % of maize farmers in Mexico sell to intermediaries, 23 % to an industry or collection center, and only 25.5 % have direct dealings with end customers (^{INEGI, 2017}). The information above shows that a significant number of maize farmers have difficulty to access directly to more profitable markets. The farm support instrument that replaced CONASUPO was a program of direct support to farmers that until 2018 had the name of PROAGRO-Productivo, which was the farm support program with the highest budget in Mexico and was led by the International Center of Maize and Wheat Improvement (CIMMYT). PROAGRO-Productivo aimed to increase the productivity of production units in priority crops, including maize, through economic incentives delivered to 1) self-consumption farmers that have five hectares or less; 2) farmers in transition with between five and 20 hectares; and 3) commercial farmers with properties greater than 20 and up to 80 hectares (^{SAGARPA, 2018}). However, the intervention has not been remarkably reflected in productivity. Some authors (^{Maximiliano-Martínez et al., 2011}) attribute it to the fact that it was assumed that the productive units of 20 hectares or less, which represent 72.1 % (^{INEGI, 2017}) and operate under a peasant logic, could adopt a business logic, that is, that they could organize as a group and govern their production taking into account the market, and thereby be competitive, but it was not like this.

Competitiveness is a concept that emerges from the classical economy with Adam Smith, took a boom with the publication of ^{Porter (1980)} and is related to the ability of companies to penetrate a market, deal with competitors and generate economic gains. In another of his publications, ^{Porter (2008)} defines competitiveness as the capacity shown by companies to face the five competitive threats: rivalry in the industry, the threat of substitute products, the threat of incoming competitors, the bargaining power of suppliers and the bargaining power of customers. Competitiveness has been measured with different approaches, and in the agricultural sector it has not been the exception. In this sense, in their review, ^{García-García et al. (2015)} find that competitiveness in the agricultural sector has been approached predominantly with an economic vision, however, interest has recently grown to include other dimensions such as social, political and environmental. The variables that have been included in the measurement of the competitiveness of companies, cooperatives, productive chains or entire industries come from the different dimensions mentioned and include financial and economic performance (^{Emam & Mohamed, 2011}; ^{Magaña Sánchez et al., 2010}), productivity (^{Girán et al., 2008}; ^{Moyano Fuentes et al., 2008}) and public policies (^{Elbadawy et al., 2013}). On the other hand, it has been identified that the main disadvantages to achieve competitiveness in agriculture derive from causes such as the reduced size of the production unit, limited innovation and low access to information, limited access to markets and less skilled labor (^{IICA, 2016}).

This research carries out an approximate assessment of the competitiveness of maize farmers benefited from the PROAGRO-Productivo program in the Central Valleys region of the state of Oaxaca, Mexico. It does so based on an indicator named Commercial Capacity, which integrates the values of four variables: yield, number of customers, sale price and proportion of the production destined for sale. Commercial Capacity is not synonymous with competitiveness, since the latter is a consolidated concept in the scientific field, which includes more dimensions than those considered for the calculation of the indicator. In that sense, Commercial Capacity does not include social and environmental dimensions, for example. However, the variables that make up Commercial Capacity are adequate to determine a portion of competitiveness because they are measures which, if favorable, help mitigate the adverse effect of some of the competitive threats described by ^{Porter (2008)}.

With the help of universities and research centers, CIMMYT evaluates the impact of its interventions which are carried out with public resources, in order to identify areas for improvement in its processes. Therefore, the objective of the study was to identify the factors that favor a higher level in the proposed indicator as an estimator of competitiveness. The variables examined as possible determinants are related to the sociodemographic profile of farmers, their degree of connection to the commercial network and the type of innovations they perform.

Material and Methods

The information used in this research comes from a survey that CIMMYT staff applied to maize farmers in the Central Valleys region of the state of Oaxaca. The purpose of CIMMYT with the survey and subsequent analysis, in collaboration with the Chapingo Autonomous University, is to monitor and evaluate the progress of its interventions, in this case within the framework of the PROAGRO-productivo program, to identify areas for improvement in its processes. Data were collected on sociodemographic profile of farmers, their level of adoption of innovations, yields, technical and commercial links, sales prices and proportion of the production destined for sale.

For the concept of innovation, we went to the Organisation for Economic Co-operation and Development ^{(OECD) (2005)}, which defines innovation as a process that implies the adoption of something new or significantly improved and that is reflected in changes that can be of a technological, commercial or organizational type. Under this perspective, a catalog of innovations was built with the help of technical advisors and trainers from CIMMYT, who listed a series of practices that improve production yield and quality, and that represent something new or significantly improved in relation to usual practices in the region studied.

Particularly, we chose to study the Central Valleys region of the state of Oaxaca, due to the high variability that the farmers of this region showed in terms of their yields, sale prices and number of customers. Specifically, 14 municipalities were studied: Ciénega de Zimatlán, Cuilápam de Guerrero, Heroica Ciudad de Ejutla de Crespo, Magdalena Apasco, Ocotlán de Morelos, San Bernardo Mixtepec, San Juan Chilateca, San Miguel Ejutla, San Pablo Huixtepec, Santa Ana Zegache, Santa Gertrudis, Tlacolula de Matamoros, Trinidad Zaachila y Villa de Zaachila (Figure 1).

Source: Own elaboration.

Figure 1 Location of the 14 municipalities analyzed in Valles Centrales region of Oaxaca. 

The sample analyzed included 266 maize farmers who were beneficiaries of the PROAGRO-Productivo program in the specified municipalities. The information was collected during April-May 2018 by CIMMYT technical personnel, the data correspond to the 2017 production cycle. For data capture and systematization of the information, Microsoft Excel ® version 2016 was used.

The commercial network was built based on the methodology of actors detailed mapping from ^{Rendón et al. (2007)}. The basic question was: whom do you sell your production? The relational information required the identifiers catalog shown in Table 1. With these references the graph of the network was constructed with the Ucinet 6.0 and Gephi 9.2 softwares.

For the determination of factors that influence on Commercial Capacity of farmers under study, three multiple linear regressions were formalized with the help of the statistical package SPSS ® version 25, in which the Commercial Capacity was entered as the dependent variable. Commercial Capacity is a composite indicator based on four variables: 1) yield (t ha^-1), 2) proportion of production destined for sale (%), 3) number of customers, and 4) sales price ($ t^-1). It is based on the idea that greater Commercial Capacity implies higher levels of yield and production destined for the market, as well as a greater number of clients and a higher sale price. Therefore, in this study, Commercial Capacity is defined as the level of competencies that farmers show to face the competitive threats of their environment.

To obtain the Commercial Capacity of each farmer, we based on the multicriteria analysis procedure (^{Saaty, 1996}), which consists of two stages: weighting and aggregation. In the weighting stage, for each variable, the farmer with the highest value was taken as a reference and the value of 100 was assigned to it, then, based on this, the proportionality ratio of the values of each farmer was obtained. This was done with the four variables. In the aggregation stage, the average of the four proportional relationships obtained previously was calculated; this average reflects the Commercial Capacity shown by each farmer in relation to their peers. Then, Commercial Capacity ranges from 0 to 100, where values close to 100 reflect high Commercial Capacity and values close to zero otherwise.

The independent variables entered in the three regressions are explained in Table 2. To assure the quality of the estimators, the assumptions of nonmulticollinearity and non-heteroskedasticity in the three regressions were checked. Multicollinearity was inspected by the ^{Klein criterion (1962)} and heteroskedasticity through the ^{Glejser test (1969)}.

Results and Discussion

Table 3 shows a characterization of the 266 farmers analyzed, based on the means and standard deviations of the variables examined in the present study. In relation to their sociodemographic profile, in general, producers are elderly people and their school preparation does not exceed primary education. Regarding their productive profile, according to the classification of ^{SAGARPA (2018)}, they are producers of self-consumption given their low scale in production, and their yield is lower than the average (1.95 t ha^-1) reported in surfaces with temporary water regime nationwide in the autumn-winter 2019 cycle (^{SIAP, 2019}). In general, the level of access they have to the commercial network is low, and on average, they adopt only one in five innovations that were listed in the catalog by CIMMYT staff.

The commercial network of the 266 producers analyzed is presented in Figure 2. It is observed that the main buyers are collection centers (CA), intermediary clients (CI) and relatives (FAM), 52 % of the producers sell to the first, 16 % to the second and 10 % to third. In addition, in the center of the network there is a significant number of buyers present, however, on the periphery there are seven sub-nets dominated by seven buyers that concentrate 51 % of the network’s commercial relations. The actors that dominate the peripheral subnets are collection centers (CA), intermediary clients (CI) and input suppliers (PI). This situation is a reflection of the oligopsonistic structure (few buyers) prevailing in some of the municipalities, which puts farmers in a disadvantageous condition. A high dependence on prices paid by intermediaries and collection centers reflects the disadvantage faced by the farmers studied, and likewise happens with the majority of individual maize farmers in Mexico (^{Ortíz-Rosales & Ramírez-Abarca, 2017}).

Source: own elaboration with information on surveys carried out in 2018.

* CA: collection centers; CF: final customers; CI: intermediary customers; ER: maize farmers benefiting by PROAGRO-productivo; FM: actors with multiple roles; IG: government institutions; IE: teaching and research institutions; MQ: maquila companies; NP: non-farms; OR: farmer organizations; FAM: relatives; PSP: professional service providers; PI: input suppliers; and PF: financial providers.

Figure 2 Commercial network of the 266 maize farmers analyzed. 

The multiple regressions through which were identified the factors that influence on Commercial Capacity of the maize farmers analyzed are presented in Table 4. The values of the coefficient of determination (R2) are low in the three regressions, therefore, models lack sufficient goodness to make precise predictions, however, the statistical significance of some estimators (b) indicates a real relationship between significant independent variables and the dependent variable. In this sense, it was found that Commercial Capacity increases as farmers are older (p < 0.05) and have a larger surface (p < 0.05); it also increases as the degree of connection to the commercial network measured through the coverage indicator (diffuse) increases (p < 0.1); and similarly, it increases with the use of improved seed (p < 0.1). On average, the farmers analyzed are 63 years old, have an area of 1.5 hectares, a degree of connection to the commercial network of 0.28 % and 25 % of them use improved seed. On the other hand, it was identified that the use of pheromones decreases the Commercial Capacity (p < 0.05). Of the farmers analyzed, 5 % use pheromones.

^{Janthong and Sakkatat (2018)} identified a similar pattrn in Thailand, the analysis they performed reveals that a greater experience in the cultivation of maize (which is largely given by age), a larger size of the production unit and a stronger link with public and private institutions are significantly associated with the adoption of good agricultural practices, including the use of improved seeds, which produces improvements in the productivity of farmers.

The use of improved seeds has demonstrated its potential through advances in the quality and yield of maize grain in different territories of Mexico, but for this, varieties that adapt to the areas must be used (^{Jaramillo-Albuja et al., 2018}), its supply must be timely (^{Luna-Mena et al., 2012}), and its use must be accompanied by an adequate use of chemical fertilizers and insecticides (^{Brenes et al., 2011}). With regard to the improved varieties developed by CIMMYT, in Pakistan, it has been shown that their use impacts grain yield in wheat with improvements of between 5 % and 17 %, compared to local varieties (^{Joshi et al., 2017}).

On farmers age, ^{López-González et al. (2019)} have generated evidence of its direct relationship with the level of adoption of progressive innovations, which improve the maize yield of Mexican farmers. In the same study, the use of improved seed is defined as a radical and non-progressive innovation, however, for the farmers of this study, it can be a progressive innovation because the technical support offered by the PROAGRO-Productivo program can be reducing the perceived size of the change that innovation implies.

With respect to the scale, it has been found in Mexican nut farmers that greater surface area and greater production increase the bargaining power of the producers, and with this, sale price and income improve (^{Espinoza-Arellano et al., 2019}); in this sense, the organization is the way through which small producers can increase their bargaining power. However, the size of the farm does not always represent a restriction. In Bangladesh, evidence has been generated that the production of maize at the farm level is profitable and competitive regardless of the amount of cultivated area, because farmers in that country show a high capacity to respond to changes in input and product prices, being the amount of labor the variable that farmers adapt to the circumstances imposed by the market (^{Rahman et al., 2016}). This adaptation of the labor factor is especially important for producers with a low degree of mechanization, such as those analyzed here, since labor is the main variable input required for the production of maize. On the other hand, highly mechanized maize production systems, like some in South Africa, show profitability with low sensitivity to salary increases (^{Saayman & Middelberg, 2014}).

Finally, a greater link is known to contribute to the improvement of the results of MSMEs, both by increases in productivity, via access to knowledge, and by improvements in the conditions on which the commercialization of production is carried out, via access to more and better clients (^{González-Campo & Gálvez-Albarracín, 2008}). In maize, it has been proven that farmers obtain higher yields as they are linked to other farmers that serve as social support, and to suppliers and customers that provide specialized knowledge (^{Sánchez Gómez et al., 2016}). Regarding the adverse effect of the use of pheromones, it may be due to the fact that its application for combating the fall armyworm (Spodoptera frugiperda) is not carried out with the appropriate distribution and maintenance specifications (^{INIFAP-CIMMYT, 2016}).

According to the above, the identified combination of factors that determine Commercial Capacity influences on the four variables considered in the calculation of this indicator. When comparing the 10 most prominent farmers in terms of their Commercial Capacity with the averages of the network, it is found that five sell at a better price, eight obtain higher yields and four of them equal or exceed four tons per hectare, eight incorporate a greater proportion of its production for sale and four have a greater number of customers (Table 5). The sale price of $7,500 per ton corresponds to a farmer that produces and sells native maize, which is better paid than the improved varieties, although it has a significantly lower yield.

The actions carried out by the most outstanding farmers mainly translate into improvements in their yields, which in turn allows them to incorporate a greater quantity of product into the market. On the other hand, the price and the number of clients are variables that present a greater restriction for their improvement, since they are limited by the oligopsonistic structure of the commercial network. A similar situation was identified in small maize farmers in Malawi and Mozambique, places where the increase in productivity is concluded is the most efficient way to improve their competitiveness, because on the side of input and product prices, individual producers have little to do (^{Mango et al., 2018}). Therefore, the promotion of proven technological innovations, such as the use of improved seeds, is a task to take into account in interventions aimed at strengthening the competitiveness of small maize farmers.

Conclusions

Despite the difficulty involved in making a profitable commercialization for Mexican maize farmers, after the liquidation of the parastatal company CONASUPO, this study revealed that some farmers have greater Commercial Capacity, and with that, improve their competitiveness by counteracting the threats of the bargaining power of suppliers and the bargaining power of customers. The factors that increase the Commercial Capacity are: a greater age, a greater surface, a greater connection to the commercial network and the adoption of improved seed. The yield is the variable included in the calculation of Commercial Capacity that is most strongly favored by this combination of elements. Having high yields and a greater quantity of grain for sale allows maize farmers to have greater possibilities of overcoming the challenges of producing in a free market framework.

Finally, it should be mentioned that the study has two limitations. Firstly, the accuracy and reliability of the results only apply to the farmers analyzed and not to all maize farmers in Mexico. However, the findings and their consistency with other studies carried out inside and outside the country indicate that the results can be retaken as a reference for future studies. Secondly, although the study is carried out with farmers benefited by the PROAGRO-Productivo program, it cannot be inferred that the findings are caused by the intervention of the program. Therefore, it is suggested that comparative research be carried out between farmers benefiting from PROAGRO-productivo and similar farmers who do not receive such benefit.