The study area

Vanilla (Vanilla planifolia Jacks. ex Andrews) is an orchid native to the rainforests of southeastern Mexico and Central America (^{Bello et al., 2015}). Due to the characteristics of the crop, it grows in subtropical, warm and humid climates, at altitudes below 600 m, although plants have been found up to 1 100 m, with temperatures from 20 °C to 32 °C, average annual rainfall between 2 000 at 3 000 mm; for optimum growth it requires 50 % shade, and up to 70 % in dry seasons to conserve soil and air moisture. It grows in land with excellent drainage, rich in humus and pH from 6 to 7 (^{Hernández, 2011}). In Mexico, the production of vanilla has been recorded in 52 municipalities (Figure 1), which are located in the states of Chiapas (1), Hidalgo (1), Oaxaca (7), Puebla (11), Quintana Roo (3), San Luis Potosí (2) and Veracruz (27) (^{SIAP, 2017}).

Producción de vainilla = Vanilla production; Municipios = Municipalities; Estados = States

Figure 1 Location of vanilla-producing municipalities (Vanilla planifolia Jack. ex Andrews) in Mexico from 2003 to 2014. 

International productivity of the vanilla sector

From the Food and Agriculture Organization of the United Nations (FAO) data were obtained on production (t), planted area (ha) and field yields (t ha^-1) of the main vanilla producing countries for the 2003-2014 period (^{FAOSTAT, 2017}). Through an analysis of variance and Tukey’s means tests, significant differences were established for each of the variables by country, and by means of the application of the methodology proposed by ^{Santillán et al. (2014)}, the type of production system (intensive or extensive) was determined.

National productivity of the vanilla sector

From the Agricultural and Fisheries Information System (^{SIAP, 2017}) variables were obtained: yield in the field (t ha^-1), area sown (ha) and production (t) of the municipalities that registered continuous operations from 2003 to 2014. Such period was divided into two stages: the first from 2003 to 2008 and the second from 2009 to 2014, as the renewal cycles of the plantations last around six years (^{Kelso-Bucio et al., 2012}).

To analyze the productivity of vanilla we used the decomposition model of growth factors of production (^{Gómez, 1994}), the model has been used effectively in different investigations (^{Cruz et al., 2013}, ^{Santillán et al., 2014}; ^{Schwentesius and Sangerman, 2014}) and precisely determines the influence of the yields in the field and the sown area, as well as their interaction in the growth of production in a base period (2003-2008) with respect to a final period (2009-2014).

The sown area, the yield in the field and the interaction (area planted - yield in field) in percentage, were plotted to calculate the growth of the vanilla production at the municipal level, if it is extensive or intensive and which of the factors explains said growth. The municipalities were classified according to the typology established by ^{Santillán et al. (2014)}, in which different combinations are made to know those in which production increases and in which it decreases.

Finally, to identify some of the factors that explain the changes in the planted area and the field performance at the municipal level, specialists in the production of vanilla from the Universidad Autónoma Chapingo (Autonomous University of Chapingo) and the Colegio de Postgraduados (Graduate Studies College) were interviewed, as well as producers and representatives of the Comité Sistema Producto de Puebla y Veracruz (Product System Committee of Puebla and Veracruz). The interview guide considered phytosanitary, climatic, social and market aspects.

International productivity of the vanilla sector

According to Tukey's means test, Indonesia and Madagascar, the main vanilla producing countries in the world, base their production on an extensive system, which compensates for the low yields in the field. The success of Indonesia and Madagascar is based on the comparative advantages associated with favorable natural factors. It is driven by the low opportunity cost of internal production factors such as labor, land and capital (^{Hernández et al., 2010}). However, they cannot be considered competitive, because according to ^{Porter (1991)} a competitive advantage is based on the ability to transform labor, land and capital into differentiated goods and services, which increases efficiency in order to increase profit. ^{Gómez (1994)} postulates that intensive systems tend to be more competitive by making production techniques and management practices more efficient, which according to ^{Ormerod (1997)} stimulates economies of scale and improves competitiveness.

On the other hand, Mexico presented the second best yields in the field, but its production is reduced by the low area sown for cultivation (Table 1). ^{Barrera et al. (2011)}, ^{Jaramillo et al. (2012} and ²⁰¹³) agree that the problem of competitiveness of the national vanilla sector is more a problem of differentiation, since most of the vanilla sold in the country is green, which reduces direct purchase prices to the producer, and causes an abandonment of the sector, which explains the reduction in the area sown. They also agree that although yields in the field are acceptable, there is a wide margin for improvement that, together with an increase in the harvested area and a marketing scheme with niche markets, would substantially improve international productivity.

National productivity of the vanilla sector

In order to determine the factors that prevent an increase in the area planted at the national level, which has a positive impact on international productivity, the factor decomposition model of production growth at the municipal level was applied (^{Gómez, 1994}). It was found that from 2003 to 2014, 52 municipalities cultivated vanilla nationwide, of which only 14 had constant production throughout the period, located in the states of Puebla (1), Veracruz (9) and Oaxaca (4) (Figure 2A). In the national context from 2009 to 2014 there was a sector growth of 2.5 % compared to 2003-2008, as a result of an increase in field yields (2.0 %) and area sown (0.5 %) (Figure 2B). However, this growth is registered only by 14 out of the 52 vanilla producing municipalities. For them, in the same periods of analysis, there is a 3.4 % rise in production in response to a 1.6 % increment in the planted area and 1.8 % in field yields (Figure 2C), which contrasts with the fall in production (-0.1%) of the remaining 38 municipalities, as a consequence of a reduction in the planted area (-2.3%) (Figure 2D). ^{Barrera et al. (2011)} assigned this effect to the cyclical nature of the crop, management practices, physical and climatic aspects, but above all to two factors: the lack of enough economic resources and of knowledge about production technology by the producer.

Producción = Production; Tasa de crecimiento = Growth rate; Crecimiento= Growth; Superficie = Surface area; Rendimiento = Yield; Interacción = Interaction

Figure 2 Growth of vanilla production in Mexico in the 2009-2014 period compared to 2003-2008. 

The growth of the vanilla production at national level was presented by an increase of the yields in the field, which counteracts the reduction of the sown area. ^{Hernández (2011)} states that vanilla yields have a positive trend but are still considered low and associated with a lack of technological innovation; this fact prevents increasing its production internationally. ^{FND (2017)} points out that the yields of vanilla have fallen around 0.35 t ha^-1 and agrees that the use of higher technology and improved processes would increase yields. The aforementioned author calculated that with traditional technology, 0.33 t ha^-1 is produced and with technological innovation it is possible to obtain up to 1.55 t ha^-1. In this regard, ^{SAGARPA (2011)} and ^{FIRA (2012)} granted support to the agricultural sector for the transfer of technology, technical assistance, training and technological development, with the aim of improving the productivity of the field of cattle raising.

From 2003 to 2014, 63 % of national production was based on 14 municipalities that registered continuous operations (^{SIAP, 2017}). Table 2 shows the productivity of these municipalities from 2009-2014 compared to 2003-2008 and their importance in the national context; Four municipalities (Gutiérrez Zamora, Papantla, Tecolutla and Colipa in Veracruz) have increased their sown area and field yield (Type A). Misantla and Juchique de Ferrer (Veracruz), and Venustiano Carranza (Puebla) have adopted an extensive model (Type B); San Felipe Usila and San Pedro Ixcatlán (Oaxaca) an intensive model (Type C). However, the most drastic scenarios are located in Atzalán, Cazones de Herrera and Tuxpan (Veracruz), and San José Independencia and Cosolapa (Oaxaca) where productivity has declined (Type D, E and F). Figure 3 illustrates, above the dotted diagonal, the municipalities that had growth in production, and below those that showed a reduction in production for the period 2009-2014 compared to 2003-2008.

Producción = Production; Tasa de crecimiento = Growth rate; Crecimiento= Growth; Superficie = Surface area; Rendimiento = Yield; Interacción = Interaction

Figure 3 Typology of municipalities based on growth axes in sown area and yields in the field. 

From 38.18 % of the planted area, 37.44 % of the production is obtained, which is concentrated in four of 52 municipalities. When considering those that have lost productivity and those that have been inconstant in their production, 48.49 % of the national production is at risk.

The production of vanilla increased in 9 of the 14 municipalities analyzed and the yield in the field is the factor that mostly influenced this tendency. From 2011, institutions such as SAGARPA and FIRA, implemented the application of technological packages to the cultivation of vanilla, which generated more knowledge in the management of the crop; this fact explains the increase in yields at the national level (^{Hernández, 2011}). However, the country has not been able to compete internationally. ^{FAOSTAT (2017)} reports that Mexico obtains higher yields than those of Indonesia and Madagascar, but that it has not been able to increase its sown area and thus, its production. Although support programs have already been created, it is necessary to train producers in crop management and design strategic programs to increase the sown area. With the productive increase that results, it is expected it is expected to improve the competitiveness of the vanilla sector (^{Hernández, 2011}; ^{Jaramillo et al., 2012}).

Finally, when interviewing specialists in the production of vanilla from the Universidad Autónoma Chapingo and the Colegio de Postgraduados, as well as producers and representatives of the Comité del Sistema Producto de Puebla y Veracruz, it was possible to establish some of the factors that explain the changes in the sown surface and yield in field by type of municipalities (Table 3).

Although in Mexico there are conditions that could encourage the production of vanilla, such as good land, guardians, plants, productive experience and quality control, climate, running water and organic matter, we also have low prices, market unstable (price variation, lack of buyers, etc.), climatic changes of prolonged drought, pests and diseases, lack of specialized and permanent technical assistance, infrastructure, equipment and work tools, financing and promotion of the establishment of plantations. To this should be added the insecurity in some of the production areas as occurs in the Totonacapan region, the main producing area nationwide. The causes mentioned are present in 38 of the 58 producing municipalities, and unfortunately, it is difficult to reverse them in the future.