Apuntes del Cenes ISSN: 0120-3053 Universidad Pedagógica y Tecnológica de (UPTC)

Arias-Gómez, Helmuth Yesid; Antosová, Gabriela Spatial Patterns of in Boyacá Apuntes del Cenes, vol. 37, no. 66, 2018, July-December, pp. 203-237 Universidad Pedagógica y Tecnológica de Colombia (UPTC)

DOI: https://doi.org/10.19053/01203053.v37.n66.2019.6013

Available in: https://www.redalyc.org/articulo.oa?id=479558722008

How to cite Complete issue Scientific Information System Redalyc More information about this article Network of Scientific Journals from Latin America and the Caribbean, Spain and Journal's webpage in redalyc.org Portugal Project academic non-profit, developed under the open access initiative Artículo de investigación Apuntes del CENES ISSN 0120 - 3053 E-ISSN 2256-5779 Volumen 37 - N° 66 julio - diciembre 2018. Págs. 203 - 237

Spatial Patterns of Agriculture in Boyacá

Patrones espaciales de la agricultura en Boyacá

Padrões de espaço da agricultura em Boyacá

Helmuth Yesid Arias Gómez * Gabriela Antosová **

DOI: https://doi.org/10.19053/01203053.v37.n66.2019.6013

Fecha de recepción: 17 de abril de 2017 Fecha de aprobación: 11 de abril de 2018

Cómo citar este artículo/ To reference this article / Comment citer cet article / Para citar este artigo: Arias, H., & Antosová, G. (2018). Patrones espaciales de la agricultura en Boyacá. APUNTES DEL CENES, 37(66). https://doi.org/10.19053/01203053.v37.n66.2019.6013

* Economist. Ph D. Applied Economic Analysis and Economic History (Universidad de Sevilla). Master in Economic Sciences (Universidad Nacional, Bogotá). Universidad Sergio Arboleda, Bogotá and University the College of Busi- ness, Department of Economics (Prague). [email protected]. orcid.org/0000-0003-0107-8611 ** Business Manager. Graduated Ph.D. in Regional Development University the College of Business, Department of Economics (Prague). [email protected]. orcid.org/0000-0001-5330-679X

203 Spatial Patterns of Agriculture in Boyacá Helmuth Yesid Arias Gómez • Gabriela Antosová

Abstract This article figures out spatial patterns for Boyacá’s agriculture and exposes the behavior of production. Applying gravitational concepts we describe relevant spatial interactions across municipalities using variables as population, output and linear distances. We perform an econometric model for detecting the class of spatial dependence showed by agricultural output as endogenous using as arguments distance to and rural population. We deploy standard tools of spatial analysis and empirical strategies for identify clusters of towns according with their performance and productivity. Statistical contrast indicates that the most suitable scheme for describing spatial dependence in production is spatial lag model. Econometrics conveys important clues demonstrating that higher scales output is conducted to urban national markets and output produced with scarce scales is sold locally in local towns. A strong subjacent idea is that lagged municipalities are badly influenced by geographical isolation and high transportation costs do hinder the social and economic development in Boyacá. Keywords: gravity model, agricultural production, productive structure, spatial interaction, spatial econometrics. Classification JEL: R12, R 14, R 58, C 51

204 Apuntes CENES Volumen 37, Número 66, ISSN 0120-3053 julio - diciembre 2018, 203 a 237

Resumen Se indagan los patrones espaciales de la agricultura boyacense analizando el comportamiento de la producción. Apelando a conceptos gravitacionales se describen relevantes interacciones espaciales entre municipios, usando como variables la población, la producción agrícola y las distancias lineales. La econometría espacial detecta la índole de la dependencia espacial exhibida por la producción en función de la distancia a Tunja y de la población rural. Se despliegan herramientas del análisis espacial y estrategias empíricas para identificar clústeres de municipios de acuerdo con su desempeño y productividad. Los contrastes estadísticos recomiendan elegir el modelo de rezago espacial como más apto para explicar la dependencia espacial. Interesantes resultados econométricos sugieren que la producción con altas economías de escala se destina a los mercados sólidos y la pequeña producción parcelaria se vende en el propio municipio. Definitivamente, los altos costos de transporte constituyen un impedimento insalvable que interfiere en el desarrollo en las zonas más apartadas y menos productivas. Palabras clave: modelo gravitacional, producción agrícola, estructura productiva, interacción espacial, econometría espacial.

205 Spatial Patterns of Agriculture in Boyacá Helmuth Yesid Arias Gómez • Gabriela Antosová

Resumo: Os padrões espaciais da agricultura de Boyacan são analisados, revisando o comportamento da produção. Apelando para conceitos gravitacionais, são descritas interações espaciais relevantes entre os municípios, utilizando como variáveis a​​ população, a produção agrícola e as distâncias lineares. A econometria espacial detecta a natureza da dependência espacial exibida pela produção em função da distância a Tunja e à população rural. Ferramentas de análise espacial e estratégias empíricas são implantadas para identificar clusters de municípios de acordo com seu desempenho e produtividade. Os resultados estatísticos recomendam a escolha do modelo de spatial error como mais apto a explicar a dependência espacial. Resultados econométricos interessantes sugerem que a produção com altas economias de escala vai para mercados sólidos e produção do pequenas parcelas de terra são vendida no município. Definitivamente altos custos de transporte são um obstáculo intransponível que interfere no desenvolvimento nas áreas mais remotas e menos produtivas. Palavras-chave: modelo gravitacional, a produção agrícola, estrutura de produção, interação espacial.

206 Apuntes CENES Volumen 37, Número 66, ISSN 0120-3053 julio - diciembre 2018, 203 a 237

As showed in Figure 1, it is clear that INTRODUCTION the geographical conditions of the state There is a myriad of technical anal- suggest a predominant presence of high ysis dealing with agricultural sector lands and cold climates that propitiate coming from academic and technical the development of agricultural crops, institutions and also pullulate policy adequately adapted to such conditions. documents emanated from national and In spite of the importance of cold climate regional governments, which diagnose production, we can also find a broad the situation of rural sector. There the variety of climates and natural environ- central topic gravitates around pro- ments that stimulate the production of ductivity, competitiveness, regional other commodities. So, from the outset and local development and planning we can presume the remarkable role of of territory. In this article we try to agriculture in the state´s economy. apply some quantitative techniques to identify economic and spatial trends of This article tries to identify the behavior agriculture in Boyacá and we consider in the spatial distribution of agricultural that by its methodology and approach, production in Boyacá state. Provided the article contributes to the discussion the remarkable role of its rural sector of productive and spatial conditions at a national scale, this article tries to of agricultural output and remarks the analyze in detail the productive role of fundamental drawbacks that hamper municipalities in the agricultural output the productive and social development and the implication of dependence in the of provinces within the state. spatial distribution of output. In order to deal with such topics the article is orga-

207 Spatial Patterns of Agriculture in Boyacá Helmuth Yesid Arias Gómez • Gabriela Antosová nized as follows: after this introduction, models are checked. Later, empirical we present a short description of the strategy is developed and spatial econo- main agricultural products in Boyacá. metric model is performed. At the end, Secondly, a summary of theoretical and some conclusions are proposed. methodological approaches on gravity

Figure 1. Map: Boyacá: Agricultural Production, Highness and Weather. Source: Own elaboration based on Instituto Geográfico Agustín Codazzi, DANE, Ministerio de Agricultura y Desarrollo Rural.

From the outset we highlight the spa- of location theory (mainly Lösch and tial dimension of production either for Christaller) and more recently by New theoretical analysis as for empirical Economic Geography theorists, who methodologies. In fact, spatial distance incorporated transport costs as a crucial becomes an important determinant in parameter in models of location and relationships and economic flows across spatial distribution of manufacturing territories. Commercial exchanges, activity. At international level, the new mobility factor, capital movements, international trade theory underlines migrations, etc., can be prompted by the upsurge of intra industry exchang- closeness between countries and re- es between neighbor countries as the gions, and geographic proximity fosters main feature of actual trade between specialization. The problematic of space advanced economies. In this kind of ex- have been treated before in the context changes, differentiated goods classified

208 Apuntes CENES Volumen 37, Número 66, ISSN 0120-3053 julio - diciembre 2018, 203 a 237 in the same statistical category are trad- According to new approaches, it is inter- ed between commercial partners who esting to tackle a regional analysis about take advantages of increasing returns behavior of agricultural production and and scale economies (Krugman, 1992, its spatial determinants, conditioned to 2008). market size and distance to consumer places. In doing so, we apply a gravity In spite of the evidence that spatial in- approach to identify municipalities that teractions make up an underlying phe- reveal different degrees of spatial inter- nomena in economic relationships, they action with neighboring markets, taking have not been acknowledged or embod- into account the magnitude of own ied in economic theory´s hard core. In agricultural output and the proximity to fact, in the conventional international main regional centers. Boyacá state is an trade theory, countries are assumed as optimal case of study due to its key role points without spatial meaning (Krug- in Colombian agricultural production, man, 1992) and on this basis, conven- and its idiosyncratic productive organi- tional theories define commodities as zation, based on small scale peasantry perfectly traded and factors completely and by the specific mechanism of har- immobile assuming null transport vest commercialization. So, using as ba- cost. Quoting Leamer and Levinsohn, sic unity of analysis the municipalities, Anderson asserts that in academic all ingredients for gravitation analysis context there is no “explicit reference to come up: distance between farms and distance, but with the very strange im- markets, size of agricultural production plicit assumption that countries are both and regional urban markets. infinitely far apart and infinitely close, the former referring to factors, and the SPATIAL DESCRIPTION OF latter to commodities” (Anderson, 2010, AGRICULTURE IN BOYACÁ p. 1). The role of Boyacá in national agricul- So, in order to analyze geographical tural landscape is marked by special- concentration of production and spe- ization and by natural condition. Jointly cialization of countries and regions, with Cundinamarca, Boyacá leads the new economic geography integrates cold climate supply for central zone of in analysis concepts as: transport cost, Colombia, and beyond, its production increasing returns and mobility of reaches further markets in mild zones at manufacturing labor (Krugman, 1991, north of the country. 2008). So, in the context of theories and empirical works that highlight space, Boyacá state is one of the most re- distance and spatial influences emerge markable agricultural producers in as powerful forces that prod economic Colombia, where the economic process processes up. is surrounded by a rural context of peas- antry and small farms (DNP, 2007). Ac-

209 Spatial Patterns of Agriculture in Boyacá Helmuth Yesid Arias Gómez • Gabriela Antosová cording to Proexport (2014), this region Natural diversity responds to complex- stands out as the first vegetables and ity in weather conditions and climate tomatoes producer. In the same way, it areas in a reduced space. According to is ranked as the first supplier in eleven Plan Frutícola Nacional (PFN, 2006), varieties of fruit and shares with Cundi- the region has all climate areas. As we namarca and Nariño the first places in will show below, topographical condi- the production of potatoes. tions are a significant determinant of rural production provided that average Also in onion crops the state is an active highness in the towns of the department protagonist. Boyacá concentrates the of Boyacá rises to 2066 m. half of onion production in Colombia, followed by Norte de Santander and Broad climatic variety and diversity Cundinamarca. At local level some in environment conditions explain the towns have an important role. Produc- productive vocation in Boyacá. Recent tion in and Samacá represents statistical information gives an idea of 16 % and 5 % of national production, the dimension of volumes in agricultur- respectively (Asohofrucol, 2013). al output, as Table 1 shows:

2010 2011 2012 2013 839,8 756,5 661,8 722,1 Cane 164,0 159,8 167,6 205,9 Plantains 28,4 39,6 22,8 41,1 Yucca 32,6 27,3 17,7 31,3 Corn 17,3 19,0 16,7 18,8 Bean 6,5 5,5 4,7 4,2 Pear 3,9 15,0 3,7 23,0 Peach 12,7 12,0 14,8 12,9 Guava 9,4 10,1 18,3 9,9 Blackberry 8,6 9,4 5,9 5,7 Plum 11,0 11,4 4,2 9,4 Orange 6,9 8,0 4,7 9,1 Golden Berry 5,5 6,4 4,5 7,8 Table 1. Boyacá: Main agricultural production (thousands of tons) Source: Secretaría de Fomento Agropecuario de Boyacá.

In fact, it is used that small peasants family. On the other hand, the main alternate crops of potato with others fruit crops are: guava, orange, pear and of corn, beans, peas, and barley plum. Moreover, the region is the lead- (Rodríguez & Bermúdez, 1997). Such ing producer of deciduous fruits due multi crop system reduces the eco- to its aptitude in plain lands, mainly in nomic risk for small farmers and offers towns as Nuevo Colón. food for self-consumption of peasant’s

210 Apuntes CENES Volumen 37, Número 66, ISSN 0120-3053 julio - diciembre 2018, 203 a 237

Figure 2. Map: Boyacá: population per municipalities. Source: Own elaboration based on Instituto Geográfico Agustín Codazzi and DANE.

A crucial aspect in agricultural produc- In and farmers mar- tion is the market orientation of com- kets are opened several days in the week, modities and the projection to nearby in particular for fruits. In this condition, consumers looking for local demand. some intermediaries can trade in differ- In terms of population Boyacá exhibits ent points of the state during a week, a few urban centers, meanwhile rural forging an oligopsony structure taking features predominates in less populated information of prices from Bogotá’s towns. market: Corabastos (Plan Frutícola Nacional, 2006). Regional government Most of the towns in Boyacá have a plans to promote local farmers markets relevant rural predominance, and cities to guarantee food security and to match concentrate urban activities. Figure 2 the productive cycle with local needs shows more populated municipalities. of consumers (Gobernación de Boyacá, Tunja represents the administrative and 2012). urban center of the whole region and concentrates bureaucratic, administra- Intermediation chains and markets tive, educational and financial activi- location are crucial for understanding ties. Secondly, other cities of a smaller the dynamics of agricultural system in size are, namely: Sogamoso, Duitama, Boyacá. A part of the output is carried Chiquinquirá and Puerto Boyacá. out to markets from farms and huge

211 Spatial Patterns of Agriculture in Boyacá Helmuth Yesid Arias Gómez • Gabriela Antosová quantities have as destination Bogotá, structures intended for storing and trans- the national main city, Santander state formation of commodities, high cost of and the Northern Coast. On the other production, scarcity of labor, emigration hand, due to relatively short distances, of young workers, not remunerative pric- production can be distributed to main es for production and high climatological markets within the state, where there are and business risks. Moreover, there are daily local transactions, according to the not sufficient coverage instruments for organization of farmer markets in each guarantee eventual losses associated to jurisdiction. climatological and market uncertainty.

In general terms, peasant economy in Potato is a basic commodity in nutritive agricultural exploitations has serious habits of Colombian household and has a profitability and productivity drawbacks big weight in Colombian Consumer Price due to lags in technological and produc- Index. Harvest has diverse destination: tive procedures and lack of credit sourc- 8 % is used as input for manufacturing, es. Regional authorities (Secretaría de processing, 10 % goes to self-consump- Fomento Agropecuario) identified all the tion, 64 % is sold in wholesale markets problems affecting the small parcel of and the remaining quantity is used as production. Features of peasant econo- seed (Superintendencia de Industria y mies are tiny size of farms, lagged infra- Comercio [SIC], s.f.).

Figure 3. Map: Department of Boyacá: Agricultural Output per Municipalities (Tons) vs. Main Highways (Primary net). Source: Own elaboration based on Instituto Geográfico Agustín Codazzi – Ministerio de Agricultura y Desarrollo Rural – Ministerio de Transporte.

212 Apuntes CENES Volumen 37, Número 66, ISSN 0120-3053 julio - diciembre 2018, 203 a 237

The potato production system is dom- in trucks and finally puts the product in inated by a myriad of small dispersed wholesale markets. In fact, the whole- farms. Technologies of production are sale market is the main link between very traditional and barely mechanized producers and consumers because it (DNP, 2007). Small size of parcels concentrates a larger quantity of potato and irregular terrain hamper an in- traded. tensive use of capital and machinery. Geographically, potato crops require In terms of symmetry of information adequate highness and specific climate and efficiency of markets, a small conditions. In general crops are located peasant operates with a big handicap at more than 2.600 meters over the sea because has no exact data on prevailing level, in temperature conditions of 11 prices neither an integral perspective of °C or 14 °C. Suitable places to cultivate the situation in wholesale markets. Sys- potato includes central region of Boyacá tem of commercialization bestows huge in towns as: Tunja, , Boyacá, market power on those intermediaries Chivatá, Toca, Oicatá, , that collect harvest from peasant with Cómbita, , Tuta, Duitama, Santa full awareness of general market con- Rosa, , Belén, Gámeza, , ditions and who are agents that know Aquitania and Sogamoso (Rodríguez & the conditions of final markets and the Bermúdez, 1997). Indeed, some of those final prices of commodities. In general, towns appear henceforth with intense small peasant ignores the prevalent cir- spatial interaction in the gravitation cumstances of highly segmented market model and some of them are promptly as potatoes (Rodríguez & Bermúdez, connected, overlapping with an area 1997). very well connected by highways1 (see Figure 3). On the other hand, fruits crops have an increasing importance in state´s agricul- The system of commercialization of po- tural activities, particularly in mild cli- tato output at a national scale operates mates. According to the National Fruit as an oligopoly mechanism. Indeed, to Plan (2006), the productive potential of convey output into consumption centers the state could be increased in specific a key figure emerges in landscape: the municipalities namely: Blackberry in collector. This agent picks up the pro- , Gachantivá, and duction from farms, transports potatoes Nuevo Colón; strawberry: Arcabuco,

1 A broader list of towns with potato crops appears in FEDEPAPA- MAVDT (2004): Tunja, Aquitania, Arcabuco, Belén, Betéitiva, Boyacá, Briceño, Busbanzá, Caldas, Cerinza, Ciénaga, Cómbita, Corrales, , Cuítiva, , , Chíquiza, Chiquinquirá, Chivatá, Chita, Duitama, , El Espino, , Floresta, Gámeza, , Gachantivá, Guacamayas, Guicán, Iza, Jenesano, Jericó, , , Monguí, Motavita, , Nuevo Colón, Oicatá, , Paipa, , , Paz del Rio, Pesca, Ramiriquí, Ráquira, Saboyá, , Samacá, Santa Rosa de Viterbo, Santa Sofía, , , Siachoque, Soatá, , Socotá, Sora, Soracá, Susacón, Sogamoso, Sotaquirá, Sutamarchán, Tasco, Tibaná, Tibasosa, Tinjacá, , Toca, Tópaga, Tota, Turmequé, Tuta, Tutasá, Úmbita, , Villa de Leiva, Viracachá and Zetaquirá.

213 Spatial Patterns of Agriculture in Boyacá Helmuth Yesid Arias Gómez • Gabriela Antosová

Villa de Leiva, Paipa, Tuta, Tunja and But all public policies with long term Duitama; guava: Moniquirá and Mira- goals, must aim to promote the effective flores; orange: , , Santa insertion of peasant in market economy María, Moniquirá, Miraflores, Páez, reaching remunerative prices covering and ; grape: the cost of production. All public efforts Villa de Leiva, Sutamarchán and Soga- must contribute to improve competi- moso; golden berry: Arcabuco, Úmbita, tiveness of peasant economy, pursuing a Santa Sofía and Miraflores (PFN, 2006). successful and sustainable expansion of rural production. In other places of the According to PFN, there are three chan- world similar goals has been proposed: nels to sell the fruit output. Part of the output is traded through wholesale mar- farmers have slowly acquired a deeper kets and neighborhood markets, aimed understanding of the importance of to the final consumption of a household. commercial channels and marketing Part of the output is sold through large mechanism for the whole process of ag- ricultural and rural development. The retailing stores. Finally, fruits can be role of national policy makers became used as input for manufacturing and crucial to speed up this process and fa- processing (PFN, 2006). Nevertheless, cilitate the way farmers should learn to the bulk of output is absorbed by a con- organize themselves around common sumer through small farmers markets in interest. (De Noronha & De Noronha, towns and neighborhoods. 2010, p. 125)

Deficiencies in commercialization GRAVITATION MODEL chain and marketing deprive peasant, hinder the generation of added value Gravitation models are used to estimate and restrict the formation of remunerat- statistically the relationship between ing prices. Many tasks must be done for variables that are supposed to be influ- improving stuff commercialization, as enced by distance and spatial factors, well as optimizing the quality, selection, when economic flows appear. From the packing and storing and incorporating beginning models have been applied any basic transformation. Regional mainly in the field of international trade, policy suggests reinforcing the storage in order to detect the importance of dis- centers intended for selecting, sorting tances and size of markets in exchange and packing, in particular for commod- of commodities. Anderson (2010) evokes th ities as vegetables, potato, carrot, onion, the works of Ravenstein in 19 century, fruits, cacao, and milk (Secretaría analyzing migration patterns in Britain, de Fomento Agropecuario, s.f.). and those of Jan Tinbergen about trade in 1962. Also, this kind of models has an enormous advantage provided that conventional theories of trade only puts

214 Apuntes CENES Volumen 37, Número 66, ISSN 0120-3053 julio - diciembre 2018, 203 a 237 in scene two countries, conversely grav- In Isard’s version of gravity principles, itation models collect interaction among analysis is applied to migration between several trade partners (Anderson, 2010). the place i and j. In these terms the mod- Other gravity models have been applied el can be presented as: to migration as in case of Tobler (1975). [2] As Isard points out: “Since 1960, a very large array of gravity and spatial interaction models has evolved. They relate to a different kind of interactions, Where: I corresponds to interaction, Pi some purely theoretical, but much more is population in I, Pj is population in j, G frequently interaction concerned with is gravitation constant and d is distance everyday problems” (Isard et al., 1998, (Isard et al., 1998). p. 249). Moreover, gravitation models detect In regional studies the gravitation friction forces that lead the forces of the approach has also been implemented model in an opposite direction. successfully. In this context, regions are assumed as masses, in such a way that In spite of simplicity in the concept of flows between regions can be under- gravitation, a key aspect must be out- stood as the interaction between masses lined. In practice, trade and economic (Isard et al., 1998), in consequence, the phenomena are influenced by forces bigger their dimension, the strong inter- outside the model that divert flows from actions between them. multiple origins and diverse destination. Taking into account the definition of According to De Noronha and De the model, people or goods can come Noronha (2010), in essence the model from several origins different to i, and applies the same principle as the uni- can have as destination diverse places versal gravitation theory of physics as excluding j. This circumstance inserts follows: into model forces of friction that must be analyzed. [1] Results of Gravitation

F being: the force of attraction between For developing this section we will ap- two objects, m two masses and r the ply the formula that appears in [1]. For distance. In these terms the spatial this calculation we will use agricultural interaction between regions is strongly output, population and crossed distanc- influenced by the size of two masses, es across towns. acting as diminishing force the distance.

215 Spatial Patterns of Agriculture in Boyacá Helmuth Yesid Arias Gómez • Gabriela Antosová

In our exercise, gravitation princi- To calculate gravitation forces in this ex- ples are used to model the interaction ercise, two masses are assumed for the throughout agricultural towns, consult- launching process. Firstly, agricultural ing data on agricultural production and production measured in ton interacts by potential markets represented by the interacts with population of towns, as a local population. Gravitation space is proxy for dimension of local markets. constrained to jurisdiction of Boyacá in On the other hand, the distance between order to control gravitation movements. the masses is determined by geo-sta- At this point a caveat is worthy. Out- tistical tools. Geo-referenced and ward markets such as Bogotá, Bucara- geo-statistical processes were carried manga or further points, in practice, are out in software ARC GIS. Data were important market niches for Boyaca’s collected from the webpage of Colom- production. Even few products as golden bian geographical office (IGAC), from berry are sold in international markets. Colombian statistical office (DANE) So, in spite of their potential relevance, and from Ministerio de Agricultura y outside markets are not included in the Desarrollo Rural (1997), including 119 analysis because it would imply geo-sta- municipalities. tistical complications. In these terms, disposable data for Boyacá will be used at the state level for 119 municipalities.

Figure 4. Map: Boyacá: Intensity of Spatial Interaction according to Gravitation Model. Source: Own elaboration based on Instituto Geográfico Agustín Codazzi & DANE.

216 Apuntes CENES Volumen 37, Número 66, ISSN 0120-3053 julio - diciembre 2018, 203 a 237

Figure 4 summarizes the forces that Chíquiza, Cómbita, harvest commodi- interact in the model, and has a similar ties of cold climate that are attracted to rendering as appears in De Noronha Bogotá (national main city). Also in this and De Noronha (2010). According to central zone, towns as Toca, Siachoque gravitational dynamics, surrounding and Viracachá confirm their specializa- Tunja cold climate agriculture pre- tion in potato crops. Tunja itself appears dominates, pushing neighbor towns to in statistics either as a producer and im- supplying urban needs with agricultural portant consumer for regional output. In commodities. In fact, in short distance previous research, Arias and Antosova to the main city, a system of fertile (2015) identified a similar central zone towns produces potato and vegetables, in this state describing a spatial pattern taking advantage of spatial proximity of high agricultural activity, suggesting and working as a crossroad in the state potential spatial autocorrelation. A rele- receiving the main regional highways vant feature is the fluent communication (see Figure 3). that has this central zone, provided that an important highway crosses longitudi- Results of gravitation model depict con- nally those municipalities and connects crete spatial zones located near to urban quickly with the national main city (as centers that spur agricultural output. shown in Figure 3). The intensity of spatial dynamics re- veals that urban markets and important With the purpose to mix the whole in- producers of agricultural commodities formation offered by the gravitational are located in short distances and are model and basic data in mass variables connected by relatively well paved (population and output), a cartographic roads. superposition of layers will be built, in order to compare the spatial behavior In productive terms, close towns, such of points of production and consump- as Samacá, Ventaquemada, Motavita, tion with the spatial intensity of their interrelation.

217 Spatial Patterns of Agriculture in Boyacá Helmuth Yesid Arias Gómez • Gabriela Antosová

Figure 5. Map: Boyacá: Agricultural Production, The Gravitational Model and Population. Source: Own elaboration based on Instituto Geográfico Agustín Codazzi, DANE and Ministerio de Agricultura y Desarrollo Rural.

A combination of three layers displayed in terms of spatial interaction around previously is rendered now in Figure Boyacá´s main city. According to the 5. Higher and lower maps render the results, important agricultural points local values for agricultural output and are concentrated geographically in the population, respectively. But the inter- state looking for supplying strong outer esting information is rendered in the demand. intermediate map, because it represents the outcomes turned out from the grav- On the Figure 5, moving to east, Aqui- itation formula [1]. There can be appre- tania and Tibasosa emerge as leading ciated more clearly the influence of the producers of onion at national level, masses (output and population) on the but nearby we find an important cen- results of gravitation model, and there ter namely: Sogamoso. Consequently, also appear some insights about the those towns show intense spatial in- distance between municipalities. It is teraction as agricultural producers. exhibited there the most important zone Although, Aquitania and Tibasosa are

218 Apuntes CENES Volumen 37, Número 66, ISSN 0120-3053 julio - diciembre 2018, 203 a 237 widely known as prominent producers products outside the state, in particular of onion at a national level, and a big to Bogotá, Bucaramanga, and other share of their output is sold in Bogotá areas beyond state´s frontiers. The well and outside of Boyacá. shaped connectivity built around potato towns helps to communicate quickly In the Northern Province of Ricaurte, with Bogotá and projects the bulk of warmer conditions allow to harvest production of the national main city. Re- commodities of mild climate. In the centlly improved highway from Briceño gravitational model appears to Sogamoso allows the connection of having an intense spatial interaction as agricultural place as Samacá and Ven- a producer and showing closeness to taquemada with Bogotá in a journey populate towns located towards east. lasting roughly two hours (as shown in Neighbors as Moniquirá and San José Figure 3). de Pare have an important share in pro- duction, but are located far away from In consequence, this geographical close- cities as Duitama, Sogamoso, and even ness to the national main city drives Tunja. a large part of production into Bogotá and reinforces some frictional forces, it The geographical extremes of the involves that there is commercial flows department show a weaker intensity and spatial interactions there are getting in spatial relationships, as can be ap- out from our model centered in the geo- preciated at Provinces of Gutiérrez graphical jurisdictions of Boyacá. and Occidente. In the former, towns as Cubará, Chiscas, El Espino, Panqueba, It could be concluded with Isard that: Guacamayas, El Cocuy and Guicán “the more clearly the gravitational inter- are very remote places, barely commu- action is discerned, the fewer the differ- nicated with the state´s main centers. ences among the elements of each mass Consequently, gravitation models detect and the larger the masses involved” an important influence of geographical (Isard et al., 1998, p. 252). distance that inhibits spatial interaction in those cases. In regard to the Province EMPIRICAL STRATEGY of Occidente, it has two populated plac- es: Chiquinquirá and Puerto Boyacá, In this section, for performing geo-sta- but its mining vocation relegates agri- tistics and spatial econometrics our culture activities to a second place. sources of information are DANE, IGAC and Ministerio de Agricultura Nevertheless, powerful friction forces y Desarrollo Rural with data available intervene through the operation of for 119 municipalities and distances commercialization systems. In fact, to Tunja are calculated by means of there are strong flows carrying out geo-statistical tools. The cartographical,

219 Spatial Patterns of Agriculture in Boyacá Helmuth Yesid Arias Gómez • Gabriela Antosová geo-statistical and econometric strate- out the spatial factors that affect spatial gies were performed in ARC MAP and distribution of economic variables, e.g. R software. distance, proximity and spatial depen- dence across geographical entities. Analyzing 119 municipalities, we generated a spatial weight matrix (W) A crucial aspect to consider in spatial defining the kind of spatial interaction analysis of agriculture is that basic data as queen. So the direction of spatial captured are not independent, because association flow is similar to the queen there is a clear relationship between movement in chess. observations and neighboring places (Bongiovanni, 2009). The fact that there Gravitation model highlights the enor- is effectively a strong influence across mous importance of markets, transport points in space drives to the concept of cost, distances, and connectivity in spatial autocorrelation. This is a regu- stimulating agricultural market output. larity that occurs when the basic unities A huge drawback comes up from the are geographically arranged. isolation of some towns in remote re- gions of Boyacá, affecting output and This phenomenon is due to the presence producer’s productivity. In such condi- of variables systematically correlated tions, farmers limit their potential mar- in space or by the evidence of depen- kets, and output is sold once in a week dence in the residuals (Moreno & Vayá, at own local farmers markets. Only high 2002). Spatial autocorrelation may be competitive products can overcome originated in omitted variables in the geographical obstacles and can be sold specified model and consequently, the at competitive prices that cover the high error term will include this kind of transport cost. spatial influence. It happens because there is dependence between the values Moreover, it is normal that geographical of the endogenous variable in neighbor- conditions and climatological influence ing areas, independent of influence of can explain clearly defined clusters of exogenous variables. In other cases, po- municipalities producing the similar litical division of data has no economic kind of products as: potato, fruits, coffee, significance because does not reflect the guava and cane. These reasons convey economic interaction across agents, and important assertions about localization economic relations go beyond artificial of production. A worthy contribution borders (Helsen, 2008). Some theorists to set out an analysis of agricultural have defined such situation in terms of sector is to inquire the economic and an administrative split of territory with spatial aspects related to the volume of no economic relevance as has been output at a local level. In doing so, spa- recognized by Krugman (1992) and tial econometrics offers tools to figure Duranton and Overman (2005).

220 Apuntes CENES Volumen 37, Número 66, ISSN 0120-3053 julio - diciembre 2018, 203 a 237

In technical terms, spatial autocorrela- it is assumed that spatial dependence is tion comes up if residual terms of differ- linked to the spatial model and both are ent geographical areas are correlated. It used when the purpose is to assess the obligates to submit data to correction of degree of intensity in spatial correlation. spatial autocorrelation because results Arfa, Rodríguez, Daniel and Shonk- can be distorted and inference becomes wiler (2010) interpret the inclusion of inaccurate. spatial lag as a way to incorporate the local economies of location. In these But this situation is normally ubiquitous cases, the presence of spatial autocor- if data have a spatial nature or if the lo- relation of substantive class entails that cation is a fundamental criterion. Once estimations will be biased and inconsis- spatial autocorrelation is detected, a tent, even if error term is not correlated careful treatment must be implemented. (Moreno & Vayá, 2002). The evidence of spatial autocorrelation conveys a difficulty for estimation of A second model assumes a structure of the classical OLS model because results dependence in the error term (E[εiεj]≠0) can turn out biased and the estimators and it is understood as a nuisance de- inefficient. Moreover, the variance of pendence, and is another useful strategy estimators can be biased as well (Dubin, for correcting the biasing influence of 1998). spatial autocorrelation (Anselin, 2001). In these terms, correction for spatial au- According to this, the process of correct- tocorrelation conveys an additional ad- ing spatial autocorrelation helps to get vantage represented by an improvement accurate estimates and to improve the in the fit of the model (Bernat, 1996). reliability of the hypothesis test. Once the identification of autocorrelation is On the other hand, in the error model achieved, its information is included if there is nuisance autocorrelation, in prediction contributing to improve estimators are inefficient and, in conse- precision. Afterwards, maximum likeli- quence, statistical inference is invalid hood techniques are commonly applied although it could be unbiased (Helsen, to model the autocorrelation parameters 2008). and estimations of regression (Dubin, 1998). In consideration of the rising impor- tance of spatial econometrics and its For correcting spatial correlation two widespread application, in this article possibilities have been defined. Firstly, a model of agricultural output is run in a model of spatial lag can be estimated order to detect the spatial structure of where a lag of the dependent variable is the endogenous variable. included in the matrix form Wy. Here

221 Spatial Patterns of Agriculture in Boyacá Helmuth Yesid Arias Gómez • Gabriela Antosová

Spatial Analysis and Geo-statistics Exogenous variables have a close relationship with agricultural output In order to apply the geo-statistical and all of them are significant. Rural and econometric methodology, munic- population shows an important explan- ipalities of Boyacá have been chosen atory capacity because commodities as primary geographical unities of are offered at the local level, mainly analysis. Data proceed from Colombian if towns are barely communicated. A geographical office (Instituto Geográfi- geographical variable is represented by co Agustín Codazzi) and Ministerio de distance regarding Tunja, the main city Agricultura y Desarrollo Rural, using of Boyacá, and shows the projection to information adequately geo-referenced. urban markets. According to Arfa et al. Agricultural output is measured in tons (2010) a topographic variables can rep- and is available for all municipalities. resent local relief.

Global and Local Spatial Analysis

Figure 6. Map: Boyacá’s Agricultural Output in tons: Local Indicator of Spatial Association. Source: Own elaboration using ARC MAP, based on IGAC and Ministerio de Agricultura y Desarrollo Rural.

Figure 6 renders an additional statistical significant contrast is applied, having as contrast intended to capture the local a null hypothesis of the absence of local spatial dependence, known as LISA. spatial association. As a consequence, For that purpose, at each spatial entity a each spatial entity can be surrounded by

222 Apuntes CENES Volumen 37, Número 66, ISSN 0120-3053 julio - diciembre 2018, 203 a 237 different or similar values of the same are surrounded by others with similar variable. features.

Concomitantly, the LISA map According with basic data and provided exhibits some clusters, indi- that we suppose spatial interaction only cating the behavior of values within the state borders, the process re- of variable for those places depending on ports the case of municipalities having the case. Our results corroborate the ex- the lesser spatial contact with neigh- istence of a cluster of variable high values bors, so we find 4 cases connected only in the central corridor of the state and in with one neighbor: Coper, Covarachía, peripheral zones of municipalities with and . low agricultural production, which

Table 2. Global Spatial Autocorrelation Test Moran I test under randomization

Data: logagr Weights: wqueen Moran I statistic standard deviate = 6.4978, p-value = 4.075e-11 Alternative hypothesis: greater Sample estimates: Moran I statistic Expectation Variance 0.397446002 -0.008474576 0.003902559

The Moran test of spatial autocorrela- prevalent in a specific geographical enti- tion makes up a very standard contrast ty depend on the behavior of variable in for verifying spatial dependence. For neighboring spaces. For our purposes, such purpose the null hypothesis is the Moran test rejects the hypothesis of defined in terms of absence of spatial the no existence of spatial dependence autocorrelation and in case of rejection, in our endogenous variable agricultural there would be evidence that values production.

223 Spatial Patterns of Agriculture in Boyacá Helmuth Yesid Arias Gómez • Gabriela Antosová

Figure 7. Moran’s Dispersion Diagram of Agricultural Output in Tons. Source: IGAC layers. Own elaboration using R software.

In spatial analysis any variable spatially axes the values of the logarithm of lagged are defined as the weighted av- agricultural production and in the other erage of the neighboring values. This hand, the spatially lagged value of the concept casts light on the possible ex- same variable. The diagram shows a istence of spatial dependence in values positive relationship, suggesting the of variable distributed across different existence of spatial autocorrelation in polygons in space. In such case we our variable of interest. According to suppose that for diverse geographical Chasco (2003), the slope of regression entities the values of analyzed variable line corresponds to statistic of Moran, must be similar if they are located suf- concomitantly, more open the angle ficiently close, suggesting that the eco- formed with the abscise, highest degree nomic phenomena behaves in a similar of spatial autocorrelation. way across neighboring spaces. For our purposes, we are interested to bear out Econometric Strategy that agricultural production has similar values across close municipalities in For empirical procedures hence we will order to confirm the existence of a con- use as endogenous variable the agricul- tagion effect across towns. tural output in tons for 119 municipali- ties during 2005, and as exogenous the In Figure 7 is showed the Moran Dis- rural population in the year 2005 and persion Diagram representing in two the distance to Tunja in kilometers.

224 Apuntes CENES Volumen 37, Número 66, ISSN 0120-3053 julio - diciembre 2018, 203 a 237

Table 3. Descriptive Statistics

In terms of econometric procedures and 4, which indicate the basic behavior of as it is commonly used, the first step is the model. Value of adjusted R2 0.328 to estimate an OLS regression assuming suggests that the fit of model not jet null values for autocorrelation parame- includes relevant information about ters (ρ=λ=0). With such assumption the some explicative variables on model, contrast of spatial autocorrelation is but it is an acceptable value in cases of performed, and in case of finding evi- models with cross section or data panel dence of this phenomenon, a Lagrange arrangement. Both exogenous variables Multiplier test must be applied in order turn out significant. to choose the most suitable structure for spatial dependence (Helsen, 2008). Mathematical sign of the coefficient of Precisely OLS estimation is useful as relative distance to Tunja means that the benchmark and Lagrange Multiplier central towns as important producers tests offer the criteria if the model re- are located at short distances regarding quires to be extended for estimating regional main city where cold climate spatial interaction effects (Elhorst, predominates. The other coefficient 2010). indicates that for small scale produc- tion, the preferential market is the local In this case the regression equation is market and that the bulk of output pro- similar to the conventional technique duced with larger scales is conducted to by OLS and results are shown in Table Bogotá.

225 Spatial Patterns of Agriculture in Boyacá Helmuth Yesid Arias Gómez • Gabriela Antosová

Table 4. OLS Regression (Model 1)

Table 5. Autocorrelation Test on Residuals

226 Apuntes CENES Volumen 37, Número 66, ISSN 0120-3053 julio - diciembre 2018, 203 a 237

The next step consists in applying stan- The Table 6 shows the contrasts com- dard Lagrange Multiplicator contrasts monly used for choosing the correct for corroborating the presence of spa- structure of spatial dependence. In tial autocorrelation, and to determine doing so, two additional tests help to criteria for evaluating the pertinence understand the character of spatial of a spatial lag or spatial error patterns dependence: LMlag and LMerr tests in order to define the kind of spatial which indicate the significant results in dependence. In case of dubitation, ro- the case of spatial lag model. The right bust Lagrange Multiplicator contrast is choice of spatial dependence is crucial promptly performed. for analysis provided that this decision could conduct to diverse interpretation of the spatial dependence.

Table 6. Lagrange Multiplicators Contrasts

227 Spatial Patterns of Agriculture in Boyacá Helmuth Yesid Arias Gómez • Gabriela Antosová

Table 6 shows the results of Lagrange SPATIAL LAG AND SPATIAL Multiplier tests as definitive criteria ERROR MODELS for choosing the best description of the spatial dependence. Elhorst (2010) In tables 7 and 8 are presented two summarizes diverse approaches for results of models, the first one corre- tackling the estimation of spatial mod- sponding to the case of the spatial error els, applying three possible strategies: model and the second reports spatial lag Keleijan and Prucha, Lesage and Pace model. and Anselin. Accordingly, with the sug- gestion we choose Anselin´s approach If the spatial autocorrelation is present composed by several steps. Firstly, we only during error term, the model to estimate the OLS model and afterwards estimate corresponds to: in concordance with contrasts of La- grange Multiplier (and for a definitive [3] decision Robust LM), the disjunctive is to add or not spatial terms to regression and finally to decide if this spatial term must correspond to spatial error or spa- tial lag. Where: u is a withe noise term and λ becomes the autoregressive parameter. In fact, according to Table 6, Lagrange Multiplicator indicates significance In the model [3] if there is no omission either for the error model as for the lag of lag in endogenous variable of model, model, so we are forced to interpret the it corresponds to the case of residual Robust LM contrasts. Now, following spatial autocorrelation, then a scheme of the last test, the contrast of error model spatial dependence in error term must loses significance and in consequence, be included (Moreno & Vayá, 2002). we consider that the best approach for describing the true data-generation pro- The standard treatment of econometrics cess corresponds to spatial lag approach. in the presence of spatial autocorrelation consists in performing a regression via Maximum Likelihood. As it is known, in this technique estimator is calculated maximizing the logarithm of a function of likelihood associated with a spatial model (Moreno & Vayá, 2002).

228 Apuntes CENES Volumen 37, Número 66, ISSN 0120-3053 julio - diciembre 2018, 203 a 237

Table 7. Spatial Error Model (Model 2)

In the case of spatial lag it is necessary In equation [4] can be observed that if to specify the following model: spatial lag is omitted in exogenous or endogenous variables, the feature of spatial dependence will be transmit- [4] ted to error term, in a case quoted as substantive spatial autocorrelation. To handle this case, it is necessary to in- Where: y is a vector (Nx1), Wy is a spa- clude a spatial lag of dependent variable tial lag of variable, X is the exogenous with spatial autocorrelation (Moreno & variables matrix, u is a white noise error Vayá, 2002). term, N is the number of observations and ρ is a autoregressive parameter. This kind of models is named also as Models of Communication or Contagion and integrates all autocorrelation struc-

229 Spatial Patterns of Agriculture in Boyacá Helmuth Yesid Arias Gómez • Gabriela Antosová ture into spatial lag, as an explicative this kind of model, estimation commits argument of the endogenous variable. In a specification error that biases estima- the case of omission of weight matrix in tors and drives to invalid inference.

Table 8. Spatial Lag Model (Model 3)

Comparing tables 4, 7 and 8, the sig- in the presence of spatial autocorrela- nificance of estimators is maintained tion because that effect invalidates the between OLS regression and two modes standard test when OLS regression are of correction of spatial autocorrelation, estimated, in a similar way as do serial but Bernat warns about such situation autocorrelation or heteroskedasticity

230 Apuntes CENES Volumen 37, Número 66, ISSN 0120-3053 julio - diciembre 2018, 203 a 237

(Bernat, 1996). In lag model (Table 7), that the influence of neighboring plac- the rho parameter indicates the mag- es is superior to exogenous variables. nitude of effect originated in neighbor This is an additional argument for the municipalities on an endogenous vari- presumption that lag model is the most able defined as agricultural output. adequate for our analysis.

According to the results of both models In the presence of spatial autocorrela- for correcting spatial autocorrelation, tion the criteria derived from r squared the adequate specification for describing regarding to fitness of the model, cannot the spatial dependence of the model is be applied trustfully, so it is necessary the lag model, so the influence of neigh- to use other criteria for evaluating the boring areas on local manufacturing good fit of the model. In this exercise, presence requires a special interpreta- we check the result offered by Robust tion. In these cases, the spatial lag mod- Lagrange Multiplicator criteria, and the el indicates that agricultural output at elements offered by test conducted to each municipality is directly influenced choose the spatial lag model. by the size of agricultural production in neighboring localities and moreover, CONCLUSIONS such spatial effect is independent re- garding the set of exogenous variables. Both gravitational and econometrical When rho parameter is high, the model exercises suggest several interesting has a substantial spatial dependence and conclusions. The gravitation model municipalities show strong interest in showed an important productive zone the performance of agricultural output in the core of the state where munici- in neighboring places (Bernat, 1996). palities share some common features. So, evidence indicates that an import- They are located in high cold lands and ant and positive impact on the size of commodities correspond to cold climate the agricultural production exists in products. But according to econometric municipalities if neighboring towns and gravitation principles the desti- are prominent producers of agricultural nation of output could be quoted as commodities. bimodal. Cold climate production per- formed with high scales is conducted to As can be appreciated in Table 8, coef- larger national markets as Bogotá and ficient rho is significant and has a high Bucaramanga. Larger productive towns magnitude, reinforcing the influence have quick geographical and logistical of neighbor agricultural towns on the connection with those markets using tons of local agricultural output. The important paved highways (as shown in magnitude of rho is neatly higher than Figure 3). Higher prices in larger cities the coefficients of relative distance to attract strongly the output from Boyacá. Tunja and rural population, suggesting

231 Spatial Patterns of Agriculture in Boyacá Helmuth Yesid Arias Gómez • Gabriela Antosová

All gravitation principles admit some ities and less influential agricultural friction forces that exert outward activities. pressure. According to econometric parameters, distance to Tunja has a neg- Farther towns in the Provinces of Oc- ative impact on agricultural production. cident and Gutiérrez show weak spatial Consequently, if we combine all infor- interaction. Firstly, distance from the mation from gravitational results and big cities and scarcity of connectivity econometrics, we can assert that Bogotá hamper commercial flows with big mar- and diverse markets beyond regional kets, in consequence, small production borders of Boyacá attract a substantial with reduced scales are sold in weak part of output produced in this state to local markets. Secondly, in some points municipalities located in central areas agricultural production performs a sec- very close to Tunja. ondary role in the local economy.

In fact, the bulk of agricultural pro- Peripheral zones in the state face an duction in tons is concentrated in cold enormous market constraint. Precari- lands in the center of the state, located at ous conditions of communication with high geographical positions. Cold lands important markets compel peasants to are located at high geographical points, put production in local markets that, by so highness is a strong factor affecting small size, inhibit the development of positively the agricultural production in sectorial scale economies. In fact, in the Boyacá. Some of the important towns vast territories of Boyacá, local markets having a preponderant role in potato are insufficient in size for developing production are: Samacá (2604 m), Toca a broad scale of production. Spatial (2750 m), Siachoque (2760 m), Ven- interaction could be intensified trough taquemada (2630 m) y Jenesano (2075 transportation, communication and as- m). sociations across farmers and peasants.

A clear interaction is specially estab- Regarding commercialization of output, lished in the central zone of the state collectors and intermediates accom- where the cold climate crops mainly plish a role of arbitrage in agricultural potato predominate. Moving to east, markets. Relative close position across strong interaction is detected in towns urban centers in Boyacá bestows upon as Aquitania and Tibasosa, typical pro- them a privileged position for getting ducers of onion. It is possible that weak full information on prices, so they can spatial intensity detected around Duita- take advantage of information and play ma and Sogamoso could be explained an arbitrage function. Hiking prices and by the presence of dominant mining, high population density attract the bulk manufacturing and commercial activ- of production.

232 Apuntes CENES Volumen 37, Número 66, ISSN 0120-3053 julio - diciembre 2018, 203 a 237

The spatial econometric model indicat- output and its spatial distribution. Mu- ed what kind of spatial dependence ex- nicipal rural population is important ists when agricultural output is used as because small peasant’s output is sold in an endogenous variable. The evidence his own town. In fact, for tiny producers of spatial autocorrelation confirmed that local weekly farmers markets and local local agricultural output is strongly in- demand can be strongly influential. fluenced by the size of sectorial output Distance to the main city, Tunja, has a in neighboring localities. negative coefficient indicating that the closer cold lands are the relatively most This kind of spatial dependence is easy productive area and emerge as a rele- to understand for agricultural com- vant agricultural cluster well connected modities because climatic and natural by fluent paved highways. conditions are shared by nearby munic- ipalities, so there must be coincidence An overwhelming conclusion that in the productive profile between mu- emerges throughout the analysis is the nicipalities and other places located in enormous negative impact of the high proximity. transport cost of the state´s economic performance. Figure 3 shows that the According to results of spatial econo- main highway system has a radial de- metric the most appropriated interpre- sign structure having as a central point tation of spatial dependence deployed Tunja and in consequence, in surround- by the variables of model is concordant ing areas highly productive crops devel- whit spatial lag. oped. But in further zones of peripher- ies we can find agricultural production, Exogenous variables in model are also although with scarce productivity and useful to understand the agricultural with reduced scale economies.

233 Spatial Patterns of Agriculture in Boyacá Helmuth Yesid Arias Gómez • Gabriela Antosová

REFERENCES Anderson, J. (2010). The Gravity Model. Working Paper 16576. Cambridge, MA: National Bureau of Economic Research. Anselin, L. (2001). Spatial Econometrics. In B. Baltagi (ed.), Companion to Theoretical Econometrics (pp. 310-330). Oxford: Blackwell Publishing. Arfa, N., Rodríguez, C., Daniel, K. & Shonkwiler, S. (2010, March 10-11). Spatial Structure of Agricultural Production in France. In OECD Workshop on the Disaggregated Impacts of CAP Reform. Arias, H. & Antosova, G. (2015). Perfil espacial de la economía boyacense. Revista Apuntes del Cenes, (59). Asohofrucol. (2013). Plan de negocios de cebolla. Bogotá: Asohofrucol. Bernat, A. (1996). Does Manufacturing Matter? A Spatial Econometric View of Kaldor’s View. Journal of Regional Science, 36(3), 463-477. Bongiovanni, R. (2009). Econometría espacial aplicada a la agricultura de precisión. Actualidad Económica, (67). Chasco, C. (2003). Métodos gráficos del análisis exploratorio de datos espaciales. Anales de economía aplicada. Almería, España: Asociación Española de Economía Aplicada. De Noronha, E. & De Noronha, T. (2010). Gravitational Models and Spatial Foresight: From Agricultural Policy to Agricultural Loss. Advances in Urban Rehabilitation and Sustainability, 124–127. DNP. (2007). Agenda interna para la productividad y la competitividad. Documento regional: Boyacá. Bogotá: DN P. Dubin, R. (1998). Spatial Autocorrelation: A Primer. Journal of Housing Economics, 7, 304-327. Duranton, G. & Overman, H. (2005). Testing for Localization Using Micro- Geographic Data. Review of Economic Studies, 72(4), 1077-1106. Elhorst, J. (2010). Applied Spatial Econometrics: Raising the Bar. Spatial Economic Analysis, 5(1), 9-28. Fedepapa & Ministerio de Ambiente, Vivienda y Desarrollo Territorial. (2004). Guía ambiental para el cultivo de la papa. Bogotá: Fedepapa.

234 Apuntes CENES Volumen 37, Número 66, ISSN 0120-3053 julio - diciembre 2018, 203 a 237

Gobernación de Boyacá. (2012). Plan Departamental de Desarrollo 2012-2015: Boyacá se atreve. Tunja: Gobernación de Boyacá. Gobernación de Boyacá. (2014). Informe de ejecución: Plan Departamental de Desarrollo 2012-2015: Boyacá se atreve. Tunja: Gobernación de Boyacá. Helsen, J. (2008). Essays on the Spatial Analysis of Manufacturing. Ph. D. Dissertation Ken State University. Graduate School of Management. Isard, et al. (1998). Methods of Interregional and Regional Analysis. Ashgate: Aldershot. Krugman, P. (1991). Increasing Returns and Economic Geography. Journal of Political Economy, 99. Krugman, P. (1992). Geografía y comercio. Barcelona: Antoni Bosch. Krugman, P. (1995). Desarrollo, geografía y teoría económica. Barcelona: Antoni Bosch. Krugman, P. (2008). The Increasing Returns Revolution in Trade and Geography. Nobel Prize. Lecture. Stockholm. Retrieved from www.nobelprize.org Ministerio de Agricultura y Desarrollo Rural & IICA (1997). Instrumentos, mecanismos e institucionalidad para la comercialización de productos agrícolas en Colombia. Bogotá: Minagricultura. Moreno, R. & Vayá, E. (2002). Econometría espacial: nuevas técnicas para el análisis regional. Una aplicación a las regiones europeas. Investigaciones Regionales, 1, 83-106. Plan Frutícola Nacional –PFN- (2006). Desarrollo de la fruticultura en Boyacá. Tunja: s.n. Proexport. (2014). La Revista de las Oportunidades. Boyacá. Bogotá: Proexport. Rodríguez, L. & Bermúdez, L. (1997). Bases para el estudio de la competitividad de la producción de papa de las agroempresas campesinas de la región central de Boyacá. Agronomía Colombiana, 14(1). Secretaría de Fomento Agropecuario. (s.f.). Política agropecuaria departamento de Boyacá. Con Visión 2020. Tunja: Gobernación de Boyacá. Superintendencia de Industria y Comercio. (s.f.). Cadena productiva de la papa: diagnóstico de la libre competencia. Bogotá: SIC. Tobler, W. (1975). Spatial Interaction Patterns. IIASA Research Report. Luxembourg. 235 Spatial Patterns of Agriculture in Boyacá Helmuth Yesid Arias Gómez • Gabriela Antosová

ANEXO 1 Own elaboration on basis IGAC. Department Map Boyacá: Political of Source:

236 CODE CODE CODE CODE CODE CODE CODE CODE NAME NAME NAME NAME REGION TOWN REGION TOWN REGION TOWN REGION TOWN 15 1 Tunja 15 238 Duitama 15 507 Otanche 15 740 Siachoque 15 22 Almeida 15 244 El Cocuy 15 511 Pachavita 15 753 Soatá 15 47 Aquitania 15 248 El Espino 15 514 Páez 15 755 Socotá 15 51 Arcabuco 15 272 Firavitoba 15 516 Paipa 15 757 Socha 15 87 Belén 5 276 Floresta 15 518 Pajarito 15 759 Sogamoso 15 90 Berbeo 15 293 Gachantivá 15 522 Panqueba 15 761 15 92 Beteitiva 15 296 Gámeza 15 531 Pauna 15 762 Sora 15 97 15 299 Garagoa 15 533 Paya 15 763 Sotaquirá 15 104 Boyacá 15 317 Guacamayas 15 537 Paz de Río 15 764 Soracá 15 106 Briceño 15 322 15 542 Pesca 15 774 Susacón 15 109 Buenavista 15 325 Guayatá 15 550 Pisba 15 776 Sutamarchán 15 114 Busbanzá 15 332 Guicán 15 572 Puerto Boyacá 15 778 15 131 Caldas 15 362 Iza 15 580 Quípama 15 790 Tasco 15 135 Campohermoso 15 367 Jenesano 15 599 Ramiriquí 15 798 15 162 Cerinza 15 368 Jericó 15 600 Ráquira 15 804 Tibaná 15 172 Chinavita 15 377 15 621 Rondón 15 806 Tibasosa 15 176 Chiquinquirá 15 380 La Capilla 15 632 Saboyá 15 808 Tinjacá 15 180 Chiscas 15 401 La Victoria 15 638 Sáchica 15 810 Tipacoque 15 183 Chita 15 403 La Uvita 15 646 Samacá 15 814 Toca 15 185 Chitaraque 15 407 15 660 San Eduardo 15 816 Togüí 15 187 Chivatá 15 425 15 664 San José de Pare 15 820 Tópaga 15 189 Cienega 15 442 Maripí 15 667 15 822 Tota 15 204 Cómbita 15 455 Miraflores 15 673 San Mateo 15 832 Tununguá 15 212 Coper 15 464 Mongua 15 676 San Miguel de Sema 15 835 Turmequé 15 215 Corrales 15 466 Monguí 15 681 15 837 Tuta 15 218 Covarachía 15 469 Moniquirá 15 686 Santana 15 839 Tutasá 15 223 Cubará 15 476 Motavita 15 690 Santa María 15 842 Úmbita 15 224 Cucaita 15 480 Muzo 15 693 Santa Rosa de Viterbo 15 861 Ventaquemada 15 226 Cuítiva 15 491 Nobsa 15 696 Santa Sofía 15 879 Viracachá 15 232 Chíquiza 15 494 Nuevo Colón 15 720 Sativanorte 15 897 Zetaquira 15 236 15 500 Oicatá 15 723 Sativasur Source: DANE & DIVIPOLA