Agricultural Production and Biodiversity in France: A Spatial Analysis

Hermann Pythagore Pierre Donfouet

CESAER-INRA, UMR 1041

Phone number: +33 07 51 39 49 55

Fax: +33 02 99 38 80 84

Corresponding author’s email:

Cécile Détang-Dessendre

CESAER-INRA, UMR 1041

Aleksandra Barczak

CESAER-INRA, UMR 1041

[Very preliminary. Please do not quote].

Acknowledgement: This paper is a component of Towards RUral Synergies and Trade-offs between Economic development and Ecosystem services (TRUSTEE) project, task 5.4: Economic development and ecosystem services bundles. The website of TRUSTEE is: http://www.trustee-project.eu/. We acknowledge funding from the European Union under the seventh framework program. We would like to express our warmest gratitude to the “Observatoire du Développement Rural” (ODR) for providing the data. We are grateful to Eric Cahuzac (ORD-INRA) and Muriel Tichit (AgroParis Tech-INRA) for data discussions. The usual disclaimer applies.

The authors declare no conflict of interest.

Abstract: Ecosystem services play a major role in the production of agricultural goods and contribute to the essential ecological functions which agriculture depends. In France, in the context of climate change, the concept of ecosystem services is experiencing a resurgence of interest among policymakers and is increasingly integrated into public policy. But, so far there is no study which investigates the effect of biodiversity on agricultural production while accounting for spatial dependence. This paper fills the gap by providing an empirical evidence of the effect of biodiversity on agricultural production and spillover effect. Results of the study based on spatial models suggest that biodiversity has a positive and significant effect on agricultural production and its marginal contribution is substantial when rainfall is low in the agroecosystem. More importantly, spatial dependence is a major issue and could be explained by the topographic, climatic and agronomic constraints. We discuss these findings.

Keywords: Biodiversity, agricultural production, spatial dependence.

JEL code: C21, Q1, Q57.

1. Introduction

In many countries, agricultural productivity is increasingly recognized as a positive driver for food security, climate change, poverty reduction and political stability. In Europe, the Common Agricultural Policy was launched to improve agricultural productivity of farmers, provide affordable food for European citizens and huge amounts of money has been released to sustain the agricultural sector to ensure food security and stimulate farmers to produce goods which are sustainable and environmentally-friendly. French agriculture is the most dominant agricultural sector in Europe and employs for the year 2012 about 3.3% of the workforce (AGRESTE, 2013). The agricultural products that place France among the world top producers are: milk, cereals, wine, and sugar beets to name a few. Most of these agricultural products are produced by the ecosystem. In the literature, there is a clear link between agricultural production and ecosystem services (Di Falco and Chavas, 2008, 2009). As outlined by Gardiner et al. (2009), Kremen et al. (2004), ecosystem within agricultural lands could provide services of biological pest control and pollination that may promote agricultural production.

Ecosystem services are defined as “the conditions and processes through which natural ecosystems, and the species that make them up, sustain, and fulfil human life” (Daily, 1997, p. 6). They contribute to the essential ecological functions which agriculture depends, including erosion control and sediment retention, soil formation, genetic resources, water regulation and supply (Costanza et al., 1997). They also offer a wide variety of aesthetic, recreational and cultural services to the human welfare. In France, in the context of climate change, the concept of ecosystem services is experiencing a resurgence of interest among policymakers and is increasingly integrated into public policy. Therefore, the French government has established the new French national biodiversity strategy over the period 2011-2020 which aims at preserving, restoring, strengthening and enhancing biodiversity in the sustainable and equitable use by involving all the stakeholders in different sectors.

As defined by UNEP (1993), biodiversity is the “variability among living organisms from all sources, including terrestrial, marine, and other aquatic ecosystems and the ecological complexes of which these are part: this includes diversity within species, between species, and of ecosystems”. Biodiversity is the foundation of ecosystem services. It is often seen as public goods which means that individuals cannot be effectively excluded from use (non-excludable) and where use by one individual does not reduce availability to others (non-rivalrous). Hence, markets do not reflect the full social costs or benefits of biodiversity and their management may be complex. Nevertheless, biodiversity valuation can help scholars and policymakers deal with this market failure by assigning a monetary value that reflects the social importance of biodiversity and this could help designing effective tools for their management. From an economic viewpoint, assessing the value of biodiversity may be done with a variety of valuation approaches (Barbier, 2007; De Groot et al., 2002; Farber et al., 2002; TEEB, 2010). The first approach is the direct market valuation approach which uses data from the actual market. The market price-based approach, cost-based approach and production function-based approach are components of the direct market valuation approach. The revealed preference is the second approach where individuals reveal their preferences for biodiversity in existing markets. The travel cost method and hedonic pricing technique are parts of this approach. The third approach is the stated preference method which simulates markets where researchers can infer the value that individuals attach to biodiversity. The contingent valuation method, discrete choice experiment and contingent ranking are the main types of stated preferences method.

In the current study, we use the first approach mainly the production function-based approach where we assume that biodiversity is an input in the production process of agricultural goods which are themselves marketed and we attempt to assess its contribution to agricultural production while accounting for spatial dependence. Furthermore, taking ecosystem resilience into account, we consider worthwhile to investigate the extent to which biodiversity can act as a catalyzer to agricultural production mostly when rainfall is scant. Biodiversity could play an important role in the ecosystem resilience. Resilience refers to an ecosystem's capacity to recover from disturbances or unexpected shocks and maintain its essential functions (Holling, 1986).

To the best of our knowledge, we do not know of any study in France which examines the effect of biodiversity on agricultural production. Another important shortcoming in the literature is the scarcity of studies which integrates spatial dependence in the analysis. The observed agricultural production may be influenced by the agricultural production of the geographically nearest neighboring agricultural land. First, farmers can be part of a large network and exchange information on agricultural practices that could improve their productivity. Thus, due to exchanges of information in the network, copy-catting and learning effect, the levels of agricultural production in a region (county) may be influenced by those in neighboring regions (counties). This effect may be relevant at a fine scale, such as the parcel or plots and may be irrelevant at an aggregated level. Second, spatial externalities may arise and spillover across boundaries or due to some unobserved factors that are spatially correlated. For instance, the cluster pattern of agricultural production may be explained by some natural, historical, socio-cultural and institutional components. Therefore, not accounting for spatial dependence may bias the estimates and lead to erroneous policy recommendations. Hence, the paper contributes to the existing knowledge by shedding some lights on the effect of biodiversity on agricultural production in France with some significant spillovers effects across the neighborhood. From policy perspective, a better understanding of the factors that may influence agricultural productivity could give more insights on how the policymakers could intervene via some incentives to protect some agricultural lands and biodiversity as well.

The overall objective of the study is to examine the effects of biodiversity and other factors on agricultural production while accounting for spatial dependence. More specifically, we aim via econometric tools at measuring the impact of biodiversity on agricultural production on a given Small Agricultural Regions[1] (SAR) and other contiguous SARs. Results of the study suggest that biodiversity has a positive and significant effect on agricultural production and its marginal contribution is substantial when rainfall is low in the agroecosystem. More importantly, spatial dependence is not at odds with the data.

The paper is structured as follows. In Section 2, we review the literature on the link between agricultural production and biodiversity. Section 3 provides the econometric model and discusses the data while Section 4 presents the results of the study. We conclude the study in Section 5 with some policy recommendations.

2. Literature review

The relevance of biodiversity in the provision of ecosystem services has been fully documented in the literature. Tilman et al. (2005) demonstrated that the plant diversity (number of plant species added to plots) improve plant primary productivity. Reich et al. (2001) had shown that higher plant diversity leads to greater carbon (CO2) storage in plant and lower levels of nitrate in ground waters. Hajjar et al. (2008) gave an exhaustive survey of the links between crop genetic diversity and ecosystem services such as: (i) pest and disease management, (iii) enhances pollination services and soil processes, (iii) proving continuous biomass cover, aids carbon sequestration and prevents soil erosion. The debate has focused on the principle mechanisms that might explain this benefit of plan diversity. Two explanations are found in the existing literature: sampling and complementarity effects. The sampling effects are thought as increasing plant diversity will eventually increase the likelihood of some species to adapt well to some particular pedo-climatic conditions (Tilman et al., 2005). A second explanation known as complementarity effects are perceived when particular species perform better in the presence of others (Chavas and Di Falco, 2012; Loreau and Hector, 2001). This complementarity leads to a form of division of labor and a better collective exploitation of available resources such as soil mineral and light. This complementarity is also better apprehended under crop rotation and diversification schemes. Hence, they might help reduce pathogens and pests that often occur when one single crop is used. Furthermore, they add nutrients to the soil (some farmers rotate nitrogen-fixing crops such as legumes with non-fixing crops such as maize which need nitrates) and protect the soil against erosion. Farmers might then be less prone to use artificial fertilizers which impair biodiversity.

Most existing studies which explored the effect of biodiversity on agricultural production used crop diversity index such as the Shannon, Margalef and Simpson indices[2] as a measure of biodiversity. Bonneuil et al. (2012) discussed the interests and the limits of these different indicators. In particular, they do not take into account all dimensions of genetic diversity, focusing on species richness and evenness of their proportional abundance. High nature value farmland (HNV) index is another measure of biodiversity in rural areas. It is widely accepted by the European commission but the data are scarce for many countries. HNV comprises the hot spots of biological diversity in rural areas, characterized by extensive farming practices. According to Andersen et al. (2003, p.4), HNV farmland is described as: “those areas in Europe where agriculture is a major (usually the dominant) land use and where that agriculture supports or is associated with either a high species and habitat diversity or the presence of species of European conservation concern or both”. A thorough search in the literature suggests that bird population seems also to be a proxy for biodiversity since they provide regulative ecosystem services such as seed dispersal, pollination and predation/pest control (Civantos et al., 2012; Mäntylä et al., 2012).

Over the past decades, some authors have investigated the effect of biodiversity on agricultural production. Di Falco and Chavas (2008) used a production function to explore the effects of crop biodiversity on the production of durum wheat in eight regions in Southern Italy from 1970-1993. Their dependent variable was the quantity of durum wheat produced and the covariates where the labor force, fertilizer, capital (expenditure in machinery and buildings), land to cereal, annual rainfall) and Shannon index for biodiversity. The authors controlled for the unobserved individual effects, endogeneity of biodiversity and state dependence of agricultural production by means of a dynamic panel model. Biodiversity is positively and significantly related with production both in current and lagged effects. By interacting biodiversity and rainfall, the results indicate that biodiversity contributes to the production of durum wheat when the agroecosystem faces lower or scarce rainfall. In a similar study conducted in Ethiopia, Di Falco et al. (2010) examined the impact of crop biodiversity on food production using panel data and instrumental variable (IV) estimator. In their model, the biodiversity[3] is considered as an input in a standard household production function. Other inputs: household labor, land, fertilizer, manure and improved seeds, climatic factors such as rainfall and farm characteristics (soil fertility and slope) were controlled for. The dependent variable is the production of all cereals and pulse crops grown by farmers in a particular year. In their two-stage least square (2SLS) technique, land tenure security, distance between plots and the farm, gender are used as instrument for biodiversity. Results of the study suggest that increasing the number of crop variety increases production and the effect of biodiversity is stronger when rainfall is lower. Though the results of the study reveal that too much of fertilizers are detrimental to agricultural production, labor, soil fertility, plot slope and improved seeds do not have any effect in the food production.

The productive value of crop biodiversity was assessed by Chavas and Di Falco (2012) in the highlands of Ethiopia. The main research question was to examine the sources that generate positive linkages between crop diversity and productivity by using a parametric production function on a large dataset from a farm survey conducted over the period 1999-2000 in Tigray region. Three outputs were used: Teff, barley and wheat. The standard inputs in the production function used as covariates were animal traction, land, labor, fertilizer, rainfall, soil fertility, soil erosion, slope, location, use of improved seeds, and the presence of soil conservation practices. Village fixed effects were also included to control for unobserved heterogeneity. The instruments used for output variables were farm agroecological heterogeneity, land share under conservation measures and distance from the input supplier (access to the seeds market). Results from the 2SLS show that the conventional inputs such as land, labor, animal traction, and fertilizer were all positive and statistically significant. The interaction term between barley and wheat was positive and statistically significant, implying the presence of positive interaction effects across crops. This finding highlights the presence of complementarity effect in the agroecosystem. Hence, each crop tends to have a positive effect on the marginal productivity of other crops. Other studies also found the positive effect of biodiversity on agricultural production (Heisey et al., 1997; Smale et al., 2003; Smale et al., 2002)