FOOD TRADE AND BIODIVERSITY EFFECTS

• Author: Cecilia Bellora,Jean‐Marc Bourgeon
• DOI: http://doi.org/10.1111/iere.12408
• Published date: 01 November 2019
• Date: 01 November 2019

INTERNATIONAL ECONOMIC REVIEW

Vol. 60, No. 4, November 2019 DOI: 10.1111/iere.12408

FOOD TRADE AND BIODIVERSITY EFFECTS∗

BYCECILIA BELLORA AND JEAN-MARC BOURGEON1

CEPII, France; INRA and Ecole Polytechnique (CREST), France

Pests create biodiversity effects that increase food production risks and decrease productivity when agri-

cultural production is specialized. Pesticides contain these effects, but damage the environment and human

health. When opening to trade, governments are tempted to restrict pesticide use because, with more food

being imported, less pesticide is needed for domestic consumption. However, pesticide restrictions hinder the

competitiveness of their agricultural sector on international markets. We show that restrictions on pesticides are

more stringent under free trade than under autarky, which reduces the gains from trade, and that trade increases

food price volatility.

1. INTRODUCTION

Agricultural prices are historically more volatile than manufacture prices (Jacks et al., 2011).

Perhaps because this stochasticity is considered to be due to factors beyond human control, such

as weather conditions, economic studies analyzing the determination of food price focus mainly

on factors related to market organization, such as demand variability and the role played by

stock management.2However, in addition to abiotic factors, such as water stress, temperature,

irradiance, and nutrient supply, which are often related to weather conditions, food production

is also impaired by biotic factors, also known as “pests”—including animal pests (such as insects,

rodents, birds, etc.), pathogens (such as viruses, bacteria, fungi, etc.), or weeds. These harmful

organisms can cause critical harvest losses: The estimations of global potential yield losses for

wheat, maize, and rice, the three most produced cereals in the world, vary between 50% and

70% (Oerke, 2006).3The impact of pests on yields is linked to the degree of specialization of

the agricultural sector, which depends on the country’s openness to trade. The more cultivation

is concentrated on a few high-yield crops, the more pests specialize on these crops and the

greater their virulence. Yields become more variable and the probability of low harvests rises.4

∗Manuscript received February 2016; revised November 2018.

1The authors are grateful to Brian Copeland, two anonymous referees, and the editor for their comments and

suggestions. They also thank seminar participants at Ecole Polytechnique, at the 2014 BioEcon conference, at the

2014 conference of the French Association of Environmental and Resource Economists, and at the 2015 CESifo

Area Conference on applied microeconomics. The research leading to these results has received funding from the

European Union’s Seventh Framework Program FP7/2007-2001 under Grant Agreement 290693 FOODSECURE.

The authors only are responsible for any omissions or deficiencies. Neither the FOODSECURE project and any of

its partner organizations nor any organization of the European Union is accountable for the content of this article.

Much of this work was done while Cecilia Bellora was a Ph.D. student at Universit´

e de Cergy-Pontoise and INRA.

Please address correspondence to: Jean-Marc Bourgeon, INRA, 16 rue Claude Bernard, 75005 Paris, France. E-mail:

bourgeon@agroparistech.fr.

2See Gilbert and Morgan (2010) and Wright (2011) for overviews on food price volatility and examinations of

its causes.

3Oerke (2006) defines “potential loss” as a loss occurring when no pests control management procedures are used

at all. Savary et al. (2000) and Fernandez-Cornejo et al. (1998) provide lower but nevertheless significant estimates of

yield losses caused by pests.

4The link between crop diversity and the spread of pests has widely been investigated in the ecology literature,

both theoretically and empirically. Zhu et al. (2000) are often cited for their empirical evidence of successful disease

control in rice cultivation: in 10 townships of the Yunnan province in China, thanks to a widespread crop diversification

program conducted in 1998 and 1999 and mixing different genotypes of rice, the severity of rice panicle blast caused

1957

C

(2019) by the Economics Department of the University of Pennsylvania and the Osaka University Institute of Social

and Economic Research Association

1958 BELLORA AND BOURGEON

These effects are very much reduced by the use of agrochemicals such as pesticides, fungicides,

herbicides, and the like: For example, agrochemicals reduce potential losses of wheat by 50%

(actual average losses are about 29%, with a minimum loss of 14% in northwestern Europe).

But agrochemicals generate negative externalities, on human health, biodiversity, water, and air

quality, which is a growing concern.5The necessity of pesticide use is called into question under

trade because of Not in My Back Yard (NIMBY) considerations. Indeed, food grown locally

that is sold abroad exposes the local population to pesticide externalities without benefiting

them personally. Besides, since some of the food consumed locally is imported, pesticides that

were used under autarky are no longer needed under free trade. When opening to trade, the

government is faced with a trade-off: Restricting the use of pesticides satisfies NIMBY concerns

but also reduces the competitiveness of the agricultural sector. Increasing awareness of the

negative externalities of pesticide use augments the weight of NIMBY considerations in public

decisions, and the use of pesticides seems to follow a decreasing trend (Bexfield, 2008; ECP,

2013; Ryberg and Gilliom, 2015).6A marked reduction in the use of pesticides would have clear

environmental benefits but could also increase food prices and price volatility, adding to the

effects linked to food demand and stock management.

The aim of this article is to analyze how crop biodiversity and environmental policies interact

with trade. This formal description of the mechanisms at stake is also the first detailed exami-

nation of the potential role of biodiversity in the behavior of food prices. We develop a simple

model of farm production affected by biotic factors that vary with specialization to represent

the impacts of crop biodiversity on agricultural productivity and on the pattern of trade in a

Ricardian two-country setup. We single out these impacts by assuming that the use of pesticides

is regulated by an environmental tax with no distributional effects, and we abstract from risk

aversion by assuming that farmers and consumers are risk neutral.7

Our analysis provides three main findings. First, although countries have different compara-

tive advantages under autarky, biodiversity effects lead to incomplete specialization under free

trade. Indeed, as specialization reduces expected crop yields, some crops are produced by both

countries because their agricultural sectors end up with the same productivity at equilibrium.

Second, when factor endowments of the countries are not too dissimilar, the trade-off in the

design of environmental policies results in restrictions on pesticides more stringent under free

trade than under autarky: NIMBY considerations eclipse the market share rivalry between

the two countries. As a result, gains from trade are reduced, and countries may even experience

welfare losses when opening up to trade. We quantify these effects for the United States and

its main trading partners. We show that NIMBY impacts are rather large in the United States.

Taken alone, they would induce a 12-fold increase in the ratio of the environmental tax to the

by a fungus was reduced by 94% and yield increased by 89% compared to monoculture. The role played by species

diversity in limiting disease and weed dissemination has also been documented (Knops et al., 1999; Mitchell et al., 2002;

Smith et al., 2008; Davis et al., 2012). For example, Smith et al. (2008) report that in the absence of any pesticide use,

yield doubled compared to monoculture due to diverse crop rotations.

5Pimentel (2005) reports more than 26 million cases worldwide of nonfatal pesticides poisoning and approximately

220,000 fatalities. He estimates that the effects of pesticides on human health cost about $1.2 billion per year in the

United States. Mammals and birds are also affected. Farmland bird population decreased by 25% in France between

1989 and 2009 (Jiguet et al., 2012), and a sharp decline was also observed in the European Union as a whole during the

same period (EEA, 2010). Pesticides also contaminate water and soil and significantly affect water species both locally

and regionally (Beketov et al., 2013).

6Many countries have recently adopted regulations that forbid the use of the most harmful molecules, have set

provisions on the use and storage of pesticides, and promote their sustainable use. For example, only a fourth of the

molecules previously marketed passed the safety assessment made during the European review process that ended in

2009 (EC, 2009). Correlatively, demand for organic farming is rapidly increasing. In Europe, sales of organic products

are estimated to be around €23 billion in 2012, a 6% increase from 2011’s level (Schaack et al., 2014) and farmland

devoted to them nearly doubled between 2005 and 2016. In the United States, sales exceeded $34 billion in 2014 and

more than tripled between 2005 and 2014 (USDA-ERS, 2015). By replacing synthetic pesticides with natural ones and

reducing their use, organic farming has a smaller environmental impact (Tuomisto et al., 2012) but also lower yields

(Seufert et al., 2012) than conventional farming.

7We elaborate on the effects of risk aversion in Section 7.

FOOD TRADE AND BIODIVERSITY EFFECTS 1959

other production costs compared to autarky and a welfare loss. However, concern about the

competitiveness of the agricultural sector strongly limits NIMBY behavior. Overall, the relative

tax increase ranges between 4% and 26%, depending on the strength of the biodiversity effects.

As a result, food trade leads to a sharp increase in the U.S. welfare thanks to a decrease in food

prices and a better environment, which compensates for the reduction in agricultural revenues.

Third, food price behavior depends on the pattern of trade. Trade increases the production

volatility of crops produced by both countries. Country-specific crops for which comparative

advantages are large could see a reduction in their volatility, but that supposes very small bio-

diversity effects. Average prices of country-specific crops are increased for consumers of the

producing country. This is because of more restrictive environmental policies and the reduced

yields resulting from the specialization of production under free trade. For crops produced by

both countries, the sharing of production determines the change in average prices.

Our work is related to different strains of literature. The link between crop biodiversity, yield,

and revenue variability is empirically investigated in Smale et al. (1998), Di Falco and Perrings

(2005), and Di Falco and Chavas (2006). These studies find sometimes contrasting results

but generally tend to show that increasing agricultural biodiversity is associated with higher

production and lower risk exposure (Di Falco, 2012). We add to this literature an economic

foundation of the mechanisms at stake.8We build on Weitzman (2000) to model farm production

with biodiversity effects: The larger the share of farmland dedicated to a crop, the more its

parasitic species proliferates and thus the more fields of that crop are at risk of being wiped out.9

Weitzman (2000) uses this model to solve the trade-off between the private and social optima,

the former tending to specialize on a few varieties while the latter aims to preserve biodiversity.

We depart from his work by considering a trade context, incorporating the use of pesticides,

and investigating the impact of biodiversity effects on production and price distributions. Our

setup is a Ricardian trade model with two countries and many goods, `

a la Dornbusch et al.

(1977, hereafter DFS). In this context, pests create external decreasing returns to scale (DRS)

in the agricultural sector, which generates increasing marginal costs in a perfect competition

setup. A number of papers have studied external economies of scale in Ricardian models. Ethier

(1982) characterizes what Grossman and Rossi-Hansberg (2010) call “pathologies” generated

by increasing returns to scale (IRS), in particular multiple equilibria and a reverse pattern of

trade.10 With IRS, higher productivity leads to an increase in the industry scale that, in turn,

improves productivity, creating a snowball effect responsible for these pathological results.

To avoid these effects, Grossman and Rossi-Hansberg (2010) assume Bertrand competition.11

This is not necessary with DRS since the industry scale reduces productivity. However, the

external scale effects we consider cause incomplete specialization. In this Ricardian setup, we

find an impact of trade on the strength of environmental policies. Previous literature has shown

that international market share rivalry tends to weaken environmental policies (Barrett, 1994):

By weakening environmental policies, the government reduces the marginal cost of domestic

firms, making them more competitive on international markets. However, governments may

also be tempted to reduce polluting activities at home when the same products are produced

abroad: Markusen et al. (1995) and Kennedy (1994) show that governments are induced to

increase their environmental tax. Both effects are at work in our context, and we show that

the latter is the main driving force in the setting of the environmental policy: When countries

8For more details on the biological mechanisms involved, see Tilman et al. (2005), who use simple ecological models

to describe the positive influence of diversity on the biomass produced and corroborate their findings with empirical

results detailed in Tilman and Downing (1994) and Tilman et al. (1996).

9Weitzman (2000) makes an analogy between parasite–host relationships and the species–area curve that originally

applies to islands: The bigger the size of an island, the more species will be located there. He compares the total biomass

of a uniform crop to an island in a sea of other biomass. A large literature in ecology uses the species–area curve that

is empirically robust not only for islands but also, more generally, for uniform regions (May, 2000; Plotkin et al., 2000;

Drakare et al., 2006; Garcia Martin and Goldenfeld, 2006; Storch et al., 2012).

10 Indeed, with IRS, production can be pushed toward the lowest cost producers as well as toward firms with higher

costs but larger size, which allows them to remain competitive.

11 Lyn and Rodr´

ıguez-Clare (2013) complete the setup by refining the way transport costs are taken into account.