Are food safety standards different from other food standards? A political economy perspective

1. Introduction

A variety of standards are increasingly affecting the global food system for several reasons (Fulponi, 2007). Consumer concerns related to food safety, food quality, and environmental and social issues in food and agricultural production processes have all contributed to the emergence of a broad set of standards. This process has been reinforced by globalisation and new technologies which also affect the global spread and increase in standards (Unnevehr, 2000; Reardon and Berdegué, 2002; Maertens and Swinnen, 2007).

An important issue is whether such standards are welfare enhancing and, if not, why they are introduced. Bockstael (1984) and Fischer and Serra (2000) argue that standards observed in reality are not socially optimal. One argument is that, as trade agreements restrict the use of traditional trade protectionist instruments, standards are non-tariff barriers to trade (Baldwin, 2000; OECD, 2001). For example, Anderson et al. (2004) and Fulton and Giannakis (2004) argue that genetically modified (GM) food standards are used as protection against imports. Another argument is that standards are suboptimal because they are part of a ‘race to the bottom’, for example in order to attract foreign investors (Wellisch, 1995; Kunce and Shogren, 2005), or a ‘race to the top’ (see Jaffe et al., 1995). Swinnen and Vandemoortele (2008) show that the politically optimal choice of standards is affected by a combination of various factors including levels of income, consumer preferences, implementation costs, the media, a standard's effect on comparative advantage, and the ability of institutions to enforce standards. Hence, while these factors may also affect the social optimality of standards set by public authorities, the political choice may not be socially efficient.

An important unanswered question is whether the findings in the literature are sensitive to the nature of standards. All existing studies have either analysed a specific case (such as GM regulations) or used a general concept of ‘food standards’. None of the studies has analysed how the results may depend on the nature of standards, for example when one distinguishes between different types of food standards. Yet this is an important question, in particular, if one considers the increasing importance of standards in trade conflicts. Some of the main conflicts in international trade policies relate to food safety regulations, such as the bovine growth hormone case, the GM food case, and the (in)famous aflatoxine case. However, other types of standards, such as social and ethical standards, also affect trade. The growing influence of quality, environmental, and social standards on trade has caused Pascal Lamy, the Director-General of the World Trade Organisation, to publicly warn that this proliferation of ‘green’ and other standards could complicate trade negotiations (Minder, 2007). For all these reasons, it is important to examine whether the nature of standards affects the political choice of the standard and how it may distort trade.¹

The objective of our paper is therefore to develop a theoretical model of how the nature of food standards affects the political economy of standards, in particular how food safety standards (FSS) compare to other types of food standards. The model in this paper builds on the analytical framework of Swinnen and Vandemoortele (2008) which is extended in several ways. We explicitly model the effects of each different type of standard. An important contribution of this paper is that it explicitly integrates risk in a political economy setting. As far as we know, this is the first paper that incorporates risk into the analysis of the political economy of standards, not only in the literature on food markets but also more generally.

The paper is organised as follows. In Section 2, we present a formal model to analyse the impact of food standards on consumers – taking risk considerations into account – and producers. We integrate the interests of consumers and producers in a political economy model in which the government sets food standards and is influenced by contributions from lobby groups. Section 3 discusses the resulting political equilibrium. In Section 4, we classify different types of food standards and compare them in terms of political optimum. Section 5 discusses the relation between the nature of food standards and trade conflicts, and Section 6 concludes.

2. The model

2.1. Consumption, risk, and food standards

Various approaches have been used to model how consumers are affected by standards. First, some authors in the literature on standards (such as Bockstael, 1984) assume that consumers can costlessly observe product characteristics, while others (such as Leland, 1979) assume that standards apply to characteristics of which the assessment by consumers requires additional costly information. In the latter case, standards can improve upon the unregulated market equilibrium by reducing the information asymmetry between consumers and producers. Second, other studies (such as Copeland and Taylor, 1995; Fischer and Serra, 2000; Anderson et al., 2004; Besley and Ghatak, 2007) model standards as having no direct impact on consumers' individual utility. Instead these authors define standards as regulations that have an impact on consumption externalities.

In the literature on (food) standards, the effect of a public standard is generally modelled as cost-increasing for domestic producers as well as for foreign producers (interested in) exporting to the country that imposes the standard (Suwa-Eisenmann and Verdier, 2002; Henson and Jaffee, 2007). Swinnen and Vandemoortele (2008) incorporate both direct consumer effects and externalities in their analysis, and they examine the effects on different parts of the economy. We extend this model by introducing different types of standards and by including risk considerations.

The second component of expression (1), Eu(c, s), is the expected utility of consuming c of the good with standard s produced in the import-competing sector. The application of expected utility in relation to product characteristics is based on the literature on product warranties (see e.g. Cooper and Ross, 1985; Emons, 1988; Elbasha and Riggs, 2003). If the product contains the characteristics desired by the consumer, utility of consumption is u(c) which is assumed to be increasing and concave in c (u′(·) > 0 and u″(·) < 0).² However, if the good has inferior characteristics, consumers incur a disutility of consumption, d(c), which is increasing and convex in c (d′(·) > 0, and d′′(·) > 0). The probability of the latter outcome is ρ(s) which is decreasing and convex in s (ρ′(·) < 0 and ρ′′(·) > 0), with a higher s referring to a more stringent standard. Hence a higher standard reduces the probability that the product has inferior characteristics and induces consumers to consume more of the product, ceteris paribus. For example, consumers who perceive health risks related to certain (potential) ingredients or production processes may increase consumption if they are guaranteed the absence of these elements through a food standard.

It follows from expression (2) that the perceived risk, which consists of the perceived risk probability λ^cρ(s) and the magnitude of the adverse consequences, u(c) + d(c), can be defined as λ^cρ(s)[u(c) + d(c)]. Such two-dimensional specification of risk is consistent with, among others, Peter and Ryan (1976), Yeung and Morris (2001), and Pennings et al. (2002).

2.2. The economy

We consider a small open economy where domestic firms are price takers and domestic prices equal world prices. The numéraire good is produced with labour (l₀) as the only input and we assume that the sector experiences constant returns to scale and an input-output ratio of one. The price of the numéraire good is normalised to unity and hence the wage w also equals one. Production in the import-competing sector is a function of the production factor labour l and of an inelastic sector-specific input k such that q = f(l,k). All profits made in the sector (Π_p) accrue to the owners of the sector-specific input.

We assume that when the country imposes a food standard, the production costs of the imported goods also rise as the standard applies to all goods sold in the country. Consequently, the price on the domestic market – which equals the average production cost of foreign importers – increases in the standard (∂p/∂s > 0).

2.3. The political model

3. The political economy equilibrium

The first term in expression (8) captures how producers are affected by a change in the level of the standard weighted by their lobbying strength (1 + α_p). This impact may be positive or negative depending on whether the marginal increase in costs, ∂g/∂s, is smaller or larger than the marginal increase in revenues, (∂p/∂s)q. When the marginal change in costs is smaller than the marginal revenue effect, the producers will contribute more to support a more stringent standard. Because the marginal revenue effect is driven by the marginal cost increase of foreign importers (implicit in the price increase ∂p/∂s), the effect of a more stringent standard on domestic producers depends on the relation between their marginal costs of the standard and those of their foreign competitors.

The effect of a more stringent food standard on consumer surplus weighted by the consumers' lobbying strength is also uncertain. Consumer surplus of the representative consumer will increase if the sum of the marginal risk reduction, −λ^cρ′(s) (u(c) + d(c)), and the marginal warm glow effect, λ^gγ ′, is larger than the marginal increase in cost of consumption, (∂p/∂s)c, which results in higher contributions from consumers. Vice versa, if the marginal increase in consumption expenditures outweighs the marginal risk reduction and marginal warm glow effect, consumers will contribute less such that the standard in the political equilibrium will be lower.

4. The nature of food standards and the political optimum

4.1. Classification of food standards

In our theoretical analysis, we distinguish between three types of food standards: FSS, food quality standards (FQS), and social and environmental standards (SES). This three-way categorisation of food standards is based on Brom (2000).

The first category of food standards is FSS. According to Brom (2000), this is the type of standards that matter to all consumers. The main purpose of FSS is to provide consumers with safe food. Examples of such standards are the limitation of pesticide residues on vegetables and, more generally, the prohibition of harmful substances in any food. We assume that FSS only have an effect on producer profits and the expected utility of consumers, and induce no warm glow effect.

The second category is FQS. These standards ensure certain product quality characteristics to the consumers. These product characteristics do not include safety, but rather concern consumer preferences about other aspects of nutritional quality, taste, colour, size, etc. These standards may be linked to personal lifestyle choices (e.g. vegetarians). As with FSS, we assume that these FQS affect producer profits and consumers' expected utility but have no warm glow effect.

The third category covers public standards that regulate social and environmental issues based on the ethical values of a society. Examples of SES include the prohibition of using child labour and the limitation of carbon dioxide emission in the production process. We assume that these SES have no direct effect on expected consumer utility (i.e. they do not reduce risk), but affect consumers through their warm glow effect (see e.g. Besley and Ghatak, 2007). As with FSS and FQS, standards concerning social and environmental issues affect producer profits because of cost-increasing requirements concerning production practices.

We should point out that this categorisation is a theoretical construct useful for our analysis which allows us to trace the impact of different effects. In reality, however, many food standards have characteristics belonging to different categories. Consider organic food standards for example (see e.g. Lusk and Briggeman, 2009). Consumers may consider food labelled as organic as safer because pesticides are not used in the production of organic food (a food safety aspect in our classification), but also as more tasteful (a food quality standard in our classification). In addition, society may benefit from reduced pesticide use (a social and environmental standard in our classification).⁷ However, for the purpose of this paper, the mutually exclusive classification adopted is useful. Obviously, it is important to take these considerations into account when applying this analytical framework for empirical research.

We use this classification to analyse the impact of the nature of food standards on the political optimum. First, the positive marginal risk reduction effect, −λ^cρ′(s) (u(c) + d(c)), only applies to types of standards that directly affect consumers by reducing the risk of consuming a good with inferior product characteristics. In our classification of food standards, this is the case for FSS and FQS, but not for SES. The marginal warm glow effect, λ^gγ ′, is zero for FSS and FQS. In the case of SES, the marginal risk reduction effect is zero, but the marginal warm glow effect is positive.

Second, we assume that the disutility d(c) of consuming goods with inferior food safety characteristics is larger than the disutility of consuming goods with inferior food quality characteristics. Because of the larger negative impact of the potential food safety hazard, an increased FSS will have a larger effect on direct consumer utility than an increased FQS, ceteris paribus. Although we are not aware of any formal statistical tests comparing consumer reactions to changes in safety and quality of food (or standards), substantial empirical evidence does suggest that consumer reactions are stronger in cases of food safety issues. In several cases of (perceived) food safety crises, strong consumer reactions have been documented. For example after the outbreak of the BSE (‘mad cow’) disease in 1996 in the UK, meat consumption in France declined dramatically (Latouche et al., 1998). Similarly, with the dioxin crisis in Belgium in 1999, consumer reactions were dramatic (Verbeke and Ward, 2001; Swinnen et al., 2005). Our assumption is also in line with the result of Lusk and Briggeman (2009) that consumers rate food safety as the most important food value compared to other values related to food quality such as nutrition and taste. In our model, these differences in valuation of food characteristics are captured by the differences in disutility d(c) from consuming less safe food and food of low quality. This reflects that the disutility of consuming a contaminated apple (food safety) is arguably larger than the disutility of consuming an apple that is not tasteful (food quality).

4.2. The nature of food standards and the political optimum

We now consider how the nature of standards affects the political economy of public standards. Using equation (8), we can analyse how the type of food standards may affect the political equilibrium and thus the level of the politically optimal public standard. Differences in the level of food standards according to which category the standard belongs to may arise for different reasons. These are captured by the different terms in the first-order condition of the government (equation (8)) which determines the politically optimal public food standard. More specifically, the political optimum of the different standards depends on: (i) the (exogenous) relative lobby weights of the different political groups (captured by the relative values of α_p and α_c); (ii) the lobby contributions of the different groups which reflect (differences in) the impact of the different standards; and (iii) whether the various standards differentially affect other variables such as the consumer perception variables.

A first result is that public FSS will be higher than public FQS. As neither FSS nor FQS induce any warm glow effect, the third term of the marginal effect on consumers in equation (8) is zero and the equilibrium for FSS and FQS is determined by the other terms. Under our assumption that the costs of implementing FSS and FQS are the same, the net impact on domestic producers of FQS and FSS is identical. With the marginal risk reduction, −λ^cρ′(s) (u(c) + d(c)), of FSS being stronger than that of FQS,⁸ceteris paribus, it follows that the political equilibrium standard will be higher for FSS than for FQS, i.e. s*(FSS) > s*(FQS).

A second result is that a comparison of the political optimum of FSS or FQS with SES cannot yield an unconditional result without more detailed information on how much the different standards marginally affect expected consumer utility and the warm glow effect. Whether the government sets SES higher or lower than FSS or FQS depends, ceteris paribus, – and given our assumption of identical implementation costs for all standards – on the relative marginal effects of the different standards on the risk of inferior product characteristics, −λ^cρ′(s) (u (c) + d (c)), (for FSS or FQS) and the marginal warm glow effect λ^gγ′ (for SES).

Third, if perceptions are biased this also affects the outcome. For example if consumers perceive food safety hazards to have a higher probability of occurrence than food quality problems (i.e. if λ^c(FSS) > λ^c(FQS)), this would reinforce the earlier result that s*(FSS) > s*(FQS) and further increase the gap between the levels of FSS and FQS.⁹ While there is little evidence on these issues, Ansell and Vogel (2006) argue that exaggerated negative perceptions are particularly important in food scares. If so, this would lead to a further relative increase of FSS compared to other food standards. Alternatively, if consumers perceive the risk effects correctly (λ^c = 1 for FQS and FSS) but perceive the environmental impact of standards as more positive than they actually are (λ^g > 1), this would increase the level of SES relatively to FQS and FSS.

Finally, one can further refine our classification of standards. For example within the class of FQS, one can distinguish between different quality characteristics that are affected by the standard. Nelson (1970) and Darby and Karni (1973) identify three types of food quality characteristics, namely search, experience, and credence characteristics. Search characteristics can be evaluated prior to purchase. Hence, a FQS does not reduce any informational asymmetry or risk of buying a product with inferior quality characteristics (see also Bockstael 1984) and the marginal risk reduction, −λ^cρ′(s) (u(c) + d(c)), is zero. In contrast, a FQS may reduce informational asymmetries and increase expected utility if it concerns experience (quality characteristics are known after consumption) or credence quality characteristics (quality characteristics cannot be known without additional costs). As a consequence, politically optimal FQS related to experience and credence characteristics will be higher than FQS related to search characteristics. The politically optimal level of FQS will thus depend on the type of quality characteristics and informational asymmetries that the FQS reduces.

5. The nature of standards and trade conflicts

To analyse how the nature of public food standards affects trade conflicts, we must start by defining when standards are distorting trade. For this purpose, we use the same basic framework of identifying (sub)optimal trade policy as is used in traditional trade theory, which is by comparing with the socially optimal trade policy. The political equilibrium is said to be suboptimal when the politically optimal tariff t* differs from the socially optimal tariff t^#. More specifically, in a small open economy, the socially optimal tariff is equal to zero while a positive tariff distorts trade and hence reduces social welfare.

The interpretation of the different terms in this first-order condition is analogous to the interpretation of the terms in the political optimum condition given by equation (8).

It is clear from comparing equations (8) and (9) that the politically optimal standard s* will equal the socially optimal standard s^# if α_p = α_c and λ^c = λ^g = 1 in the political equilibrium.¹¹ If one of these conditions is not fulfilled, this may result in ‘over-standardisation’ (s* > s^#) or in ‘under-standardisation’ (s* < s^#), with uncertain effects on trade. For example, if the cost of implementing the public food standard is lower for domestic producers than for foreign producers (∂g/∂s < (∂p/∂s)q at s^#, i.e. the domestic implementation cost is lower), domestic producers will lobby in favour of increasing the standard since it gives them an advantage over their foreign competitors.¹² This will result in over-standardisation (s* > s^# if α_p > α_c) which may reduce trade and potentially cause trade conflicts. However, if the cost of implementing the standard is lower for foreign importers (∂p/∂s)q < ∂g/∂s at s^#), or if enforcement of the standard on foreign producers is less effective than on domestic producers, then domestic producers will lobby (more) against the standard and the opposite, i.e. under-standardisation will occur (s* < s^# if α_p > α_c). This will result in reduced trade as lower standards reduce the cost advantage that foreign producers would have with more stringent standards.

The issue we are particularly concerned with in this paper is whether one type of standard is more likely to serve as a trade protectionist instrument than others.¹³ First, as we discussed above, there is some limited evidence that the bias in perceptions differs between standards. In particular, exaggerated negative risk perceptions of food safety problems and their impact on consumer welfare (and thus exaggerated positive effects of FSS) may be higher than such biases in perceptions for FQS. In this case, politically optimal FSS would be higher than their social optimum. For example, many argue that European consumer reactions to GM food are based on biased risk perceptions of the impact of GM technology on food safety. Note, however, that this observation in itself does not provide evidence for the argument that FSS will differ more from their social optimum than other food standards since GM food opposition is only partially related to the ‘food safety’ aspect of GM food. Perceptions on environmental and quality aspects of GM food are important as well and it is not obvious whether these perceptions are less or more biased than those on the food safety aspect.

Furthermore, whether a relatively stronger bias in food safety risk perceptions would effectively imply (stronger) protection for domestic producers depends on other characteristics of the economy. For example, if domestic producers and foreign producers are equally affected by a FSS, then this standard is not protectionist. However, in case higher FSS would effectively counteract a comparative disadvantage of domestic producers, then the standard could be considered protectionist. The latter case is related to the argument of Anderson et al. (2004) who argue that trade policies excluding GM foods from the EU market are hurting foreign producers as this technology would reinforce their comparative advantage.

A second element in the analysis of whether various standards affect trade differently relates to the power of the different lobbies. It is clear from comparing conditions (8) and (9) and from the discussion above that if one of the lobby groups has a relatively larger weight, this would lead to a divergence of the politically optimal standard towards a level of the standard preferred by that particular lobby group. However, there is no ex ante reason why these weights should different for varying types of standards. Moreover, different lobby weights for varying types of standards would only lead to differences in trade protection under certain conditions, as explained above.

Finally, notice that one of the main findings of the previous section, i.e. that due to the consumption effects the level of a FSS would be higher than the level of a FQS in the political equilibrium, does not necessarily imply that there would be more trade distortions with FSS than with FQS. It is evident that when for both types of standards the political equilibrium coincides with the social equilibrium there is just as much trade distortion with FSS as with FQS, since then there is no over- (or under-) standardisation for neither standard. Notice that this result would hold regardless of whether or not the politically optimal FSS is higher than the politically optimal FQS. Only if the lobby weight of domestic producers (α_p) is higher than the lobby weight of consumers (α_c), and under specific conditions of relative implementation costs such that the standard provides benefits to domestic producers over foreign importers would FSS be associated with more protectionism than FQS.

6. Conclusions and implications

Standards are increasingly important in global food markets and are increasingly the subject of trade conflicts. Only a few studies have analysed the political economy of standards as the causal mechanism of suboptimal standard setting in food markets. In this paper, we have analysed whether the nature of standards affects the level of the politically optimal standard and whether the nature of standards affects the likelihood of trade conflicts. Specifically, we have examined whether FSS are different from other standards, such as FQS or SES in a political economy setting. In studying these issues, we looked at situations when there are only public food standards set by the government.

The politically optimal level of various standards depends on their impact on consumers and producers, and on the relative lobbying strengths of those affected by the standards who thus lobby to increase or reduce the standards. An important result is that, in an environment with only public standards, FSS will be set at higher levels than FQS because the former, ceteris paribus, have a larger marginal effect on consumer welfare. The relative level of SES compared with food safety or quality standards depends on the specific characteristics of the different food standards which determine how the standards marginally (and differently) affect consumer welfare.

Another central result is that the relative level of the various standards does not imply that the most stringent ones are also necessarily the principal trade barriers – or protectionist instruments. The relationship between food standards and protectionism is complex and depends not just on the nature of the standards but also on the interaction between the standards and differences in implementation costs, production costs (comparative advantage), and enforcement between domestic and foreign producers. An important result from our analysis is that the relation between the type of food standard and protection is conditional on other factors. For example, if domestic producers have relatively more political influence – in combination with a set of production-side conditions that imply that more stringent standards provide benefits to domestic producers over foreign importers – one should expect FSS to be relatively more important trade protectionist instruments than FQS. Another example is when risk perceptions of consumers are biased. If risk perceptions are biased in such a way that they reinforce the perceived impact of standards on consumers, consumer lobbying will be more intensive and this may lead to the implementation of a trade distorting standard, but again conditional on a set of structural conditions on the supply side.

An overall conclusion from this analysis is that the relationships between the nature of food standards, public decision-making concerning food standards, and the trade effects of standards are complex. An important implication is that there is considerable room for more theoretical as well as empirical research in this area. Obviously, the theoretical models can be further elaborated and refined. For example, one line of research would be the modelling of the relation between private standards set by individual firms and public standards set in this political economy framework. Importantly, the analysis in this paper also identifies an important need for better empirical work in this field. Systematic data on various types of food standards, their level, and how public decision-making results in these standards are missing and would be very useful to provide further guidance (stylised facts) on refined theoretical research.

Acknowledgements

The authors gratefully acknowledge the valuable comments and suggestions by three anonymous referees and the guest editors. This research was financially supported by Research Foundation – Flanders (FWO) and the K. U. Leuven Research Fund (EF).

References

Author notes