Lecture 36.
The economics of food safety

---

What we want to learn about this topic

Economists are interested to ask two question about food regulations:

• are they efficient? In other words, can we show that the benefits exceed the costs of regulation? Is current legislation the most efficient way of achieving the desired level of risk?
• are they barriers to trade? In other words, does consumer protection pose a risk of becoming trade protectionism?

This lecture addresses the first of these questions. We ask first what is the rationale for government intervenion, and we examine the range of intervention options open to governments to pursue. We also discuss some of the issues in the economic evaluation of existing or proposed interventions. Lecture 35 discusses the protection versus protectionism issue.

Short introduction to the issues

The market for food safety

Food is produced with different safety attributes (i.e. different levels of risk). For example, sushi (raw seafood) or steak tartare (raw mince) is more likely to contain food-borne disease that a well-done steak. It is in this sense that markets exist for food safety or, more precisely, markets exist for foods with varying safety attributes (Antle). So why can we not rely on the market alone to produce the level of food safety we desire?

Consider first an idealised description of the market for food safety. Consumers are willing to pay for food safety reflecting the benefits they derive through reduced pain and suffering and longer life expectancy. There is a downward-sloping demand curve, indicating that if food safety can be provided cheaply, consumers will demand a lot, but if it is expensive (e.g. organic food) they will demand less. There is an upward-sloping supply curve, because it is more costly for firms to provide increasing levels of food safety. There is an equilibrium food safety level, indicating that rational consumers will generally not demand 100% safe food. Even if this was technically possible, it would be likely to be prohibitively expensive.

However, this equilibrium will only be an optimum if the market is perfect and, in particular, if consumers correctly evaluate the risks they are facing and have full information about the quality of the food product they are buying. Quality characteristics are often divided into three:

• search characteristics. Search dimensions are those where the quality can be ascertained by the buyer at the time of purchase, like the appearance of an apple. Both the producer and the consumer know exactly what they are selling or buying. The purchase decision is made with certainty and there is no market failure.
• experience characteristics. Experience characteristics are those where the quality can be ascertained only after the purchase, like the taste of an apple. Accordingly, here the consumer is faced with a decision under risk. In the case of risk, the 'lemons problem' occurs. In this, the market for food safety is like the market for used cars. Sellers have more information about the quality of the car than buyers. Because buyers often cannot tell the difference between a good and a bad used car, both good and bad cars must sell at the same price and the seller of a good car is unable to extract a premium for quality. Therefore there is less willingness to supply quality. In the same way, there is a tendency for food safety to be undersupplied by the market because consumers are not always able to distinguish between high and low food quality. Note, however, that markets may still function if there are repeat purchases, and producers have an incentive to invest in reputation. Indeed, even with single purchases, market mechanisms can work provided consumers can exchange information at low cost (consumer magazines). Mechanisms which may be used to build up reputation include brands, marks, labels or warranties. Where these conditions hold, public intervention may be needed to prevent consumer deception and to protect brands, but no direct public intervention is needed to improve the efficiency of markets for experience attributes.
• credence characteristics. Credence characteristics are those where the average consumer can never ascertain the quality by himself or herself, but has to trust the judgement of others - like whether the apple is healthy or organically produced. Consumers are not be able to assess quality even after purchase. There may be a long time lag between ingestion and effect, e.g. when quality involves the chemical composition of the food, contamination with toxic chemicals or the presence of microorganisms. Under these conditions, it is difficult for firms to establish reputations for quality. There is a danger of Gresham's law applying and bad products driving out good. This is the most clearcut case for public regulation.

These arguments suggest food safety will be undersupplied by the market and that public regulation may help to create a more efficient market. Another, different, information argument is that consumer perceptions of risk may differ markedly from those of experts. Experiments show that consumers routinely underestimate risks from certain diet-related cancers or coronary heart disease while systematically overestimating the risk of botulism or food additives.

There may also be strong externalities that justify government intervention, either on the production side (one rogue producer who fails to meet adequate food standards can put the reputation of an entire national food industry at risk) or on the consumption side (an infectious food-borne illness imposes wider costs on society that transcend those incurred by the individual consumer).

Systems of food quality control

A wide range of safety control systems has evolved for the typical food product being offered for sale (Henson and Caswell). Public interventions include:

• direct ex ante regulation in the form of standards, inspection, product testing and other programmes attempts to ensure the quality of the product by specifying how it is produced and/or its final quality. Businesses that are found not to meet the standard are penalised, for example, through fines or court actions.
• ex post regulation such as product liability that punishes companies that produce products of insufficient quality through damage awards to those harmed by their actions.

Private systems include self-regulation and various forms of certification by other parties.

• Self-regulation includes internal control systems that assure product quality, where the company sets, monitors and self-certifies the control parameters. It can take place at the level of the individual firm or be instituted by trade organisations that cover the majority of market supply.
• Certification involves the setting of product quality standards and their monitoring and certification by third parties, for example, customers, industry trade associations, or standards-setting organisations such as ISO. Such certification can be voluntarily sought by the company or required by those with whom it does business.

Approaches to public food safety regulation

Consumer information (labelling). Appeals because it does not restrict consumer choice (you can be informed that there is a risk of eating T-bone steak but you are not prevented from eating it) and is efficient in that it recognises that risk preferences are heterogeneous across the population. But labels can require a lot of time and expertise on the part of the consumers to become adequately informed (e.g. labels on food additives - do we know what they mean in terms of health?) and we may prefer that the government simply makes these decisions on our behalf. Also, labelling and education address the consumers' lack of information, which is one of the market failures which lead to food safety problems, but it does not address the second form of market failure - that consumers may demand less food safety than society would like.

Food standards. Food safety standards can take three main forms.

• Target standards do not prescribe any specific safety standards for a supplier's products or the processes by which they are produced, but impose criminal liability for pre-specified harmful consequences that arise from their products. For example, a target standard lays down the requirement that food sold for human consumption must be safe.
• Performance standards require certain levels of safety to be achieved when the product is supplied, but leave suppliers free to choose the mechanisms through which they meet such conditions. Examples of performance standards include a requirement that a product have residues below a specified maximum level, or a definition of the levels of contamination that are deemed unacceptable.
• Specification standards are applied both to products (product standards) and the processes by which those products are made (process standards) and can take positive or negative forms by either compelling products to contain certain ingredients or the use of particular production methods (pasteurisation for fresh milk), or prohibiting the use of particular ingredients or production methods.

Prior approval. Approach is used for drugs and certain food ingredients like flavourings and colourings. Producer has to show that the product is safe before it can be allowed on sale.

Product liability (litigation). Product liability relies on demonstrating that a food product has caused harm to
the consumer and the victim obtaining redress through the courts (DEFRA, 1999). Three elements must be present:

• Harm of some kind has been caused to the consumer of the product.
• There is a defect in the product - it has not performed as it should be expected to under normal usage conditions.
• A causal relationship can be demonstrated between the product defect and the harm to the consumer.

In the case of safety, product liability normally takes the form of either a negligence or strict liability standard (DEFRA, 1999). A negligence standard is where a supplier is liable to pay damages where it can be shown that harm has been caused by a negligent action on their part. With strict liability, however, it is sufficient to show that harm has been caused to the consumer, regardless of whether it can be demonstrated that the supplier was negligent.

In the case of food safety, product liability is generally not regarded as a sufficient mechanism for regulating food given that the courts are only able to act retrospectively to compensate an individual who has suffered harm. It may be difficult for consumers to demonstrate that any harm they have suffered has resulted from the consumption of a particular food product. This is particularly the case with chronic diseases where there may be a long period of time between consumption of the product and the onset of ill health. Consumers may not have sufficient resources to obtain redress through the legal system. For example, 'no win-no fee' has not developed widely as a system of legal funding in the UK. This contrasts with the United States where 'no win-no fee' is commonplace and tort liability is a more important element of food safety regulation (DEFRA, 1999).

Evaluation of alternative public food safety interventions

The key concept in this literature is the idea of 'regulatory efficiency'. Four regulatory principles can be derived:

• there are often market mechanisms which can be encouraged to plug the market failure. In the case of food safety, we have seen that these include liability and product quality reputation, or simply providing additional information. The argument is made that these incentives-based approaches allow consumers to choose the most efficient level of food safety.
• regulations are justified only if they pass a benefit-cost analysis test. There is an institutional issue of who should do this - the agency that produces the standards or an independent body?
• informed individual choice of safety is preferred to statutory safety standards when risk preferences are heterogeneous. No one wants to buy contaminated meat but some people are prepared to take more risks than others. Thus, rather than set high standards for pesticide residue, set minimum standards, but allow foods that are produced without pesticides (organic products) to be labelled as pesticide free. Labelling allows people who prefer a higher degree of food safety to choose such foods without imposing the costs on everyone.
• even where regulation is justified, the costs of regulation can be minimised by the appropriate design of regulation. In particular, the argument is made that performance standards are less costly than product/process standards (also called design standards). Don't tell firms how to do it, set targets they must achieve and leave them decide. The argument is made that this encourages technical innovation to try to achieve the standards at least cost.

Empirical studies

There is now a growing number of studies which attempt to measure the benefit-cost ratio of individual food regulations. These studies are encouraged by the growh in regulatory impact assessment (in the US, such studies are mandated by law where the regulatory impact on the economy is likely to exceed $100 million). The issues which arise in these studies include:

• measurement of benefits. The benefits of food safety regulation are reductions in risks of morbidity and mortality associated with consuming foods that could be contaminated. For exporting countries there may also be market benefits in terms of access to higher priced export markets. There are several approaches to valuing health risks. The most basic and most often used approach to value morbidity is to estimate the cost-of-illness (COI). The COI approach is based on the medical costs of an illness plus the foregone market income due to lost work time. From a theoretical perspective, although the approach is appealing, it is not the same as people's willingness to pay (WTP) to avoid illness, although it can be shown to be a lower bound of this WTP under certain conditions. Risks of death can be valued using the value of a statistical life. Various methods have been used to infer the value individuals place on risk of death, such as discounting foregone income, and using wage differences between occupations with different risks. The different methods produce a wide range of values of a statistical life. Finally, contingent valuation studies can be used to estimate the willingness to pay for food safety. Apart from the general problems with the validity of this approach, specific issues arise in their use for food safety valuation. For example, they often measure willingness to pay to avoid specific risks that are not generalisable. A further issue is the distinction between tangible and intangible costs. The latter include things like pain and suffering arising from a food-borne illness, and are very difficult to measure, and are often ignored.
• measurement of costs. Here the key issue is to ensure that all costs are appropriately measured. These will include not just the direct costs of control, but also the indirect costs borne by firms who must make adjustments to production processes.

Industry often opposes government reguations because they increase costs and reduce production. However, with regulation, consumers may be more willing to buy the food products, since they are now getting a safer food for their money. Individual firms may also be able to turn high safety standards to a marketing advantage, extracting a safety premium from consumers which is sufficient to remunerate the higher costs involved.

Recent developments

In recent years governments have reviewed their approaches to food safety regulation . There has been a shift towards performance-based measures that afford suppliers greater flexibility to achieve the desired level of food safety in the most efficient manner. Simultaneously, recognising that end-product testing is an inefficient form of food safety control, requirements have been progressively put in place for process controls based on the principles of hazard analysis critical control point (HACCP). HACCP is widely recognised in the food industry as an effective approach to establish good manufacturing practices for the production of safe food. This is achieved by establishing process controls through the identification of points in the production process that are most critical to monitor and control.

Reading suggestions

Core readings

DEFRA, 1999, 'The rationale for government intervention', Section 2 of University of Reading, Economic Evaluation of UK Policy on Chemical Contaminants in Food, London.

Mitchell, L., 2003, Economic theory and conceptual relationships between food safety and international trade, in Buzby, J., ed., International Trade and Food Safety: Economic Theory and Case Studies, Agricultural Economic Report 828, Economic Research Service, US Department of Agriculture.
(concentrate on p. 10 -17 which discusses the economics of food safety and is similar to Swinbank below; the second half of the paper discusses trade implications)

Crutchfield S. et al., 1997, An Economic Assessment of Food Safety Regulations: The New Approach to Meat and Poultry Inspection. Economic Research Service, United States Department of Agriculture, Agricultural Economic Report No. 755.
(an easy-to-follow example of a case study estimating the benefit-cost ratio of food safety regulation in the US meat industry)

Henson, S., 2003, The Economics of Food Safety in Developing Countries, Rome, FAO
(read only Sections 4 and 5, Markets for Food Safety and Food Safety Regulations, pp. 19-32)

Antle, J., 1999, Benefits and costs of food safety regulation, Food Policy 24, 6, 605-623.

Supplementary readings

Swinbank, A ., 1993, The economics of food safety, Food Policy 18 (2) 83-94 (not available through Library e-Journals but can be accessed in Periodicals).

Antle, J. 1995. 'The market for food safety', Chapter 3 in Choice and Efficiency in Food Safety Policy, American Enterprise Institute Studies in Agricultural Policy, Washington, DC (ARTS 363.11 N51 there is also a LEN copy).
(covers much the same ground as Swinbank and Mitchell )

Henson, S. and Caswell, J. 1999. Food safety regulation: an overview of contemporary issues. Food Policy 24, 6, 589-603. (only accessible from a TCD address)
(this issue of Food Policy has a number of other articles on the economics of food safety).

Valeeva, N., Meuwissen, M., and Huirne, R., 2004, Economics of food safety in chains: a review of general principles, Netherlands Journal of Agricultural Science 51-4, 369-390.
(focuses on methods for measuring costs and benefits)

USDA, 2003, Weighing incentives for food safety in meat and poultry, Amber Waves, April.
(Discusses the different mechanisms at work to improve food safety in the US meat industry)

New Approaches to Food Safety Economics, Wageningen University
(on line book with case studies of different types of food regulation measures )

Web resources

For developments in US food safety policies, check out the resources in the US Department of Agriculture Economics Research Service Briefing Rooms on the economics of foodborne disease and food safety policies.

The USDA ERS has a neat little foodborne illness cost calculator where you can get an idea of the way in which the costs of foodborne diseases are calculated.