1 Benefits begin to accrue five years after the HACCP rule is enacted, and extend over 20 years.

2Landefeld and Seskin VOSL estimates after averaging across gender and updating to 1995 dollars using BLS usual weekly earnings. 3 VOSL = value of a statistical life.

Source: Crutchfield et al., 1997. See http:¡lwww.ers.usda.govjbriefingjFoodSafetyPolicy!'features.htm for updated estimates.

1 Benefits begin to accrue five years after the HACCP rule is enacted, and extend over 20 years.

2Landefeld and Seskin VOSL estimates after averaging across gender and updating to 1995 dollars using BLS usual weekly earnings. 3 VOSL = value of a statistical life.

Source: Crutchfield et al., 1997. See http:¡lwww.ers.usda.govjbriefingjFoodSafetyPolicy!'features.htm for updated estimates.

Instances of acute adverse reactions are extremely rare. Levine's (1991) survey of the literature on pesticides from 1930 through the late 1980s turned up 42 instances of outbreaks of pesticide poisonings related to ingestion of contaminated food and water. Almost one-third involved cases in which peasants in poor countries facing starvation knowingly ate seed treated with pesticides and marked as not fit for consumption. The majority of the remaining cases also came from poor countries and involved inadvertent consumption of pesticides under the belief that they were flour or sugar, consumption of cooking oil stored in pesticide containers, and similar instances involving poor sanitation. Consumption of meat from animals fed illegally with treated seed accounted for several cases, whereas consumption of fish from polluted waters accounted for a single case (i.e., methyl mercury in Canada). The only recent case in the United States occurred in 1986 and involved the illegal application of the insecticide aldicarb to watermelons, despite the prohibition on its use on food crops. Recent cases of acute illness in other developed countries have similarly involved illegal uses, for example, the recent cases of meat raised on feed with excessive levels of dioxin in Belgium (see Buzby et al., 2001 for more on the dioxin incident) and soft drink cans contaminated with fungicide in western Europe. Cases of acute illness involving food additives have mainly been allergic reactions, such as the sweetener aspartame causing adverse reactions in those unable to digest the enzyme phenylalanine.

A number of studies have attempted to quantify the contributions of controllable substances to known long-term health effects, notably cancer (Doll and Peto, 1981; Henderson et al., 1991; Lutz and Schlatter, 1992; Ames et al., 1995). These studies combined information from animal bioassays with epidemiological information to estimate the numbers of annual cancer deaths attributable to various causes. The principal causes of cancer associated with diet are tobacco, fat, and, possibly, overnutrition. All food additives taken together were assigned a token amount of less than 1% of annual cancer deaths because the epidemiological evidence indicated no significant correlation between ingestion of these substances and elevated rates of any cancers for which laboratory studies and physiological analyses had suggested a possible causal connection.

The low incidence of illness and death related to chemicals in foods is testimony to the stringency of regulation of food additives by the FDA and pesticides by the U.S. Environmental Protection Agency (EPA). By law, the FDA is required to ascertain that food additives are safe before approving them for use. Food additives shown to cause cancer in animals cannot receive approval. The EPA is similarly required to set tolerances (maximum allowable limits) for pesticide residues on foods that ensure a reasonable certainty of no harm. Surveillance data collected by the FDA as part of its enforcement effort indicate that most domestic fruits, vegetables, cereals, meat, eggs, and dairy products sold in the United States have no detectable pesticide residues and that only about 1% have residues exceeding tolerances (Food and Drug Administration Pesticide Program, 1987-1998).


As noted above, the value of an incremental improvement in food safety is typically estimated as the product of two factors. The first factor is the change in risk, that is, the change in the probability of illness or death or, equivalently, of the incidence of illness or death in the population. The second factor is the average value of saving a life or avoiding illness. Estimation of both factors has been controversial.

Risk Assessment of Chemicals in Foods

The risk of illness or death from chemicals in foods cannot generally be estimated directly from human data. Past experience may be an insufficient guide to the risks of new chemicals. Regulation is prospective and seeks to avoid adverse consequences, so that human data may be simply unobtainable for new chemicals. Regulatory assessments of the risks associated with exposure to chemicals in foods thus tend to rely on animal studies to assess toxicity. These toxicity results are adjusted to account for physiological differences between humans and test animals and are then combined with assessments of exposure to yield an overall quantitative characterization of risk.

In general, the results of these procedures are not appropriately characterized as estimates of risk. EPA imposes a number of assumptions designed to produce "conservative" figures. Its underlying rationale is a desire to avoid type II error, that is, declaring a compound to be safe when it in fact poses a risk, possibly to an especially susceptible subpopulation. Thus, rather than using the average toxicity obtained from animal studies, EPA uses the upper limit of a 95% confidence interval of the toxicity of a substance to the most vulnerable test species. The highest physiologically defensible number is used to convert the dose from the test animal to a human equivalent. The highest possible figures are similarly used to estimate exposure.

These procedures have a number of undesirable consequences (Nichols and Zeckhauser, 1986; Lichtenberg, 1991):

• First, they tend to overestimate the benefits of regulation and understate the costs, both in total and at the margin. As a result, they indicate the desirability of levels of regulation that are actually excessively stringent.

• Second, they make it impossible to compare quantitative characterizations of risk across substances, making it impossible to determine whether substances are regulated under comparable degrees of stringency. Each quantitative characterization of risk can be characterized as an upper limit of a confidence interval, but the type of confidence interval varies in an unknown manner because of the arbitrary nature of the assumptions imposed.

• Third, they tend to overstate the net benefits of ex ante regulatory actions relative to surveillance, monitoring, and other ex post enforcement methods (Lichtenberg, 1991).

• Fourth, they tend to exaggerate risks from chemicals in foods and thus unjustifiably undermine confidence in the safety of the U.S. food supply. The EPA refers to its quantitative characterizations of risk as risk estimates, and they enter policy discussion as such.

Thus the EPA's risk assessment procedures suggest that chemicals in foods pose much greater risks than the data indicate. Moreover, research in cognitive psychology has shown that people consistently overestimate rare events like cancer from chemicals in food (see for example Fischoff et al., 1981). This bias in risk perception lends additional credibility to the EPA's exaggerated risk estimates. Overall, concern about chemicals in foods is much more prominent in food safety policy discussions than the incidence of food safety problems attributable to them would appear to warrant.

Concern over chemicals in food may have fallen somewhat over the past decade, however, after rising during the preceding decades. A number of surveys conducted between 1984 and 1990 in localities scattered across the U.S. indicated that most Americans had serious concerns about pesticide residues on foods (Sachs et al., 1987; Jolly et al., 1989; Food Marketing Institute, 1989; Porter/Novelli, 1990; Dunlap and Beus, 1992; Weaver et al., 1992). Sachs et al. (1987), comparing the results of their survey of Pennsylvania households with those of a survey done 20 years earlier, found much greater concern over pesticides in 1985 than in 1965. A national poll conducted in 1994, however, found that a minority of Americans (35-38%) believed pesticides were very dangerous for themselves or for the environment, roughly half the percentages reporting such concerns only a few years earlier (National Opinion Research Center, 1994). Food Marketing Institute surveys of public perceptions indicate that during the mid- to late 1980s, the majority of Americans considered chemicals in foods as the top concern related to food safety. By 1995, only about 14% reported chemicals as their top food safety concern (Buzby and Ready, 1996).

Valuing Avoidance of Chemicals in Foods

The value of avoiding illness or death from exposure to chemicals in foods may differ from the general value of avoiding illness or death. It is possible, for instance, that people have special fears about the types of illness or death resulting from chemical exposure. Several different types of evidence suggest that this is not the case, suggesting that values of life saving derived from the general literature are applicable to cases involving chemicals in foods. Overall, most consumers wanted assurance that their food was safe but were willing to pay little extra for small increments in safety beyond the level set by regulators.

One recent study tackled this question directly by comparing the implicit value of saving lives from risks posed by pesticide residues on foods and auto mobile accidents. It found no significant difference between them (Horowitz, 1994).

A number of other studies have examined consumers' willingness to pay for lower levels of pesticide residues on foods, in particular, complete elimination of all such residues (Ott, 1990; Misra et al., 1991; Weaver et al., 1992; Eom, 1994; Buzby et al, 1995; Buzby et al., 1998). All of these studies used CVM. Eom (1994) and Buzby et al. (1998) used a discrete choice format in which survey participants were asked which of two types of produce they would purchase at a given price differential. Eom (1994) found that willingness to pay was insensitive to the level of risk participants were told they were facing. Buzby et al. (1998) found no significant difference between respondents' willingness to pay for produce that met government standards for pesticide residues on foods and produce certified to be residue-free. The remainder of these studies asked participants to report the highest premium they would be willing to pay for produce certified to be residue-free. This latter format tends to generate excessively high reports of willingness to pay. Moreover, none of the surveys were constructed to replicate actual choice situations, that is, respondents knew that the questions were hypothetical and that there was no chance their answer would have direct financial consequences such as actually paying more. Hypothetical survey formats of this kind tend to generate excessively high reports of willingness to pay. Thus one would expect these studies to generate excessively high estimates of consumers' willingness to pay to eliminate pesticide residues on produce. Even so, few consumers reported being willing to pay more than 5% more for certified pesticide residue-free produce. Between 20 and 40% of respondents were willing to pay nothing extra, whereas an additional 25-60% were willing to pay no more than 5%) extra. In most studies, only about 10% of respondents reported being willing to pay 10% or more extra. Even fewer reported being willing to buy certified residue-free produce with lower cosmetic quality or more surface defects.

Baker and Crosbie (1993) obtained similar results using conjoint analysis on a small sample of produce shoppers at two San Jose, California supermarkets in 1992 to explore their relative preferences for price, cosmetic quality, and pesticide residues. Cluster analysis indicated that these shoppers could be divided into three subgroups. About 30% cared about price and quality but not pesticide residues. The majority (55%) cared about price and quality and whether the produce met government standards for residues. The remainder (about 15%) wanted stricter government regulation of pesticide use on the farm.

Other corroborating evidence comes from studies of demand for organic food. Some have argued that the growth of organic food sales is an indication of the public's willingness to pay to avoid pesticide residues on foods. Prices for organic produce average 25-35% higher than comparable conventional produce and have been observed to be as much as double or triple the prices for conventional produce in nearby stores (Hammitt, 1986; Morgan and Barbour, 1991; Thompson and Kidwell, 1998). These price premiums measure demand for reductions in chemicals in foods to the extent that demand for organic food is driven by concerns over chemicals in foods. It appears, however, that concerns about pesticides do not account for most of the motivation for buying organic foods. Most purchasers of organic foods believe that they are more nutritious and flavorful than conventionally grown foods (Hammitt, 1986; Jolly et al., 1989), whereas others purchase organic foods for worker safety and/or environmental concerns. Certification as pesticide residue-free did not influence demand for organic broccoli, carrots, or lettuce during the period 1985-1989 (Park and Lohr, 1996).

In sum, it appears that the great majority of the U.S. population wants assurance that produce is safe but has little or no demand for additional reductions in chemicals in foods. Thus the average willingness to pay for reductions in chemicals in foods should be treated as equal to willingness to pay for other reductions in illness or death. There does appear to be a small segment of the population willing to buy organic food at a high price premium. This segment values the overall manner in which produce is grown rather than the absence of chemicals per se. Absence of chemicals appears to be a relatively small part of the motivation for buying organic food. The price elasticity of demand for organic produce appears to be extremely low, suggesting that purchasers of organic produce do not consider conventional produce as much of a substitute for it (Thompson and Kidwell, 1998). The income elasticity of demand for organic food is quite high, suggesting that organic produce is a luxury good (Park and Lohr, 1996; Thompson and Kidwell, 1998). Thus there appears to be no reason to treat this segment of the population differently than the general population in estimating the benefits (avoided costs) of reductions in chemicals in foods.


Food additives and pesticide residues on foods are regulated under the Federal Food, Drug, and Cosmetic Act (FFDCA). In both cases, regulation is driven solely by health criteria. In general, food additives can be used legally only if the FDA has determined that they are safe. However, the FDA can approve the use of additives that can cause adverse health effects if it finds them to be safe at sufficiently small concentrations. In such cases it issues a tolerance specifying the maximum allowable concentration, which is generally 1/100 of the maximum concentration at which adverse health effects are observed. The FFDCA also specifically forbids the use of additives found to cause cancer in humans or animals.

Regulation of pesticide residues on foods is carried out by the EPA under the FFDCA as amended by the Food Quality Protection Act of 1996. This legislation directs the EPA to set tolerances for pesticide residues on foods at levels that create a reasonable certainty of no harm from aggregate exposure, including all dietary exposures as well as other exposures for which reliable information exists. It also directs the EPA to make a special determination of safety for infants and children and requires the use of an extra 10-fold margin of safety for all substances with threshold effects that pose some risk to infants and children. The EPA must take into account the vulnerability of special subpopulations (including infants and children) in estimating exposure and health effects and must use safety factors recognized by qualified experts as appropriate. Data on the actual use of the pesticide on crops and actual residue levels can be used for this purpose only if the EPA determines that the data are reliable and do not underestimate exposure for significant subpopulations. Tolerances last for five years, at which time they must be reviewed.

The legislation does allow limited use of economic criteria for determining appropriate levels of pesticide residues on food by allowing tolerances to be set at levels necessary to "avoid significant disruption in domestic production of an adequate, wholesome, and economical food supply." Even so, economic considerations can be used to assess the appropriateness of regulatory decisions ranging from approval of a tolerance to the adequacy of surveillance and enforcement programs. Moreover, economic considerations frequently enter regulatory decision making implicitly even in cases where statutes give them no explicit role.

Buzby et al. (1995) provide an example of economic analysis of changes in food safety in a case involving pesticide regulation, specifically, postharvest treatment of fresh market grapefruit with sodium orthophenylphenate (SOPP). Grower surveys were used to identify likely alternative postharvest treatment methods for grapefruit, to estimate the packinghouse level changes in treatment cost, and to estimate changes in spoilage losses. These estimated changes in treatment cost and spoilage losses were then used to estimate a shift in the supply of grapefruit provided to the fresh market (for a discussion of methodology see Lichtenberg et al., 1988). A model of grapefruit supply and demand was used to estimate changes in grapefruit consumption and price and thus changes in grapefruit consumers' and producers' incomes. Changes in price serve as a mechanism for shifting a portion of the costs of the ban from producers onto consumers. Consumers will respond to price changes in part by substituting consumption of other commodities for grapefruit. The overall cost of the regulation as measured by changes in consumer and producer incomes will generally be less than the additional cost of treating the preregulation grapefruit crop because of these substitution possibilities. CVM was used to estimate consumers' willingness to pay for the reductions in risk induced by the regulation, as estimated by the EPA. The incremental benefit of the regulation derived from the willingness to pay estimates was then compared to the losses in consumer and producer incomes, which together comprised the incremental cost of the regulation. The net impact was positive, indicating that banning SOPP would increase societal net income. This result should be taken as illustrative: The net benefits were likely lower than those estimated because the EPA's risk assessment methods overstated the reductions in risk effected by the proposed regulation.

0 0

Post a comment