A field experiment on consumer willingness to accept milk that may have come from cloned cows

Food Policy 74 (2018) 1–8

Contents lists available at ScienceDirect

Food Policy journal homepage: www.elsevier.com/locate/foodpol

A field experiment on consumer willingness to accept milk that may have come from cloned cows☆ Kofi Britwuma, John C. Bernardb, a b

MARK

⁎

School of Economics, University of Maine, Winslow Hall, Orono, ME 04469, United States Department of Applied Economics and Statistics, University of Delaware, 213 Townsend Hall, Newark, DE 19716, United States

A R T I C L E I N F O

A B S T R A C T

Keywords: Animal cloning Willingness-to-accept Milk Field experiment

FDA policy states that meat and milk from cloned animals is as safe as those from conventional animals and thus such products do not need to be labeled. Therefore if the voluntary moratorium on these foods were to end, consumers would be uncertain which items were products of cloned animals. This study examined consumer willingness to accept (WTA) milk that may or may not have come from cloned cows to judge the market impact of this policy. Non-hypothetical field experiments at four diverse locations were conducted to determine the amount consumers would require to switch from a cup of conventional milk to one with potentially cloned cow milk. Nearly a third of participants submitted the maximum request of $5, indicating a complete unwillingness to switch, while another quarter requested no compensation. Tobit model results showed opinions, attitudes, and shopping habits highly influenced consumers’ WTA. Accompanying survey results found a neutral opinion and low knowledge of cloning, but strong support for labeling despite being informed of the FDA conclusions. Given these findings, in the absence of policy changes voluntary labeling of milk as not being from cloned cows could be a likely future outcome.

1. Introduction

Agriculture (Knight, 2008). If or when they do, however, the above FDA ruling means consumers will be unable to tell if the milk or meat they are purchasing was the product of a cloned animal. There could be ramifications for these markets and overall consumer welfare depending on the public’s reaction. A key consideration is how willing consumers will be to accept products from cloned animals. Opponents of animal cloning mention issues from ethics to environmental impacts and concerns regarding the health of cloned animals and the safety of food products from clones. Nonis et al. (2010) reported that consumers expressed moral and ethical misgivings towards the adoption of cloning for food purposes. Even when consumers were provided credible information from the FDA about animal cloning, attitudes towards cloning had a strong influence on purchase intentions. Other studies have examined consumers’ attitudes towards and willingness to pay (WTP) for cloned animal products. Brooks and Lusk (2011) surveyed US consumers and reported that approximately 31% were willing to consume meat and milk products from cloned animals whereas about 43% were unwilling. Butler et al. (2008) found only slightly lower purchasing interest for milk from cloned cows when offered at a discounted price. However, when consumers were initially unaware the milk was from cloned cows and not given a price discount,

Advances in agricultural biotechnology have made animal cloning another viable reproductive technology for farmers and ranchers. As the science has moved forward since the first cloned sheep, Dolly, in 1996, animal cloning has created the potential to transform the beef and dairy industries by offering farmers the option to clone their best breeding stock or milk producing cows. Many countries besides the US, such as China, Germany, Australia, and Japan participate in animal cloning, with cattle, pigs and goats the most popular cloned species. The potential for meat and milk from these animals to enter the US food supply passed one hurdle when, in 2008, the Food and Drug Administration (FDA) concluded that meat and milk from clones of cattle, swine and goats, and the offspring of clones from any species traditionally consumed as food, are as safe to eat as food from conventionally bred animals (FDA, 2008). In conjunction with this, the FDA announced that there would be no mandatory labeling requirement for foods from cloned animals. Such products do not currently exist in the marketplace as companies with cloned animals in their stock continue to follow the voluntary moratorium, a position supported by the United States Department of

☆ ⁎

The authors are grateful to the Applied Economics and Statistics Department of the University of Delaware for providing funding for the study. Corresponding author. E-mail addresses: [email protected] (K. Britwum), [email protected] (J.C. Bernard).

https://doi.org/10.1016/j.foodpol.2017.10.006 Received 10 April 2016; Received in revised form 16 October 2017; Accepted 30 October 2017 0306-9192/ © 2017 Elsevier Ltd. All rights reserved.

Food Policy 74 (2018) 1–8

K. Britwum, J.C. Bernard

omission. One could argue that there are always details omitted in food experiments. For instance, many omit brand names to avoid such biases. Ellison et al. (2016) showed that for organic foods, even the purchasing venue, which is nearly always omitted, influences WTP results. Within the taxonomy, deception by omission is in the least severe category and both pairs of survey respondents voiced little concern over the practice. Thus it was believed the benefits of the following design in gaining non-hypothetical WTA values, while matching future market conditions (in which this omission will be allowed by companies under FDA policy), outweighed possible concerns by some. The view of the researchers though was that this study did not use deception as, following Hey (1998), subjects were never told the “wrong things.” Indeed, the design carefully avoids doing this. Following the design outline and guidelines above, field experiments were conducted at four different locations in Delaware in October 2012, generating a total sample of 148 participants. The locations and days were selected to help capture a wide variety of backgrounds and opinions to aid in modeling and make the results as generalizable as possible within the typical limits of field experiments. The locations were a public park in New Castle County, a natural foods store (which also had a farmer’s market), an urban farmer’s market in Wilmington, and the University of Delaware campus in Newark.1 The first two were visited on weekends and the latter two during the week. The park was chosen as the best representation of the general population as a popular place for many in the surrounding community and was indeed the most demographically diverse group in the sample. The natural food store in the college town setting gave access to those perhaps more knowledgeable, and potentially more opposed, to animal cloning. An urban population was captured with the sample in Wilmington, the largest city in the state. Most participants were there to have lunch on their breaks from the surrounding office buildings since food service was a substantial part of the urban farmer’s market. The location of the market near a busy city sidewalk allowed other participants to be pulled in from the street by signs advertising a food study. The campus location consisted of students and gave a view of the acceptance of such products from the next generation of consumers. As most students were from the agricultural college, it was expected their knowledge would also be relatively high. Approximately four hours was spent at each location, with each participant spending about five minutes on the experiment. Typically, a session began with setting up a tent, a table, and signage announcing a University of Delaware research project. Near the table was a cooler with ice containing two gallon-size containers of 2% milk. The brand name for one of the gallons was removed to represent the “may or may not” milk that could potentially have originated from a cloned cow. The second gallon of milk that still had the brand name represented conventional milk.2 Potential participants were approached and asked if they would be willing to take part in a short study on milk where they could earn between $2 and $7. The only requirements noted were that they drink milk and were over 18 years old. For those agreeing, they were read the following statement regarding the FDA’s ruling (FDA, 2008) and potential for milk from cloned cows:

their purchase interest dropped substantially when they were informed. They underscored the importance of consumer education on animal cloning prior to introducing milk from cloned cows to avoid a backlash, noting that lower retail prices were not sufficient. Jones et al. (2010) found about 59% of their survey participants were willing to pay more for meat with a label indicating it was not a product of a cloned animal, while around 40% were not. A disadvantage of the previous studies has been their hypothetical nature. This study instead utilizes a non-hypothetical field experiment, with subjects faced with an actual consumption decision. By requiring consumption of a real product, subjects were encouraged to be more thoughtful about the compensations they requested to mitigate potential bias (Fox, 1995; Lusk et al., 2004). A second contribution of the design was in having the selected product, milk, be described as possibly being from a cloned cow to mirror the future scenario should products from cloned animals be introduced without mandatory labeling. Lastly, rather than the typical WTP set-up, a willingness to accept (WTA) design was used to more directly gauge aversion to the uncertain milk product. This technique has been noted to resemble market conditions where consumers make the choice to accept products if they are compensated with lower prices (Lusk et al., 2004; Moon et al., 2006). Using the WTA approach, the primary goal of this study was to determine the minimum compensation consumers would need to be paid to exchange a cup of milk from conventional cows for a cup of milk that may or may not have originated from cloned cows. Part of this was to conduct the study at four diverse locations to best capture potential differences in attitudes across various segments of the population. Secondary goals included examining consumers’ opinions and knowledge of animal cloning, their views on labeling, and whether they believe the technology should be used. 2. Experimental design A prominent issue in the experimental design was a desire to avoid deception. Deception, while allowed in experiments in fields such as psychology, has traditionally been unacceptable in economics. However, there has developed a good deal of debate over the precise meaning of deception and what designs within that broad context might be allowable. Rousu et al. (2015) and Colson et al. (2015) proposed a deception taxonomy and looked at both researchers’ and student subjects’ opinions of various forms. Clearly not acceptable was selling a product that was not as the experimenters defined, meaning that claiming conventional milk actually came from a cloned cow would be viewed as a severe form of deception. Obtaining milk from cloned cows would alleviate these concerns and indeed we contacted businesses with cloned cows and found that obtaining milk from them, while difficult due to transportation issues, would be feasible. However, a second important issue in the design was to match the actual market conditions if the voluntary moratorium on cloned products were to end. Given that labeling would not be required under current policy, it could not be expected that any milk container would voluntarily be labeled as including cloned cow milk. Thus, the value of someone’s WTA milk certain to have come from a cloned cow would not be especially useful from a policy perspective. Rather, each container would be accurately described as “may or may not” contain milk from a cloned cow in the absence of the moratorium and mandatory labeling. Using this market-based design additionally increased the potential contribution of this study as the “may or may not” scenario has not been explored elsewhere. The issue then is if use of the description “may or may not” in the situation where the researchers know the answer remains a form of deception. While this could be considered deception by omission, an open question could be if any food study is completely devoid of such deception. As noted by Colson et al. (2015), all papers examining the influence of information on consumer demand use deception by

“The FDA has decided that meat and milk from cloned cows is as safe to eat and drink as meat and milk from conventionally bred cows. Due to this, if milk from cloned cows enters the market, it will not need to be labeled and you will not be able to tell if you are drinking it or not. While milk from cloned cows is not currently in stores, it does exist and we have previously contacted some farms and companies that have cloned cows.”

1 We considered having the experiments in grocery stores, but no grocery store we approached was willing to grant us permission, wary that their animal products might be associated with cloning. 2 Both milk containers were the same store brand and were purchased right before each session to be sure they were fresh.

2

Food Policy 74 (2018) 1–8

K. Britwum, J.C. Bernard

Narrating the FDA’s position to participants was done to give everyone the same base of understanding of the potential market situation. Given that trust in regulating institutions like the FDA is varied (Huffman et al., 2004; Brooks and Lusk, 2011), different reactions about the FDA as a trustworthy source of information were expected. After reading this part, participants were shown the two milk containers in the cooler. To determine participants’ WTA milk that may have come from a cloned cow, the incentive compatible BDM mechanism was employed (Becker et al., 1964). The BDM was selected over other auction formats since it is a one-person mechanism, making it uniquely adaptable to field settings (Lusk and Fox, 2003; Wertenbroch and Skiera, 2002). Specifically, participants were endowed with a cup of conventional milk and asked their minimum compensation necessary to switch to the potentially cloned cow milk. This design was similar to Lusk et al. (2004, 2005), who gathered WTA a GM food product after first endowing participants with a non-GM version. To explain the experimental goal and procedure, participants were told the following:

Table 1 Descriptive statistics for subjects’ WTA the uncertain milk. Survey location

Obs

Min $

Min%

Max $

Max%

Mean $

Std.dev $

University campus Natural foods store Urban farmer's market Public park Overall

13

0

15.38%

5

7.69%

1.44

1.35

46

0

23.91%

5

52.17%

3.18

2.14

47

0

12.77%

5

19.15%

2.86

1.59

42 148

0 0

38.10% 23.65%

5 5

28.57% 31.08%

2.19 2.65

2.10 1.97

respondents that requested the minimum possible compensation ($0) or the maximum compensation ($5) as well as the mean requested amount and standard deviation. Differences between the locations were quickly apparent. While the overall mean WTA was $2.65, it was greatest at $3.18 for respondents at the natural foods store and lowest among university participants at $1.44.6 This appears to follow Faass and Lahr (2007) who noted that younger consumers tend to be more accepting of food products from biotechnology. More revealing were the percentages of subjects at each of the extremes. Nearly a quarter of the overall sample did not request any compensation and willingly switched to the cup of milk that may have come from a cloned cow. The largest group here, at 38%, was from the general public in the park, with, somewhat surprisingly, the second biggest group (24%) at the natural food store. The arguably more meaningful finding was in the large numbers of people that requested $5, an offer that guaranteed them they would not have to consume the possibly cloned cow milk. This group accounted for nearly a third of the overall sample and was over half of the natural food store subjects. At over 28%, the park sample also seemed very hesitant to try the uncertain milk. Only for the small student sample were almost all willing to at least allow a chance that they could receive the possibly cloned cow milk. The large groupings at the extreme ends of the WTA scale led to a clear bimodal distribution when displayed graphically. Differences do appear however when examining the distributions by location, as displayed in Fig. 1. The park sample was perhaps closest to matching the overall bimodal pattern, with a smaller grouping near the center of the scale but few observations elsewhere beyond the extremes. The natural food shoppers were similar in form, but with the obvious large number already noted at $5. A higher percentage of on-campus subjects relative to those two locations requested at least some compensation, although their distribution was heavily skewed to the right. No individual from this location that was willing to have the chance to switch milks requested over $2.50. While the urban farmer’s market subjects were not especially unwilling to switch milk, they tended to request much higher levels of compensation than the others to do so. The nearly 30% requesting between $3 and $3.49 was the notable high level in their distribution. One finding in common across the locations was the lack of any offers in the $4.50 to $4.99 range. While further analysis would be required, this could suggest that $5 was an appropriate maximum as only those making offers to ensure they would not need to consume the possibly cloned cow milk requested over $4.49. There was also evidence across locations that if people want compensation, very small amounts would not be enough. For two of the four locations, no one requested compensation between $0.01 and $0.49. Most notable was the natural foods store group where the lowest of those asking for

“We are trying to determine your value for conventional milk compared to milk that may or may not have come from a cloned cow. To do this, we are offering you a cup of conventional milk, but would like to know how much money you would require to instead be given a cup of milk that may or may not have come from a cloned cow. Your amount needs to be between 0 and $5.” The importance of participants stating their true value for exchanging their cup of milk was emphasized as the best approach in the BDM and the workings of the mechanism were carefully detailed. It was explained that a compensation of $0 indicated indifference between the conventional milk and the potentially cloned cow version, whereas a compensation of $5 indicated maximum aversion and meant there would be no chance they would have to consume the potentially cloned cow milk. Participants were told the type of milk they would ultimately consume depended on a random number draw. The draws were accomplished by having the participant select from a series of envelopes on the table, each of which contained a small piece of paper with prearranged values between $0 and $5 in $0.25 increments. A participant’s stated offer was compared with the random number from their selected envelope. If the participant’s offer was less than the random amount, they were paid an amount equal to the number drawn and offered a cup of the potentially cloned cow milk to drink. Alternatively, if a participant’s offer was equal or greater than the number drawn, they were offered a cup of conventional milk to drink but no compensation was paid. This set-up ensured that participants who requested a compensation of $5 did not have to consume a cup of the potentially cloned cow milk.3 Either way, participants were given a cookie with the milk and $2 for participating alongside any compensation earned. While consuming their milk and cookie, each participant filled out a short questionnaire on their knowledge and opinions of animal cloning, and their demographics. 3. Data and descriptive statistics Table 1 displays descriptive statistics for the 1484 subjects’ WTA offers by location and overall.5 These include the percent of 3 This is a key reason behind using the WTA framework where, similar to Lusk et al. (2004), we did not want to force someone to consume a product they strongly wished to avoid. 4 A power test confirms that the sample size of 148 is adequate. Specifically, with a population mean compensation of $3 and a standard deviation of $2, the minimum sample required to detect a mean compensation of $2.50 with a 0.8 power given a 0.05 Type 1 error rate is 128. 5 While the University sample is too small to make generalized claims about students, results from this location are still considered separately to aid in discussion and suggest future study.

6 Care must be taken in drawing conclusions from the means since the $5 maximum was often entered. While it had been expected that $5 would be a high enough limit for trying a small cup of milk, it appears many would need a much greater compensation or may be unwilling for any amount.

3

Food Policy 74 (2018) 1–8

K. Britwum, J.C. Bernard

Fig. 1. Subjects’ WTA distribution by location.

4. Empirical specification

compensation requested $1. Overall, compensation requests mostly fell between $1 and $3.49. Table 2 shows the descriptive statistics for all variables in the dataset which included details on demographics, food purchase decisions, knowledge and opinion of animal cloning, and whether respondents wanted milk from cloned cows allowed and labeled. The demographics compared very closely to Delaware census data for 2010, especially race and gender, suggesting the sampling procedure performed well. Approximately 49% of respondents were male and about 70% were white. The majority of respondents (73%) had at least a college degree, which was expected given the area. In terms of shopping habits, most indicated being the primary shopper in their households and often read food labels. Food purchasing questions involved asking participants if they made their milk buying decisions on health, environment, ethical, animal welfare, or cost concerns. Cost and health were both important factors, with ethics noted the least. Opinions and attitudes showed clear support for labeling, but many components showed substantial differences based on location that will be examined in the results section.

4.1. Tobit model To investigate consumer characteristics that influenced their WTA milk that may have come from cloned cows, a two-limit tobit model with upper and lower censoring was used. For this, it was assumed there exists a latent variable yi∗ representing subject i’s offer to accept milk that may have come from a cloned cow. Censoring exists in that yi∗ is unobserved for values greater than or equal to 5 and values less than or equal to 0. Thus, these latent variables are related to the observed offers, yi by:

0 if yi∗ ⩽ 0 ⎧ ⎪ ∗ yi = yi = xβ + εi if 0 < yi∗ < 5 ⎨ ⎪5 if yi∗ ⩾ 5 ⎩

(1)

where x is a vector of independent variables, β is a vector of estimable parameters and εi is a normally and independently distributed error

Table 2 Definition of variables and descriptive statistics. Variable Shopping habits Drink conventional Gallons purchased Read labels Primary shopper Health Environment Ethics Animal welfare Cost Opinions and attitudes Knowledge cloning Opinion cloning Allow cloning Label from clones Demographics Male White College plus Income Age Children in home Liberal views

Description

Mean

Std Dev

Frequency of consuming conventional milk, 1 = never to 4 = frequently Gallons of milk purchased in a week, 1 = half or less to 4 = more than two Frequency of reading food labels, 1 = never to 4 = always 1 if subject does most of household grocery shopping; 0 otherwise Milk purchase decisions based on health, 1 = never to 4 = always Milk purchase decisions based on environment, 1 = never to 4 = always Milk purchase decisions based on ethics, 1 = never to 4 = always Milk purchase decisions based on animal welfare, 1 = never to 4 = always Milk purchase decisions based on cost, 1 = never to 4 = always

2.94 1.97 3.16 0.78 3.01 2.75 2.37 2.60 3.04

1.10 1.04 0.83 0.41 0.94 0.90 1.12 1.02 0.91

Knowledge of cloning, scale from 1 = have none to 4 = excellent Opinion of cloning, 1 = negative 0 = have none/neutral/positive 1 if subject wants milk from cloned cows allowed; 0 otherwise 1 if subject wants milk from cloned cows labeled; 0 otherwise

2.19 0.36 0.66 0.88

0.91 0.62 0.48 0.33

1 if subject is male; 0 otherwise 1 if subject’s ethnicity is white; 0 if nonwhite 1 if subject had some college education or more; 0 otherwise Household income, in thousands Age, in years 1 if subject lives with children less than 18 years; 0 otherwise Personal disposition, 1 if liberal, 0 if moderate or conservative

0.49 0.70 0.73 66.90 41.88 0.36 0.39

0.50 0.46 0.45 55.19 15.78 0.48 0.49

4

Food Policy 74 (2018) 1–8

K. Britwum, J.C. Bernard

term with mean of 0 and constant variance. It was likely, however, that the model would suffer from heteroscedasticity as is typically the case with consumer studies of this type. Following Bernard and Bernard (2009), both the mean and variance portions of the model were estimated jointly to account for any such issues. The variance portion allows the specification of variables related to any heteroscedasticity of the residuals and is expressed as σ2(exp(ɀiγ)), where ɀi represents a second vector of relevant independent variables, γ is a second vector of coefficients, and σ2 is the variance when ɀiγ is zero (SAS, 2014). By modeling the heteroscedasticity, it is removed as a concern from the results of the mean portion. In forming the variance portion, it was assumed that variation in the error variance could be present in any of the variables considered in the mean portion, and as such all variables in the model were checked. Those found significant at the 5% level or better were included in the final model to adjust for heteroscedasticity. The joint modeling had the additional benefit that the parameter estimates in the variance portion could be interpreted in terms of each variable’s specific influence on the variance, creating additional insights. The final mean portion of the econometric model consisted of the variables in Table 2 as well as location dummy variables (with Public Park serving as the reference). In considering variations of the model to determine robustness of the findings, variables related to shopping habits, subjects’ opinions and attitudes as well as their demographics were considered essential in the model and omitting them would have led to inconsistent estimates (Lu and White, 2014). A second model without the location variables was run to serve as a robustness check for the original model.7 Both models were estimated using maximum likelihood with the QLIM procedure in SAS. Expectations a priori of all regressors were hypothesized based on intuition and past studies. Of the shopping habits variables, it was expected that frequent consumers of conventional milk, food label readers and those that based their purchasing decisions on various issues would request greater amounts of compensation, while those concerned with cost would request less. The opinion and attitude variables were anticipated to be especially meaningful with the hypothesized sign intuitively obvious for three of the four variables in this category; participants with a positive opinion of animal cloning, those who wanted milk from cloned cows allowed, and those who wanted them labeled were all expected to have a greater acceptance of milk from cloned animals. Subjects knowledgeable about animal cloning technology were also expected to be more accepting of products from cloned animals. Hypotheses were uncertain for a number of the demographic variables, although having children in the home was expected to increase compensation requested while liberal views would lower it. Higher income consumers were also expected to have a lower WTA based on studies such as Puduri et al. (2005) showing that this group had a higher approval of biotechnology. Lastly, for the location variables, those at the urban farmer’s market and natural foods store were hypothesized to request significantly more compensation than those in the public park.

Table 3 Knowledge, opinion and attitudes on animal cloning. Location Question

Category

Knowledge of animal cloning Have none Fair Good Excellent Opinion of animal cloning Have none Negative Neutral Positive Should milk from cloned allowed? Yes No Should milk from cloned labeled? Yes No

Public park

University

Natural foods store

Urban farmer’s market

Overall

30.95%

0.00%

19.57%

23.40%

22.30%

38.10% 23.81% 7.14%

38.46% 61.54% 0.00%

50.00% 15.22% 15.22%

53.19% 12.77% 10.64%

46.62% 20.95% 10.14%

7.14%

0.00%

6.67%

17.02%

8.84%

30.95% 54.76% 7.14% cows be

15.38% 76.92% 7.69%

57.78% 28.89% 6.67%

25.53% 51.06% 6.38%

36.05% 48.30% 6.80%

71.43% 28.57% cows be

92.31% 7.69%

48.89% 51.11%

71.11% 28.89%

66.21% 33.79%

85.37% 14.63%

76.92% 23.08%

93.48% 6.52%

84.78% 15.22%

86.99% 13.01%

have excellent knowledge. Natural foods store consumers were the most likely to claim excellent knowledge, followed by the urban farmer’s market group. The majority of subjects in the park reported having a fair knowledge of animal cloning. Overall, with only about 10% of subjects reporting excellent knowledge of the concept, consumer education should be considered an important component of any move to market meat or milk from cloned animals.8 In terms of opinion, the major share of respondents, about 48%, expressed a neutral opinion of cloning. A further 36% of respondents held a negative opinion, with only about 7% having a positive opinion. Those with no opinion measured around 9%, suggesting room for many people to be convinced one way or the other on cloning. No one in the University group fell in this category, likely a result of all subjects in the group citing some knowledge of the technology. This knowledge though translated primarily into a neutral opinion on cloning. Those most negative, not surprisingly, were the natural foods store shoppers with well over half in this category. Despite this, all groups had similar percentages noting positive opinions. At barely 7% however, it could take considerable effort by cloning proponents to make their technology positively viewed. A great deal could depend on how companies introduce foods from cloned animals or their offspring into the food system; lessons from the early introduction of GM food products and the apparent consumer mistrust can inform producers of cloned animals. As noted by Bratspies (2007), issues relating to consumer trust in regulatory institutions and the food industry are part of the reason for consumer concerns and low confidence in GM food products, leaving perhaps a key area of improvement for policy makers. Study participants were notably open to having milk from cloned cows as an option in the food supply, with two thirds agreeing that milk from cloned cows should be allowed. The natural foods store subjects were the most divided on this, with slightly over half disagreeing with the statement, corresponding to the large number with negative opinions just noted. The park and urban farmer’s market subjects matched

5. Results and discussion Prior to examining the model results, it is worth considering subjects’ knowledge and opinion of, and attitudes towards, animal cloning. Table 3 contains the results of the four questions on the survey designed to capture these aspects, by location and overall. Knowledge of animal cloning was fairly low, with about 22% of subjects indicating they had none while the largest group, at over 46%, indicated having only a fair knowledge. The student population was the only one where all subjects expressed at least fair knowledge of cloning, although none claimed to 7 Other specifications such as one-limit tobit models yielded similar results. Due to the high degree of censoring, ordinary least squares regression was not appropriate for the data and unlikely to yield similar results (Long, 1997).

8 While no testing questions were asked to assess subjects’ actual knowledge, Crespi et al. (2015) found self-reported cloning knowledge to match well with results of followup questions.

5

Food Policy 74 (2018) 1–8

K. Britwum, J.C. Bernard

conventional milk, the less willing they were to accept the potentially cloned cow milk. While purchasing frequency had no significant influence, those who read labels more frequently requested an additional $1.17 in compensation for each level up the scale to try the uncertain milk. These types of consumers likely care more about what is in their food and may be more hesitant to purchase products from cloned animals. In terms of factors affecting milk purchasing decisions, there was a contrast between those with environmental concerns and those with animal welfare concerns. The former requested very high rates of compensation to consider the uncertain milk ($2.54 for each level on the scale) while the latter were more willing to accept milk from cloned cows the higher their animal welfare concerns. A potential hypothesis for this is that consumers who rate animal welfare issues highly may not believe cloned cows are treated any differently than conventional cows. In the model without the location variables, subjects whose food shopping habits were influenced by ethical concerns requested an additional $0.45 for each level up the scale to consume the potentially cloned cow milk. Very likely, these would be subjects who may harbor ethical misgivings towards the cloning technology. Consumers who frequently considered cost in their food choices were more accepting of milk from cloned cows, perhaps seeing the potential for lower prices with more milk available in the market. Primary shoppers also requested less compensation. All four of the opinion and attitude variables were significant at better than the 1% level, although the Label from clones variable was not significant in the second model without location variables. In a positive finding for those planning on introducing products from cloned animals, the more knowledgeable a person viewed themselves as being about cloning, the more accepting they were of the uncertain milk. Education on animal cloning therefore may be enough to reduce the apparent misapprehension felt by many. Similar findings have been seen with different aspects of plant and animal biotechnology where consumers stating greater familiarity viewed the technology more positively (Faass and Lahr, 2007; Govindasamy et al., 2008; Jones et al., 2010). However, Aizaki et al. (2011) found Japanese consumers did not change their attitudes on cloning when given scientific information. Regardless of knowledge, subjects’ opinion of animal cloning exerted a great influence on their WTA. From the two models, subjects who held a negative opinion about cloning requested anywhere from $3.81 to $4 additional compensation relative to those holding positive or neutral opinions. Whether extra knowledge could overcome the negative views of some would be an avenue for further investigation. As expected, consumers who wanted milk from cloned cows allowed in the market asked for less compensation. However, there was a large difference with consumers between allowing a food product on the market from a cloned animal and allowing one without a label. One of the more controversial issues in food produced from genetic modification has been that of labeling, where many consumers’ preferences for labels have been at odds with the original FDA policy mandating labels only in certain situations. The parallels here may be that some segments of consumers may not necessarily be convinced by the FDA’s conclusion of no significant differences between conventional versions and cloned ones. This might sync with previous research findings (Lusk, 2008; Huffman et al., 2003) that suggest that consumers who strongly clamor for labels on non-conventional food products might avoid uncertain milk in the future. Turning to demographics, subjects who lived with children under 18 years were less accepting of the uncertain milk. Consumers may not necessarily want their children to be the first to be trying food products based on new technologies. Some participants had children with them (and curious about the milk and cookie), perhaps making this factor even more cogent. The remaining significant demographics, Male, College plus, Income, Liberal views, and Whites in the case of the second model all indicated lower WTA offers. The higher acceptance at higher

Table 4 Tobit regression results of WTA the uncertain milk.

Parameter Means Intercept Shopping habits Drink conventional Gallons purchased Read labels Primary shopper Health Environment Ethics Animal welfare Cost Opinions and attitudes Knowledge cloning Opinion cloning Allow cloning Label from clones Demographics Male White College plus Income Age Children in home Liberal views Location Urban farmer’s market Natural foods store University campus Variance Gallons purchased Read labels Primary shopper Children in home Liberal views Animal welfare

With location variables

Without location variables

Estimate

p-Value

Estimate

p-Value

−0.4274

0.8222

4.8282

0.0210

0.4458 −0.2526 1.1742 −2.7260 0.1670 2.5371 −0.2209 −1.2702 −0.5045

< 0.0001 0.1985 < 0.0001 < 0.0001 0.3887 < 0.0001 0.3168 < 0.0001 < 0.0001

0.2692 0.3043 0.6274 −1.8390 −0.2265 2.0477 0.4582 −1.6766 −0.4345

0.0668 0.1101 0.0165 < 0.0001 0.3930 < 0.0001 0.0214 0.0014 0.0180

−0.7582 3.8116 −0.4525 1.3330

0.0046 < 0.0001 0.0021 0.0002

−1.0959 4.0081 −0.4165 0.4173

0.0115 < 0.0001 0.0422 0.3846

−1.4535 −0.2306 −1.1568 −0.0190 0.0074 1.1170 −1.3138

< 0.0001 0.2236 < 0.0001 < 0.0001 0.4133 0.0013 0.0184

−1.1127 −0.8169 −1.3026 −0.0124 −0.0143 0.6336 −0.4437

0.0001 0.0119 < 0.0001 0.1323 0.3786 0.1248 0.7014

0.8965 2.0874 −0.9480

0.0023 0.0039 0.0580

3.1771 −2.8959 5.5740 −7.6617 5.6215 −1.4440

< 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 0.0008

2.9609 −1.2354 4.5972 −6.5707 5.6468 −2.5749

0.0378 0.1211 0.0354 0.0353 < 0.0001 0.0080

very closely, with over 70% in favor. Overall there seemed to be little interest in outright banning the technology from the marketplace, similar to the findings of Lusk (2012). The real issue for subjects was being able to tell if the milk they were buying was the product of a cloned animal. For the subsequent question on whether milk from cloned cows should be labeled, as noted above, the vast majority answered in the affirmative. The park and urban farmer’s market again yielded similar results, with university subjects slightly less interested in an added label and the natural foods store participants substantially more so. It can be inferred that many consumers may be unhappy with the FDA decision to not require labeling, although they may just be indicating a preference for having the information. Tobit model results presented next examine the characteristics of consumers and the amount of compensation they may require to accept potentially cloned cow milk.

5.1. Tobit results Results of the tobit regression analysis, with and without location variables, appear in Table 4. While the QLIM procedure in SAS does not provide an R2 or a pseudo-R2, the former was calculated as 0.3505 in the first model (with location) and 0.3204 in the second model, indicating reasonable goodness of fit for both. Beginning with the shopping habits’ variables in the means portion, six of the nine included were found to be significant in the first model, all at better than the 1% level. This was nearly the case with the second model as well, which had the Ethics variable additionally emerging significant at the 5% level. As expected, the more frequently a person consumed 6

Food Policy 74 (2018) 1–8

K. Britwum, J.C. Bernard

education fit again in suggesting educational materials may be a useful strategy for those wanting to see products from cloned animals succeed in the marketplace. Lastly, all location variables included in the model had significantly different WTA than the public park reference location. Compared with the park, the urban farmer’s market respondents requested $0.89 more in compensation while those on campus were different in the opposite direction, requesting $0.95 less. However, the more drastic difference was with the natural foods store respondents. These participants requested $2.09 more compensation than those in the park. This group would clearly be the most difficult to convince to even try a food that may have come from a cloned animal. These results demonstrate the relative differences in subjects’ compensation request that can be observed by conducting experiments in different locations aimed at capturing diverse populations within communities. For the variance portion, also included in Table 4, the disturbance variance in compensations was a function of several shopping habits and opinion variables in the first model, and three of these variables had a positive effect on variance. Participants who purchased higher volumes of milk weekly, primary grocery shoppers, and those with liberal views exhibited a wider variability in the compensations they requested to consume the possibly cloned cow milk. Thus, while the last two had significantly lower compensation requests, there was still significant variability within these groups on the whole. Three other variables exerted a negative effect on the variance. Subjects that lived with children less than 18, often read food labels and consider animal welfare in food decisions had a narrower variability in the compensations they requested to consume the possibly cloned cow milk. It can be inferred therefore that consumers within these groups may have a more homogeneous position on animal cloning.

Costanigro and Lusk (2014) tested the signaling concerns of mandatory labels (GM) while comparing them with the influence of voluntary labels (non-GM). While in the majority of their tests mandatory labels did not lower safety ratings for the products, they did significantly increase the WTP to avoid GM. This increase was also significantly greater than that for the voluntary label they tested. It is likely that a policy of mandatory labeling for products of animal clones would yield similar reactions and should be considered in any thought of a policy change, especially without other confidence enhancing regulations or education. Absent a policy of mandatory labeling, the level of consumer support for labeling strongly suggests the possibility of niche markets for voluntarily labeled non-cloned food products. Such markets would still be of interest to policy makers who may have to help define and certify these claims to support credibility among consumers (Roe and Teisl, 2007). Lastly, as noted by Lockie et al. (2005), new regulatory regimes can stem from the concerns of retailers reluctant to stock certain products or wishing to be able to accurately inform their customers of its absence. In this sense, non-governmental agents could have a key role in food governance and policy. Results here and past experiences noted above with GM foods suggest such outcomes are likely. 7. Conclusion Similar to conclusions two decades earlier with the first GM foods, the FDA has stated that meat and milk from cloned animals are not different from their conventional counterparts and therefore do not require labeling. Such products are currently kept from the market solely under the USDA’s voluntary moratorium. Under these policies, when and if the moratorium ends, consumers will be unable to tell whether the milk they purchase is from a cloned cow or a conventionally bred cow. This study used field experiments at four distinct locations to examine the reactions of various market segments to the introduction of potentially cloned cow milk. A key finding was that nearly a third of participants were completely unwilling to try even a small cup of milk that may have come from a cloned cow. This level of avoidance, just from uncertainty over the source of the milk, was somewhat unexpected. It is easy then to imagine a market response where the labeling of milk guaranteed not to be from cloned cows could appear soon after the end of the voluntary moratorium as some companies may seek to capture this group of consumers. This would be reminiscent of the appearance of voluntary rBST-free labels on milk. Given how the original non-rBST niche expanded to include nearly all milk on store shelves, producers with cloned cows and policy makers may wish to consider the experience of that technology’s rollout to see what lessons it may have for them. Results did suggest some potentially positive aspects for those interested in introducing products of cloned animals. Nearly a quarter of participants did not require any compensation to switch. Additionally, consumers were not necessarily opposed to milk from cloned cows but simply wanted to be able to identify such products when making purchasing decisions. Large percentages of participants at all locations wanted labeling required in contrast with the conclusion of the FDA and regardless of their willingness to accept such milk. There may thus eventually be pressure for them to reconsider their current policy. The remaining slightly less than half the sample was willing to switch their conventional cup of milk for one that may have come from a cloned cow if compensated. These numbers combined with those not requesting any compensation show more potential acceptance than previous studies. For instance, Brooks and Lusk (2011) found that barely over 30% of consumers were willing to consume products from cloned animals while Jones et al. (2010) found that 59% were willing to pay a premium to avoid products from cloned animals. None of the previous studies however were non-hypothetical nor did they use the “may or may not” phrasing that will exist in the marketplace. Concerns over consumer acceptance may thus have been overestimated prior to

6. Policy implications The results suggest important aspects for policy makers to consider, and that current policy may not be sufficient. As Lusk (2012) discovered, most consumers want more government policy involvement in food markets, particularly for technologies such as cloning and GM foods. One obvious area from the results above is the potential for government help in education that could be aid in consumer acceptance and confidence. The number of consumers that had neutral opinions on cloning suggests opinions can still be shaped. The source and type of education about cloning could be instrumental in the success or failure of cloned animal products in the marketplace. Policies such as education that aid acceptance would be important if such products increase social welfare. The majority of participants in this study were willing to consume a small cup of milk that may have come from cloned cows if compensated, with the scale of acceptance suggesting consumer welfare may be enhanced if the introduction of milk from cloned animals is accompanied by lower milk prices, similar to benefits from GM crops to consumers (Brookes et al., 2010; Qaim, 2009). More research, however, would be needed to examine the possible extent of this. Similar to GM foods, perhaps the most poignant policy concern is the question of labeling food products from cloned animals. The developing changes in GM labeling requirements demonstrate that further debate is possible despite the lack of a current mandatory label. Several consumer groups have already noted their aversion to the FDA’s position and, compellingly in this study, more than 80% of participants want mandatory labels on milk from cloned cows. Concerns though have arisen that mandatory labels may serve as warning signals to consumers who lack understanding of the safety of a technology. Artuso’s (2003) model showed labeling a food as being GM, in the absence of other regulatory measures to aid consumer confidence, could be detrimental to demand. For products with the potential to improve social welfare, such labeling could instead lower it, again stressing an educational need. 7

Food Policy 74 (2018) 1–8

K. Britwum, J.C. Bernard

this study. Additionally, the low level of knowledge regarding animal cloning appears to offer an avenue for consumer education. While this wouldn’t necessarily lead to greater acceptance or willingness to consume, the finding that nearly half were beginning with a neutral opinion of the technology suggests an opening to shape consumer opinion in either direction. It would perhaps be best for producers and policy makers to educate and inform consumers about animal cloning before such products are introduced. Results here suggest this would need to entail more than noting such food products are “as safe” as their conventional counterparts which was the FDA conclusion provided here. Examining differences across the four locations further showed how various segments of the population could be expected to react. Those segments tending to be more averse to the cloning technology, such as natural food store shoppers, would be most negatively impacted from the uncertainty inherent in the current policy. It is doubtful, however, that a price reduction alone would appeal to this consumer group, especially for those whose opposition to the technology is strong. This again suggests a form of labeling may appear, introduced by producers looking to avoid the need to give a price discount, or hoping for a nonclone premium, rather than from policy makers. It would be beneficial to expand upon this study. Two obvious avenues would be in examining other products of cloned animals and in conducting non-hypothetical studies in other states or countries as the sample here does not necessarily represent all cloned animal food products or all consumers in the U.S. Also important would be to see if an extra generation from a cloned animal would reduce concerns, for instance, whether milk from the offspring of a clone would be more acceptable to drink. As a final note, there was one subject that made an impression on the research team, suggesting there may be other concerns for the success of cloned animal products, perhaps more in the realm of psychology. After consuming the conventional milk, this consumer requested and was given the milk that may have been from a cloned cow. Having done so, and in spite of the milks being the same date and brand, they announced their verdict that “it did taste a lot different than the conventional milk.”

Colson, G., Corrigan, J.R., Grebitus, C., Loureiro, M.L., Rousu, M.C., 2015. Which deceptive practices, if any, should be allowed in experimental economics research? Results from surveys of applied experimental economists and students. Am. J. Agr. Econ. 98, 610–621. Costanigro, M., Lusk, J.L., 2014. The signaling effect of mandatory label on genetically engineered food. Food Policy 49, 259–267. Crespi, J.M., Lusk, J.L., Cherry, J.B.C., Martin, L.E., McFadden, B.R., Bruce, A.S., 2015. Neural activations associated with decision-time and choice in a milk labeling experiment. Am. J. Agr. Econ. 98, 74–91. Ellison, B., Bernard, J.C., Paukett, M., Toensmeyer, U.C., 2016. The influence of retail outlet and FSMA information on consumer perceptions of and willingness to pay for organic grape tomatoes. J. Econ. Psychol. 55, 109–119. Faass, J., Lahr, M., 2007. Towards a more holistic understanding of American support for genetically modified crops: an examination of influential factors using a binomial dependent variable. Munich Personal RePEc Archive (MPRA) Paper No. 6124. http:// mpra.ub.uni-muenchen.de/6124/ (Accessed 16 April 2013). Food and Drug Administration, 2008. Press Release, January 15. http://www.fda.gov/ NewsEvents/Newsroom/PressAnnouncements/2008/ucm116836.htm (Accessed 6 February 2013). Fox, J.A., 1995. Determinants of consumer acceptability of bovine somatotropin. Rev. Agric. Econ. 17, 51–62. Govindasamy, R., Onyango, B., Hallman, W.K., Jang, H., Puduri, V., 2008. Public approval of plant and animal biotechnology in South Korea: An ordered probit analysis. Agribusiness 24, 102–118. Hey, J.D., 1998. Experimental economics and deception: A comment. J. Econ. Psychol. 19, 377–395. Huffman, W.E., Shogren, J.F., Rousu, M., Tegene, A., 2003. Consumer willingness to pay for genetically modified food labels in a market with diverse information: Evidence from experimental auctions. J. Agric. Resour. Econ. 28, 481–502. Huffman, W.E., Rousu, M., Shogren, J.F., Tegene, A., 2004. Who do consumers’ trust for information: The case of genetically modified foods? Am. J. Agr. Econ. 86, 1222–1229. Jones, G., Jennings, S., Ibrahim, M., Whitehead, J., 2010. Will consumers pay a premium for clone-free labeled meat products? Evidence from the 2009 Sunbelt agricultural exposition. In: Paper Prepared for Presentation at the Southern Agricultural Economics Association Annual Meeting, Orlando, Florida, February 2010. Knight, I.B., 2008. U.S. Department of Agriculture Statement on FDA Risk Assessment on Animal Clones. USDA Office of Communications, Washington DC. Lockie, S., Lawrence, G., Lyons, K., Grice, J., 2005. Factors underlying support or opposition to biotechnology among Australian food consumers and implications for retailer-led food regulation. Food Policy 30, 399–418. Long, J.S., 1997. Regression models for categorical and limited dependent variables. Adv. Quant. Tech. Social Sci. 7. Lu, X., White, H., 2014. Robustness checks and robustness tests in applied economics. J. Econometrics 178, 194–206. Lusk, J.L., Fox, J.A., 2003. Value elicitation in retail and laboratory environments. Econ. Lett. 79, 27–34. Lusk, J.L., House, L.O., Valli, C., Joeger, S.R., Moore, M., Morrow, J.L., Traill, W.B., 2004. Effect of information about benefits of biotechnology on consumer acceptance of genetically modified food: evidence from experimental auctions in the United States, England and France. Eur. Rev. Agric. Econ. 31, 179–204. Lusk, J.L., House, L.O., Valli, C., Joeger, S.R., Moore, M., Morrow, J.L., Traill, W.B., 2005. Consumer welfare effects of introducing and labeling genetically modified food. Econ. Lett. 88, 382–388. Lusk, J.L., 2008. Final Report on Consumer Preferences for Cloning. U.S. Department of Agriculture Economic Research Service. Lusk, J.L., 2012. The political ideology of food. Food Policy 37, 530–542. Moon, W., Balasubramanian, S.K., Rimal, A., 2006. WTP and WTA for non-GM and GM food: UK consumers. In: American Agricultural Economics Association Annual Meeting, Long Beach, California, July 2006. Nonis, S.A., Hudson, G.I., Hunt, S.C., 2010. Should we label products from clones? An exploratory study of beliefs, attitudes and food safety information on consumer purchase intentions. J. Mark. Dev. Competitiveness 5, 95–106. Puduri, V., Govindasamy, R., Lang, J.T., Onyango, B., 2005. I will not eat it with a fox; I will not eat it in a box: What determines acceptance of GM food for American consumers? Choices 20, 257–261. Qaim, M., 2009. The economics of genetically modified crops. Annu. Rev. Resour. Econ. 1, 665–694. Roe, B., Teisl, M.F., 2007. Genetically modified food labeling: The impacts of message and messenger on consumer perceptions of labels and products. Food Policy 32, 49–66. Rousu, M.C., Colson, G., Corrigan, J.R., Grebitus, C., Loureiro, M.L., 2015. Deception in experiments: Towards guidelines on use in applied economics research. Appl. Econ. Perspect. Policy 37, 524–536. SAS Institute Inc., 2014. SAS/ETS® 13.2 User’s Guide. Cary, NC: SAS Institute Inc. Wertenbroch, K., Skiera, B., 2002. Measuring consumers’ willingness to pay at the point of purchase. J. Mark. Res. 39, 228–241.

Acknowledgements The authors wish to thank Lenna Schott for her tremendous help with the field experiments and give a special thanks to Amalia Yiannaka for her review of the manuscript. References Aizaki, H., Sawada, M., Sato, K., 2011. Consumer attitudes toward consumption of cloned beef. The impact of exposure to technological information about animal cloning. Appetite 57, 459–466. Artuso, A., 2003. Risk perceptions, endogenous demand and regulation of agricultural biotechnology. Food Policy 28, 131–145. Becker, G.M., DeGroot, M.H., Marschak, J., 1964. Measuring utility by a single-response sequential method. Behav. Sci. 9, 226–232. Bernard, J.C., Bernard, D.J., 2009. What is it about organic milk? An experimental analysis. Am. J. Agr. Econ. 91, 826–836. Bratspies, R.M., 2007. Some thoughts on the American approach to regulating genetically modified organisms. Kansas J. Law Pubic Policy 16, 101–131. Brookes, G., Hsiang, T., Tokgoz, S., Elobeid, A., 2010. The production and price impact of biotech corn, canola, and soybean crops. AgBioForum 13, 25–52. Brooks, K.R., Lusk, J.L., 2011. U.S. consumers attitudes toward farm animal cloning. Appetite 57, 483–492. Butler, L.J., Wolf, M.M., Bandoni, S., 2008. Consumer attitudes toward milk products produced from cloned cows. J. Food Distribution Res. 39, 31–35.

8