|| Home || SEARCH|| Contact || Free Research Publications  ||  Eshop || Privacy Statement ||

Download full report (PDF 377k - HINT: right click and save the file to your hard disk before opening)

Impact of Genetic Engineering on Consumer Demand

by Dr Kate Owen University of Technology,  Sydney,  Professor Jordan Louviere, University of Technology, Sydney

Mr John Clark, Decided Advantage

February 2005

RIRDC Publication No 05/015  RIRDC US-92A / UTS-7A

Executive Summary
The research outlined in this report was prompted by the continuing uncertainty over consumers' responses to genetically modified foods both in Australia and overseas. This uncertainty has placed producers in a dilemma given the real benefits presented by adoption of gene technology. As a consequence, most industries have chosen to avoid pursuing commercialisation of GM crops, instead, preferring to adopt a “wait and see” strategy. However, in following such a strategy they face the risk of declining markets, and lost potential markets, as other exporting countries gain cost and other advantages through adoption of gene technology.

Improved understanding of the expected response of consumer markets to genetically modified products is considered necessary to assess the relative import of these conflicting risks and to provide economic modellers with a reference for modelling expected trade flows. In response to this, the objectives for the research reported here were to evaluate the level of consumer concern over genetically modified foods in Australia, and then to quantify consumers' response to modification of selected foods within a purchase context. Although the concern for many producers is the response of foreign consumer markets, the domestic consumer market is influential in setting the regulatory environment in which gene technology development and commercialisation occurs. As such, this project focuses on the attitudes and responses to GM of Australian consumers.

The research was conducted in two stages. In the first stage we focused on consumers’ attitudes and concerns in relation to the introduction of genetically modified foods. In the second stage our aim was to quantify these concerns by providing estimates of consumers' likely demand response to three foods that had been genetically modified: milk, potatoes, and potato crisps. The studies, their methodology and findings are outlined below.

Stage 1: Concerns and awareness
The research in Stage 1 was based on a survey of 120 consumers in metropolitan Sydney to determine the level of concern over GM foods and to ascertain their awareness of, and confidence in, organisations involved in their distribution and regulation. A small pilot survey of 48 retailers covering the same issues was also conducted to determine their congruence with consumers' concern levels.

Consumers completed a questionnaire that first embedded the GM food issue in a list of 15 known public concerns and then repeated the task for two specific applications of gene technology to food against 13 common food safety issues. In both tasks respondents were presented with repeated subsets of the issues and asked to identify the one of most and least concern in each subset. By embedding the GM food issue in this way we were able to obscure our research focus and so to provide some context in which the level of concern over GM foods could be evaluated.

Following these tasks we made explicit our interest in respondents’ perceptions of GM foods by asking them to indicate their view on GM foods in general and then to indicate their level of concern under two scenarios. In one of these, the scenario was an inaccurate labelling of a food as “contains no genetically modified ingredients”. In the other, they responded to an incident in which a food accidentally contains GM ingredients that are only approved for animal feed.

The next section focused on consumer knowledge of, and confidence in, organisations involved in the distribution and regulation of GM foods. These were first elicited through open-ended questions to gauge unprompted knowledge and then via respondents’ knowledge of a list of such organisations.

In the final section respondents were given hypothetical scenarios in which genetic modifications were made to three products: milk, potatoes, and potato crisps. These products were selected at a workshop of government and industry representatives as products that were commonly consumed, were realistic examples of possible genetic modifications, and that might be expected to evoke different responses given their association with factors such as “nurturing”, “natural”, and “processed”. Respondents were asked to anticipate their purchase response based on five categories: would not buy, buy less, the same, more, or don’t know.

Key results of the survey were as follows:

• Concern over GM foods relative to other public issues is minor. On the scale of 15 issues GM foods was 11th. Issues such as the cost of living, drugs, and unemployment and poverty, were of far greater concern.
• Of 15 food safety issues the two gene technology applications were listed at 6th and 7th. Their level of concern was on par with concern over pesticides and chemical preservatives. Upper most in the mind of respondents were diseases in beef that can be passed to humans and bacteria and disease in foods (note: survey was in March 2001)
• Around 52 per cent of respondents indicated they were indifferent or were interested in the possible benefits of GM foods. However, when presented with the GM scenarios outlined above more than half of these claimed they would be concerned enough to call for something to be done. This was particularly the case when food was found to accidentally contain GM ingredients that are only approved for animal feed. The actions respondents expected included: recall all produce, place advertised warnings, and impose higher standards for sale and heavy fines.
• Only 10 per cent of respondents were aware of the Office of Gene Technology (OGTR) and 26.7 per cent of ANZFA (now FSANZ). Retailers had a higher level of awareness of ANZFA but similar low awareness of OGTR. Respondents placed the burden of responsibility for regulation and for dealing with problems on an undefined “Government” and “Health Departments”.
• Type of product influenced the acceptance of the hypothetical product offerings:
• 75.8 per cent of respondents would reduce or never purchase products with traces of the genetic modification and no perceived benefits.
• 64.2 per cent would buy the same or more of milk modified for low cholesterol and 54.2 per cent would buy the same or more potatoes for low fat absorption.
• 68.3 per cent would reduce or never purchase milk from cows modified to grow faster.


Stage 2: Consumer demand responses
The second stage comprised a large scale survey of 1008 consumers across mainland metropolitan Australia. The aim was to quantify consumers' likely demand response to milk, potatoes and potato crisps that have been genetically modified and was led by the following questions:

1. What is the response to genetic modification without benefits and does this differ where there are tangible consumer benefits?
2. Will consumers be more inclined to purchase a GM product if the price is lower or the product has other attributes they favour (eg preferred brand)?
3. Do responses to genetic modification differ across the three products?
4. Does negative or positive information on GM have a significant impact on consumers’ demand response? The study employed the stated preference method and latent segment analysis to address these questions and identify underlying segments of respondents with different tastes/responses to genetic modification. The modelling results were then used to derive basic Willingness to Pay (WTP) estimates to quantify potential demand responses.


The stated preference method involves simulating the context in which consumers would normally make choices between competing products. This is achieved by designing scenarios in which product characteristics and prices are systematically and independently varied. Consumers are then given multiple scenarios and asked to indicate their preferred alternative in each. These choices are then analysed to develop probabilistic discrete choice models for relevant samples of consumers.

In this study, for each product we designed a set of shopping scenarios based around a choice between an "existing" product and a newly introduced product. Both the "existing" and "new" products could be genetically modified. However, in the new product the genetic modification has a clear benefit while the existing product has no stated benefit when it is modified.

In the scenarios for potatoes and crisps respondents had three options, they could choose either: an existing product, a new product, or not buy. The existing product varies in price and type (brand, variety) and is labelled "GM free" or unlabelled in some scenarios, and "genetically modified" in others (for crisps, the genetic modification switched between oil and potato). The new product varied in price and type but was labelled "organic" in some scenarios and "genetically modified for low fat absorption" in others.

The scenarios for milk had four options: existing product, new product, organic or an option to buy no milk. The organic product varied in price and fat by scenario. The existing product varied on price, brand, fat and vitamin content, and could be "genetically modified" or unlabelled. The new product also varied on price etc but sold itself on being "cholesterol free". However, in some scenarios the label indicated that "cholesterol free" was achieved by replacing fat with oil while in others it was through genetic modification.

During the survey respondents were randomly assigned to one of three information conditions: negative, positive and no information. The negative and positive conditions were presented in the form of a mock press release at the beginning of the questionnaire. Each respondent then completed 16 milk scenarios and 16 of either potatoes or crisps. They also provided information on current purchase behaviour and demographics, and indicated their level of concern over nine issues related to food. We later reduced the results of the issues to two measures that placed respondents on continuums of concern for "food quality and environment" and "fat and cholesterol". These were used with other demographic information as covariates to produce latent segment models.

Key results from this study include:

• Consistent with previous research, in the absence of benefit consumers will avoid genetically modified products. A segment of the sample representing 21, 25 and 34 per cent of the sample in crisps, potatoes, and milk would purchase the GM product only if a discount of between 30 to 60 per cent was given.
• In the same segment, the required discounts were significantly lower for a GM new product, with milk requiring an 11 per cent discount, crisps 33 per cent and potatoes 39 per cent. However, in most segments consumers were less likely to purchase the new product than the existing one across all food types, indicating that for a significant proportion of the sample the “benefit” offered was immaterial or of no perceived value.
• Lower prices did increase the probability of choosing potatoes and crisps that were genetically modified and the effect was significant in most cases but this was not the case for milk. In the milk experiments we also tested whether there was a significant interaction between brand and GM. No interaction was found; branded milk had no greater probability of being chosen than unbranded milk when either was genetically modified (with no benefits).
• Milk was the only one of the three food types where negative information significantly affected purchase responses. These were to avoid the new product altogether or an increased WTP for its oil-based alternative (47 per cent), and greater substitution of the existing product for the new non-GM product.
• Purchase behaviour of crisps was unaffected by the different information environments but potato purchases were marginally affected by positive information. A general effect was increased variance in respondents’ choices. Also WTP to avoid GM in the existing product decreased by around 19 per cent and in the new product reduced by 47 per cent. There was also a substantial drop in WTP for a GM-free product (45 per cent).


Key points and implications
The results of Stage 1 of this research lend qualified support for continued research and development of genetically modified foods, particularly those with real consumer benefits. However, together with the results of Stage 2 it is likely that to commercialise these foods they will need to be heavily discounted, at least in the early stages of market development. This may render commercialisation of some GM foods unviable, particularly if produce is subject to higher costs as a result of compliance with GM regulation (Hatwell and Pluske, 2004; ABARE, 2003) and given the evidence from Stage 2 of little prospect for extracting price premiums for genetic modifications carrying a tangible benefit.

An apparent anomaly in the results of the two stages is the change from a positive intention to purchase milk/potatoes with low fat qualities to negative purchase responses in the choice experiments. These differences are best understood if we consider their contexts. In the first study respondents were simply asked for an indication of their purchase intentions given a particular genetic modification. No context was provided and so it is not possible to discern whether their responses reflect real purchase intentions or are a vote for or against particular applications and/or outcomes (Noussair, 2002). As a vote, positive intentions to purchase a GM food with clear benefits may be more a vote for a desired benefit than the application of GM to achieve this. The choice tasks in Stage 2 are quite different. Here respondents were presented with situations that more closely resembled the shopping context. They could choose between alternatives and were faced with a number of potential trade-offs in price, brand etc. Under these conditions it is less likely that their responses simply reflect votes. Placed together, the implication of the two studies in broad terms is “yes, we like the benefits” but we’re still not sure of the technology and will stick to what we know.

Having said this, it is worth keeping in mind, first, that concerns over GM do not rank highly among everyday consumer concerns and, second, that consumers make many daily purchases using minimal information. A common scenario is shoppers on a busy Saturday morning with one or two children in tow. Their foremost concern is to get through their shopping. Familiar products are chosen rapidly and minimal attention is given to labels. Choice experiments, on the other hand, elevate attribute awareness because their very purpose is to examine the effect of manipulating these. A consequence of this is that the negative responses in Stage 2 may be greater than would be the case in the actual shopping context.

Consistent across the food types that were examined in Stage 2 were three segments. One segment consistently avoided all products and is essentially not in the market (or at least was not participating in the survey). Another had a strong preference for organic product and actively avoided GM. The “organic” group represented 30 per cent of the samples for potatoes and crisps and 21 per cent of the milk sample. The final group (between 21 to 34 per cent of the samples) was price sensitive and actively purchased all products. This is the segment most likely to respond to discounts for GM.

The remaining groups were not always interpretable but showed strong preferences for type of product (eg low fat or full cream milk) or were particularly price sensitive and their responses to other attributes were generally insignificant. An implication from these results is the need for careful market research and segmentation to identify just how economically viable development and commercialisation of a GM food will be, particularly where the benefits may not be exclusive to the technology.

A major concern to producers considering producing genetic modification in relation to crops or animals is the impact on demand of negative publicity. The response in milk indicates that this can be significant. It is not possible from this study to say whether all negative incidents or publicity would result in significant rejection of a product or for how long consumers would withdraw from its purchase. Given the positive response to optimistic information on GM potatoes it may be that demand could be restored reasonably quickly with appropriate assurances on a product's safety and benefits.

Contributing to the potential for fluctuations in response is the apparent low latent concern over GM foods. While this means there is unlikely to be resistance to ongoing development of GM foods, it is a double-edged sword since it also reflects the level to which consumers will inform themselves on gene technology and about the existing regulatory framework to ensure the safety of GM foods to health and the environment. The results from the first consumer survey indicated minimal awareness of the specific organisations responsible for regulating GM release and distribution. The recent Millward Brown (2004) report indicates that this has not increased significantly despite concerted efforts by government and industry. Although confidence was reasonably high for some government organisations, the level of confidence in farmers and producer groups, scientists, and in State and Federal politicians was generally low (less than 50 per cent had confidence).

A related problem facing producer groups is that specific GM incidents are often generalized indiscriminately. Some form of grading or categorisation of different genetic manipulations so that the consumer can easily evaluate the product they are consuming may assist to reduce this. No doubt such a system would be difficult to devise and has implications for those developing or producing foods using technology with less certain outcomes. Further research is needed to determine the nature and scope of generalisations and discussion is required to identify possible methods for their reduction and the merit of these vis a vis living with likely demand fluctuations.

The move from full sample models to segments improved explained variance significantly but did not provide clearly interpretable segments in terms of consumer groups. Rather, the segments represent coarse purchase behaviors and psychographic predispositions. Kontoleon (2003) among others has found segments with individual characteristics, such as, optimists, safety concerned and ethical opponents; the latter segments representing over 45 per cent of his sample. Although we can draw some parallels in respondent’s behaviours in this study the segments do not have such distinct psychographic or cognitive interpretations. Further research on the nature of underlying segments would add to our understanding of adoption of GM products. One avenue that might provide additional insights to expected demand for GM is product adoption behaviour. Kontoleon’s optimists could be seen as early adopters. Baker and Burnham (2002) identified one of their segments as consistent with such a group and suggested that, as leaders of new product adoption, this group could be targeted for specific GM applications.

Although the results of the choice modelling appear generalisable to other foods, and the WTP estimates are in line with previous studies, there is a need to further substantiate the findings through examining other food products in a similar manner. In particular, for potatoes and crisps we used an organic product as an alternative to "new" GM products which might have influenced respondents' choices and so other combinations need to be explored. Also, the results from the milk scenarios suggest that the product benefit must be substantial relative to non-GM options and unique to the gene technology. Choice scenarios where there are shades of differences between products are common situations for consumers and GM benefits need to be assessed against these.

Finally, we have employed the stated preference method and discrete choice modelling using scenarios that mimic as closely as possible the choice situation. In this regard our GM labelling was limited to “genetically modified” except where it had direct consumer benefits which were promoted.

Our reasoning was that the food industry would simply comply with FSANZ guidelines on labelling and avoid any further explanation or justification for using a GM ingredient. A useful addition to the current study would be to determine whether including information on product labels that emphasises the production or environmental benefits of the GM has a positive effect on purchase probability.

 

Site map Contact us Feedback Links Disclaimer Privacy Statement