Submission
INTRODUCTION AND SUMMARY
Australian and New Zealand Health Ministers meeting as the Australia New Zealand Food Standards Council (ANZFSC) on 22 October sought submissions on the labelling provisions of the draft standard A18 covering Food Produced Using Gene Technology.
Regulation should be tailored in response
- first, to the level of risk of the untoward occurring;
- secondly, to the degree that the incentives in the remedies of the normal operations of law inadequately encourage appropriate care; and
- thirdly, to the degree that a government expert body can be better informed of dangers and how to avoid them than commercial organisations.
Our recommendations are made cognizant of public concerns on the GM technology but in the light of well founded principles adopted in all jurisdictions designed to reduce regulation. The recommendations are as follows:
First, the definition of GM products falling under the regulatory oversight of the Gene Technology Regulator (GTR) should exclude those products that involve modification solely within an existing species. Confining the regulatory oversight to where genes are introduced from a different species will reduce the regulatory costs and improve the regulatory focus of both Government and industry. It will do so while retaining oversight on those product modifications that pose risks not seen with traditional breeding techniques.
Secondly, we agree with the eminently sensible provision in the Editorial Note in Clause 1 which says,
"a food derived from an animal, which is not itself a product of gene technology but which has been fed on a food produced using gene technology, is not a food produced using gene technology merely because of the feed used."
Thirdly, we favour no mandatory requirement on food suppliers to indicate whether their produce contains GM inputs. This will allow suppliers to respond to consumer needs rather than regulators determining these needs.
Fourthly, a label "Not sourced from GM products" could be adopted where there is an agreed upper level of possible GM content (say 3% or the level that is adopted in the EU or Japan).
Fifthly, GM-free could be maintained as a label if the supplier is confident that the goods do not contain any GM sourced inputs.
FOOD LAWS INFORMATION IN CONTEXT
Regulatory Approach to Food
The buying and selling of food, like that of most other goods is a transaction offering mutual advantages to both parties. There is no requirement on buyers, confronted by many alternative offerings, to purchase a particular product. The seller is in a highly competitive environment and must appeal to sufficient numbers of buyers.
Sellers, if left to determine for themselves the content of the label on their products, subject to the requirement of accuracy, will respond to the needs of their actual and prospective customers. Their labels will seek both to inform and entice the buyer who has some 30,000 lines to choose from in a supermarket.
Food regulations are primarily aimed at ensuring the product is safe to consume. Labelling provisions reinforce this. Thus, with packaged foods, in the interests of consumer health, information must be included on the label about additives to the food which enhance its taste, improve its texture, facilitate a longer shelf life etc.. However, in general, a therapeutic claim cannot be made (1).
Food regulation in legal terms has migrated over the past century from caveat emptor "let the buyer beware", to caveat venditor, "let the seller beware". As with most shifts in legal liability, there are good common-sense reasons behind this change. Chief among these is the change from the preponderance of fresh food purchases to packaged food.
In the case of fresh food, it was just as easy for the buyer to satisfy himself of the quality, wholesomeness and freshness of the product as it was for the seller. By contrast, packaged food has undergone processing. Its components cannot easily be determined by the customer and it makes sense for the liability for selling goods of below the mutually expected quality to fall upon the seller.
Food as an Experience Good
Food can be described as an "experience" good. When a particular line is first bought the consumer is unlikely to be convinced it meets her taste and other requirements. It might be argued, given this initial state of ignorance, that a seller has an incentive to mislead the consumer about the product. Indeed, although only truthful claims are permitted regarding a product's qualities, some "puffery" (statements like "the world's best chocolate bar", "the tangiest lemon drink") is perfectly acceptable.
However, a new product line involves considerable marketing and formulation expense and no seller could survive on one-off sales. Hence the seller must attempt to gain repeat sales.
This need to gain repeat sales means there is less chance of major disappointment where a consumer buys food than with durable consumer goods. It follows that the need for government regulations requiring information is correspondingly reduced. In addition, where mandatory information is required, this may reduce the available space for information that the seller considers to be more important to the buyer. In such cases mandatory labelling requirements can detract from consumer benefits.
Labelling Information: Regulation and Voluntary Provision
Notwithstanding the relatively low costs of a disappointing food purchase, the demand by the regulatory authorities for mandatory information on a food label has been increased considerably over recent years. This has extended beyond standard physical characteristics and, largely in response to pressure groups, it is also to enhanced consumer awareness and concern with dietary matters. The focus of the information requirements demanded has been to guard against deceit (food contents, weight, water content) or to ensure people with allergies to certain components are warned in advance of their inclusion in the product.
There are sound reasons both for verifying that a particular food offering is wholesome and for placing particular requirements on its label regarding ingredients and additives. Some manufacturers are likely to have less knowledge than a specialist government health agency about consumer reaction to an additive. In some cases, a manufacturer may wish to market a combination of products, which, though wholesome in themselves, are known to have adverse effects when blended together. Although in such cases liability law would provide redress and therefore cause suppliers to exercise appropriate caution (and potential loss of reputation is a very effective incentive), it is considered preferable to prevent the mishap before the event and require prior approval of the product.
Similar considerations apply with labelling to prevent inadvertent consumption of those with allergies to the product.
Nevertheless, regulations add costs to production that are eventually paid for by consumers. They also bring costs for regulatory policing that are largely paid for by taxpayers. In recognition that much -- perhaps most -- regulation detracts from, rather than adds to, national well-being, regulatory reviews under National Competition Policy, adopt a presumption against the need for regulation.
The mutuality of the buyer and seller relationship and the highly competitive environment for food sales means that sellers will respond to market requirements both in the content of their offerings and in describing them. Beyond a requirement that claims be truthful, there need be little regulation of this to prevent consumer deceit.
If consumers have a need for a particular feature from a food, in a competitive market suppliers will provide this. It might be argued that this response will be inadequate in some circumstances, hence the requirement of labelling to provide fat, carbohydrate, and protein content.
However, in the main these requirements simply led a labelling trend that was previously in train in response to consumer wishes. Thus, various forms of sugar free confectionery have been brought to the market to respond to consumer needs as have a variety of "healthy" breakfast cereals.
Moreover, suppliers have over time responded to consumer needs even where these are of a minority nature and cover features of no scientific worth. Thus, halal and kosher meat are readily available to those who (in Australia) comprise a tiny minority wishing to avoid certain products or preparations on religious grounds. In these responses, suppliers weigh up the additional costs of the labelling (and more importantly the product segregation and modification that accompanies it) with the benefits that consumers wish to see.
In line with the strictures set out in this submission's Introduction, we consider regulatory oversight should be confined to those products where there may be genuine risk and where regulation might make a difference.
LABELLING COSTS
The KPMG Report's Findings on Costs
It is necessary to preface any discussion of the proposed labelling laws with the reminder that the products concerned are perfectly safe. Either they have been developed by the same processes that have given the world a safe and abundant food supply or they have been specifically approved on scientific advice as safe. Some jurisdictions have adopted criteria of "substantially equivalent" as a test beyond which specific regulatory approval may be required. Australian and New Zealand Health Ministers have indicated that they wish to adopt a more stringent regulatory approach than this. That disposition reflects the view that the public wishes to be informed, even if the products are safe.
The work undertaken by KPMG evaluates the cost of mandatory GM labelling. It estimates that the requirement would mean auditing and testing between 18,000,000 and 48,000,000 streams of assessment. KPMG identify several different types of cost:
- Set-up costs
- On-going monitoring costs
- Indirect costs like training and levies to pay for industry supervision
- "Opportunity" costs including affect on market shares from making a labelling statement, need to redesign products, and loss of confidence among consumers.
Product costs are estimated to increase by 5-15% and an additional 3% for compliance. Such estimates are not inconsistent with those published by MacKenzie in the New Scientist (2).
The bottom line estimate for Australia from the KPMG study is $3 billion in the first year, reducing to $1.5 billion per year.
Other Costs Estimates from Mandatory Labelling
MacKenzie, quoting research from Wye College, says US soya bean processors charge 6-9% for segregating beans for their protein and oil content, and Brazilian soya beans are distinguished between GM and non-modified at a premium of 10-15%. However, these costs are based on a tolerance for GM contamination of around 1%.
The original data on which much of Mackenzie's report was based estimated that Identity Preservation (segregation) can add as much as 150% to costs. It also cited cases where some tolerance is permitted, with costs as little as 0.1%. These cost variabilities underline the difficulties involved in attempting to quantify the effects of a regulatory solution and to ensure that solution is compatible with any benefits it might bring in terms of consumer choice.
All the evidence therefore points to a wide range of costs likely to emanate from mandatory labelling. These reflect the great array of products and the manner in which they are presently handled.
Some products are carefully graded and monitored from grower to final consumer. McDonald's fries and hamburger beef and foods offered by many other fast food outlets are cases in point. The total segregation of these products' inputs is required through the chain for quality control to ensure as close a consistency as possible in the final product.
In such cases, the administrative costs of mandatory labelling could be low. The final seller is already undertaking a considerable surveillance of the product at all stages of its production. However, even in these cases the firms set their own levels of tolerance for blemishes and presumably do not involve themselves in total testing at each stage. They certainly would not test for material that is substantially equivalent in all features to the required input.
Costs Stemming from Market Versus Regulation Induced Requirements
The KPMG report may have overstated some costs. Costs are of two kinds: those that the consumer and producer willingly incur (and the consumer eventually pays), and those that the consumer is made to incur. Where producers in response to consumer demand provide more information and as a result need to undertake more control and additional gradings, the costs are offset by the value the consumer places on the increased information. These costs are recouped in higher prices and all parties benefit.
Information provided as a result of this consumer-oriented outcome is no more a (wasted) cost than the costs involved in flavour enhancers, preservatives and robust packaging. By contrast, information mandated by government has no such market test of its worth. Commonly, the information is mandated in response to contemporary and ephemeral health fads, as was the case with additives. Usually it is associated with prevention of deceit. More often than not the mandatory information is not valued by the great bulk of consumers and is a deadweight cost.
Kellow (3) points out that many of the claims made of foods being "organic" are absurd since the point of obtaining purity is to eliminate the organic residues. That said, if the consumer genuinely demands a particular feature, suppliers will meet these demands. "Genuine" demand is that which is backed up by a capacity and willingness to pay. We see no end of products the demand for which our predecessors would have considered frivolous, even antinomian, like bottled water and range fed beef. But demand by the consumer backed by a capacity to pay any extra costs entailed is the ultimate arbiter of worth.
In this respect there are benefits in the ability of suppliers to make the claim "GM free". Even if the most authoritative scientists and the vast bulk of consumers see little or no value in such a claim, if some consumers value it then it has worth. We see analogous statements on food that is halal or kosher. Consumers of these foods are willing to pay a premium for them.
There is however an important difference between this and the GM food issue. In the case of people wishing to specify a specific form of killing of their meat, a form that would not be objectively considered superior to modern slaughter, they do not insist on all meat being segregated. They merely insist, quite appropriately that the meat they consume corresponds to these production methods. As such the increased costs are borne largely by themselves. Only to the unlikely extent that their requirements deny some scale economies can there be said to be any costs imposed on other users. And even if there were some such costs, the minorities' requirements would remain legitimate.
The voluntary labelling of halal and kosher products raises issues addressed in Clause 10(e) of the proposal. This argues that,
"the provisions ... may specify additional labelling or other information requirements in relation to food produced using gene technology listed in column 1 of the Table where ... the genetic modification may raise significant ethical, cultural or religious concerns regarding the origin of the genetic material used in the genetic modification".
With regard to religious objections to GM foods, there appears to be no concern among mainstream Christians. Indeed Bishop Sgreccia, Vice President of the Vatican's Pontifical Academy for Life, following a two year study by his members has said,
"We are increasingly encouraged that the advantages of genetic engineering of plants and animals are greater than the risks. The risks should be carefully followed through openness, analysis and controls, but without a sense of alarm." (4)
Similarly, the premier Indian geneticist, Professor M S Swaminathan has not indicated religious concerns with the technology on behalf of Hindus, while two predominantly Muslim nations, Malaysia and Indonesia, have embraced the technology.
We have previously published a ball-park cost estimate for GM segregation of around $1 billion per annum. This was based on an estimated average 6% increase in the cost of food handling largely representing the dual system required to ensure segregation of foodstuffs. The estimated 6% average is a blend of figures ranging from negligible costs where rigorous quality control of the product already occurs to costs approaching 50% of the basic food where the product is homogenous and presently has no need for segregating by source.
In few cases will the costs be negligible: even where agricultural produce is subject to strict gradings -- and the number of wheat grades has risen from 3 to two dozen over the past 20 years -- there is seldom a requirement for strict segregation. With grading all that is normally important is that the produce meets certain standards, like protein, moisture content etc. The proposed labelling provisions for genetically modified foods are much more rigorous:
- there is no provision for minor impurities
- unlike grade specifications, there is no possibility of offsetting grades
- in many cases there is no objective test to demonstrate the food is GM free and in other cases specific tests will need to be developed
For these reasons the costs may be greater than the industry expects and some major producers have indicated that they can never make a categorical statement, "GM free", since this would entail constant tests due to there being no possibility of verifying the precise source of every grain.
It is not good regulatory practice to insist on absolute levels of purity. All things have some level of impurity and food may contain inputs in trivial quantities that might be harmful if consumed in large quantities. GM is becoming widely adopted across a great many plant species and few processed foods will in future be able confidently to carry a claim "free from genetic modification", especially once food additives and processing aids are included. Given that much GM derived food is "substantially equivalent" to non-GM food, it may even be difficult to make such absolute claims about the main product inputs.
The Response of Health Ministers to the KPMG Report
Health Ministers did not accept the KPMG cost analysis, arguing that it did not have regard to:
- changes to the draft standard made after KPMG was commissioned; and
- the exercise of due diligence in compliance.
The Health Ministers said, "The draft Standard clarifies the definition of genetically modified food for the purpose of labelling.
"Due diligence in compliance with the Standard relies on appropriate documentation throughout the supply chain. Testing would only be required where verification of documentation was needed (i.e. where the paper trail is unreliable)."
The effect of these two disagreements is difficult to assess. The former statement is unlikely to have a substantial impact, especially since, as discussed later, the standard as presently defined is extremely onerous.
With respect to the latter, the effects are likely to be minor: whether or not testing is required a reputable business would need to fully satisfy itself of the accuracy of a statement prior to making it. And if there is to be no tolerance of GM product that might inadvertently be part of the overall mix (because a supplier has included it with a more substantial load, it has been transported in containers that previously transported GM product, traces of which remain, and so on) the supplier will need to continuously test the product at each stage.
This gives rise to an unintended consequence, which was not addressed by either KPMG or the ministers. The spotlight on GM foods means the penalty for making no statement, or making a statement "free from genetic modification" incorrectly, is likely to be severe in terms of publicity and consequent opprobrium. Thus, major firms with strong brand names and reputations are likely to be risk averse in sourcing product inputs. They are likely to avoid smaller suppliers who usually have less investment in reputation than larger well known firms. Aside from reducing opportunities for small businesses -- and especially start-up businesses with no reputation -- the result may be a reduction in competition. As the Productivity Commission has demonstrated (5), competition is vital to productivity growth.
Implications of Onerous Labelling Laws on Competitiveness
It is therefore quite onerous to require foods to include a "genetically modified"/"may contain" label if it does, or may, contain GM ingredients. Such provisions may require all the ingredients to be tracked from field to plate. It may result in the new technology being forestalled.
Denial of lowest cost inputs to Australian food producers and denial of the most productive varieties to growers is particularly serious for Australia. Some argue that "GM free" could become a positive marketing tool for Australia. Some even maintain, quite incorrectly that specific sales have been made on the back of this claim (in particular canola to Europe which displaced Canadian canola, allegedly because the latter could not be guaranteed to be GM free).
The enhanced competitiveness of GM products means denial of the technology to a world class and low cost agricultural producer like Australia, and would mean a loss of competitiveness. The vast variety of minor changes to productivity from GM crops and the importance of agriculture to this country means failure to adopt the technologies would have an outcome well beyond the local level. Indeed, the impact would be evident beyond the agricultural sector.
Although there is no suggestion in official circles of banning the technology (though many of the voices most strident in the debate seek this), onerous labelling provisions could have an effect similar to a ban. If the labelling were to require attention to be drawn to a feature of the product's derivation that is totally irrelevant to its composition, consumers may seek to avoid the product. Consumer avoidance may be even more likely to the degree the information is required by Health Ministers, with the inheritantly strong inference this contains that the product is not wholesome.
REGULATORY PROPOSALS
There are some specific matters covered in the proposal which we wish to address.
First, the regulation should be confined further than is proposed. The definition in the first part of Clause 1 is too wide. That definition says
"a food produced using gene technology means a food which has been derived or developed from an organism which has been modified by gene technology, and includes any substance regulated as a food additive or processing aid.
It should be confined only to those aspects of the technology that pose risks not commonly seen with traditional breeding techniques.
It may be useful to classify GM into two sorts of technologies: those involving manipulation of genes within a species and those involving the introduction of genes from another species.
In the main, Genetic Modification is expediting the traditional cross breeding that has allowed mankind to feed itself. GM employing genes from the same species does nothing that traditional breeding methods cannot do in a laborious and time consuming way by cross breeding and back-crossing. In short, it effects changes to a species much more rapidly, more precisely, usefully, and surely even more safely.
All plants have their natural poisons to deter predators and the intensification of such natural poisons through selective breeding or GM techniques could have a harmful effect on consumers. But the risks are less from manipulating the genome of a particular organism.
Where GM involves introducing genes from an entirely different species, arguably this means a more radical change to the species DNA. Under such circumstances additional oversight with more comprehensive labelling is more readily justified.
Not only is the intra-species modification no more likely to create risk through GM technologies than through traditional approaches, but public concern with the technology stems from well publicised anxieties that GM products will be developed containing genes introduced from alien species. Pig genes in GM tomato paste is a case often cited.
It follows that the definition of GM products falling under the regulatory oversight of the GTR should exclude those products that involve modification solely within an existing species. Confining the regulatory oversight to where genes are introduced from a different species will reduce the regulatory costs and improve the regulatory focus of both Government and industry. It will do so while retaining oversight on those product modifications that pose risks not seen with traditional breeding techniques.
Secondly, we agree with the eminently sensible provision in the Editorial Note in Clause 1 which says,
"a food derived from an animal, which is not itself a product of gene technology but which has been fed on a food produced using gene technology, is not a food produced using gene technology merely because of the feed used."
To do other than this would open an unmanageable vista of regulatory oversight and offer no possible advantages other than a simple pandering to ill-informed opposition to any innovation in this field.
The digestive systems of animals bombard their with sufficient potency that it is radically transformed prior to forming part of the organism which humans, in turn, would consume. Animals' feed content, unless it was a poison, could not adversely impact on its nutrition to humans and certainly could not, in the definition the draft uses, "alter the heritable genetic material of living cells or organisms".
Thirdly, the Standard needs to offer consideration of a threshold of where a good is GM. Earlier we pointed out that it is bad regulatory practice to insist on absolute levels of purity. In the case of GM inputs into food, the requirement for purity is not even targeted at a health concern.
The bulk of the costs involved in making a claim "GM free" is verification of the paper trail. GM products will only be adopted where they are superior to the alternatives. But the superiority is unlikely to be overwhelming and, like organic produce, non-GM products will continue to be available if there is a sufficient market segment prepared to pay a premium price.
Thresholds are commonly used in all regulatory procedures. Whether it is "pure wool" or "handicraft" goods or even "Made in Australia", some toleration is essential if consumers are to be able to exercise their preferences without being obliged to pay an intolerable price. Without this, an absolute standard will severely constrict the availability of products as the technology is extended into increasing varieties of basic foods. Allowing thresholds will enable the products to be marketed under a variety of banners including an absolute "GM free".
All things have some level of impurity and food may contain inputs in trivial quantities that might be harmful if consumed in large quantities. GM is becoming widely adopted across a great many plant species and few processed foods will in future be able confidently to carry a claim "free from genetic modification", especially once food additives and processing aids are included. Given that much GM derived food is "substantially equivalent" to non-GM food, it may even be difficult to make such absolute claims about the main product inputs.
We favour the adoption of a spectrum of labels being permitted, largely driven by market rather than regulatory requirements. The labelling issue under control of the Health Ministers should properly be confined to the area of their Ministerial accountability: that the food on offer within jurisdictions is safe. The movement of Health Ministers into consumer and industry policy takes them outside of their area of expertise and could undermine their credibility within that core area.
We favour no mandatory requirement on food suppliers to indicate whether their produce contains GM inputs. This will allow suppliers to respond to consumer needs rather than regulators determining these needs.
Beyond this we favour the possible use of different labels should suppliers wish to use them.
A label "Not sourced from GM products" could be adopted where there is an agreed upper level of possible GM content (say 3% or the level that is adopted the EU or Japan).
"GM-free" could be maintained as a label if the supplier is confident that the goods do not contain any GM sourced inputs.
ENDNOTES
1. This provision is becoming increasingly strained both as a result of producers responding to the consumer's increasing health consciousness and because some foods are "healthier" than others. Some even have demonstrable therapeutic qualities like the prevention of osteoporosis; GM techniques are likely to create more such products.
2. How to Price What We Put on Our Plate, Debora Mackenzie, New Scientist, 27 February 1999.
3. Risk Assessment and Decision-Making for Genetically Modified Foods, Richard J. Wood, Biotechnology Backgrounder, Oct 1999
4. St Louis Review Oct 12 1999
5. Macreconomic Reforms and Productivity, Productivity Commission, November 1999.