Don’t Take That Bet

lottery

The Yale-based Environment 360, an excellent, careful and well edited on-line magazine, ran a story recently on GoodGuide and related efforts to guide consumer behavior by providing information of every item in the stores. Written by Marc Gunther and headlined, “Betting on Technology to Help Turn Consumers Green,” the article tells this story:

U.S. consumers tell researchers they want to buy environmentally friendly products, but so far they haven’t been doing that on a large scale. Now a host of companies and nonprofits are trying to use new technology — from smartphones to social networking — to make it easier for buyers to make the green choice.

The theory is quite simple: tell the customer everything know the product’s ingredients, and they will follow by making the rational choice and buy the least harmful or most beneficial item. Standard market economics depends on something like this, as it is based on decisions made with “perfect’ knowledge. Otherwise, each transaction creates an externality: some outcome not present in the buyer’s calculating machine, situated in his or her mind. I have written that unsustainability is one giant externality--an unintended consequence--of acting in ignorance of the harms buried in every cultural act.

Standard regulatory economics uses this theory in designing specific instruments to shift consumption to products with the least impacts on people and the planet. The shifts in behavior may be direct, at the point of purchase, or indirect, back at the point of manufacture, for example the publication of a plant’s emissions of all the hazardous substances in the Toxic Release Inventory listings. The theory in this case is that the publication of such data, a putative measure of a company’s environmental performance, will motivate managers to voluntarily install measures to reduce the numbers and the potential for some kind of stigma. Two of my former graduate students, Andy King and Mike Lenox, found only a weak connection. They argued that, if some form of reductions were needed or desirable, voluntary efforts based on publication of performance data are uncertain and variable. More official forms of sanctions (standards and penalties) would be more effective. Other work done at my former research group at MIT showed similar results for other self-regulatory (free market) approaches.

GoodGuide is based on the same theory; knowledge of good and bad characteristics of any item will guide the purchaser to the right or rational choice. Rational, here, means some balance of the properties sought in the item and the undesirable content indicated by the information provided. The values of the agent, the actor standing in front of the crowded shelves of products in a market, frame the balance, again, in theory. Here the information is presumed to have an immediate and direct influence on the purchasing decision. More about this in a moment.

The developer of GoodGuide, Dara O’Rourke, also believes the information will have a similar indirect effect.

“There definitely is a growing percentage of consumers who are aware and who care and who are seeking out products that have better environmental, social, and health attributes,” O’Rourke says. “We view those consumers who care as point of leverage over these big, big systems.” These “conscious consumers,” as they’re sometimes called, are important to the work of activist groups who bring pressure on corporations to reform their environmental or social practices; companies feel compelled to respond because they don’t want to alienate even a small share of their customers or potential customers.

The direct effect of GoodGuide is produced by the information contained in an index with values from 1 to 10 describing the relative “goodness” of the product on that scale. The website advertises its index, thus enabling purchasers to “Find safe, healthy, green & ethical products based on scientific ratings.” I have argued in this blog that there are several fundamental problems in this approach. (I need to disclose I am a former colleague of Dara O’Rourke.) The index is a composite of values in three categories: health, environment, and society. In the default setting, the three are weighted equally in calculating the single composite index. The system to compute the score now allows the user to select the categories most valued and override the default set. These items can be selected: Nutrition; Organic (Food); Scientifically proven hazards; Controversial ingredients; Fragrance-free; Animal Welfare Certifications; Energy Efficiency; Recycled Materials; Climate Change; Pollution; Resource Conservation; Fair Trade; Labor & Human Rights. (I could not find the details of the methodology on the website and may not be up to date here.)

I will freely admit that the ratings probably do guide purchases toward healthier products, but with far less accuracy than the index, which reports scores to one decimal place, implying far more precision than the system is capable of. As others and I have written, widely divergent scores, say 4.5 vs 9.3, illustrate significant differences and should lead to well informed decisions. Decisions based on two products with scores, say, of 8.7 and 9.3, are probably not justified, given the precision of the elements of the index.

My reservations about the effectiveness of this and similar systems, such as the one that the Sustainability Consortium, started by Walmart, is developing come from three sources: questions about the rationality of the agent, the disconnection of single transactions from the whole system; and the so-called moral hazard inherent treating the symptoms so as to ignore the roots.

First, rationality. It is becoming more and more apparent that agents do not act in a classic rational manner. We are creatures of habit and prejudice. We do not in most relatively familiar circumstances, like shopping, base our actions on some sort of maximizing or optimizing calculus. Current cognitive scientific research has produced much evidence to this end. Evolutionary ontogeny models describing the development of the cognitive system of individual humans, increasingly demonstrate that the neuronal networks of our brains from which action is directed are built through the experience of living itself. As, Humberto Maturana, whom I so often quote, says, “All doing is knowing, and all knowing is doing.

Our actions are largely conditioned by what patterns are already embedded and are in charge. The contribution of more information is limited. Maturana and his co-worker Francisco Varela argue against the conventional rational model connecting “knowledge” and action. Here knowledge refers to inputs such as the GoodGuide index. The role of prior acts, embedded in memory and influenced by other past inputs such as advertising, may overcome what is arguably the rational decision. If this science is correct, the actual functioning of this index and all others like it is not so clear. It is only a single input to a cognitive system that has already retained many other memories of past actions that may be dominant in determining the immediate act.

Next, the disconnect between single act and the systemic result. The health-related part of the index refers to potential health impacts of the user. These have a more or less one-to-one correspondence. Less harmful or more beneficial ingredients should, in theory, produce changes for the better. Promising safer products is not so direct. Safe is a relative standard, whose magnitude generally is context-sensitive. So advertising the index as producing safety is almost as misleading as the claims made by the companies GoodGuide hopes to expose as bad actors. The same concern applies to the claim of "ethical" made on the GoodGuide homepage. What is ethical cannot be collapsed into the factors used to generate the index.

Green and society are very different targets. Here the objective of GoodGuide is to improve the quality of the environment and of society as a system. Both are complex systems and do not respond linearly to changes in flows of materials entering or leaving, in the case of green, or labor policy, for example, in the case of society. It is extraordinarily unlikely that any buyer understands this connection, or, better, lack of connection. And this leads me to the third issue, the moral hazard embedded in this whole approach.

People using this system to make choices may and do come to believe that they are doing their part in combating unsustainability and increasing sustainability as a result of advertising and public relation efforts of GoodGuide and others like it, and of the companies whose products are listed. They may let lapse or ignore other activities essential to creating the conditions enabling sustainability to emerge. They may ignore the fact that their individual efforts, when buying detergent, can be overwhelmed by growth of the whole economy. Systems dynamicists have a phrase, shifting-the-burden, to describe this very common pattern. By continually focusing on the symptoms of a problem, attention is diverted from attacking the roots of the problem. In buying the safer product to protect one’s children, as was the case that motivated O’Rourke to create GoodGuide, parents may stray from asking questions about the need for the product, the nature of the economy, the predominance of relying on technology, or the indices that we use to measure our well-being.

I will be exaggerating here, but reliance on a single index, GDP, to guide our political economical actions is no different than relying on a single index to produce a healthy, green and socially just world. When the index is used for more than it can represent, it can and does lead to trouble. We know now how the use of GNP masks and distorts the condition of the social world. I have lumped the three categories, used in GoodGuide, and more into a single emergent systemic quality, flourishing. as the one we want the Planetary system to produce. When we learn, if we can, the rules by which this Planetary system works, then, and only then, will we be able to create a set of corresponding indicators that might (note the uncertainty here, indicative of the nature of complexity) be used to guide decisions. In the meantime, it is very important not to be lulled by the apparent precision and significance of indices that are said to have a direct relationship to a complex system. Use them. They have value, but don’t bet on them to turn the world into a green paradise.

|