The Economics Of Green Consumerism (1993)

THE ECONOMICS OF GREEN CONSUMERISM Alberto Cottica University College London

The author gratefully acknowledges the vital contribution of Professor David Ulph to bringing order to this paper. We wish to thank Raffaele Miniaci and Giovanni Ponti for the long, inspiring late-night discussions held together, and Francois Leveque for his encouraging and helpful comments. The usual disclaimers apply. This paper was funded by the Fondazione Luigi Einaudi. The Fondazione also took care of publishing it, in the Italian translation. The reference is A. Cottica, 1993, L’economia del consumismo verde, Annali della Fondazione Einaudi, Torino, Vol. XXVII.

INTRODUCTION CHAPTER 1 1.1. Chapter contents 1.2. Environmental friendliness as product quality 1.3. Quality Assessment Modes 1.4. Trust Goods in Consumer Theory CHAPTER 2 2.1. Chapter Contents 2.2. Repeat Purchase, Quality Premia and Market Structure 2.3. Signalling Quality: Brand and Green Reputation 2.4. Reputation and Market Structure CHAPTER 3 3.1. Chapter Contents 3.2. The Firm 3.3. Green Watch 3.4. The Model. 3.4. The model in steady state Appendix: a Fixed Costs Model

2

CHAPTER 4 4.1. Chapter Contents 4.2. Environmental Policy by Circulating Information 4.3. Information-Based Environmental Policy Instruments: A Proposed Taxonomy 4.4. Green Consumerism 4.5. Voluntary Agreements 4.6. Eco-Labeling 4.7. Possible Consequences on Foreign Trade CHAPTER 5 5.1. Concluding Remarks 5.2. Directions for Further Research BIBLIOGRAPHY

3

Introduction In 1992, the British branch of the environmentalist group Friends of the Earth waged a war on mahogany import. Mahogany, of which the UK. is the largest world importer, comes generally from the Amazon rainforest; its trade has very serious consequences over the Amazonian environment and the human rights of the forest peoples. The persuasion technique adopted by FOE was to organize demonstrations in the parking lots of the key showrooms of the four main British furniture retail chains, trying to hammer into consumers the idea that they, by buying mahogany furniture, are helping destroy a unique ecosystem. This won the adoption of company policy to sell only "sustainable" mahogany from three out of the four chains, and their commitment to lobby the Government for a legal ban of unsustainably managed mahogany. The latter point is easily understood. As a FOE spokesman pointed out, such a ban keeps high profile companies, that are being targeted by environmentalists, from losing sales to smaller businesses that their small size saves from being watched as closely, and can therefore sell mahogany furniture at a lower price. On the 16th of January 1993, activists of the Italian environmentalist group Legambiente marched to the biggest supermarkets in their cities. They bought what they needed, regularly paid, and dumped on the supermarkets themselves empty boxes and bottles, residuals of earlier purchases. The aim of Legambiente is to get the Italian parliament to adopt an anti-packaging waste law, similar to the German Packaging Ordinance. This law sets regulatory targets for packaging recycling rates and leaves to industry and commerce the choice (and the financial and organizational burden) of how to reach them. In October 1992, the Belgian detergent manufacturer Ecover opened, in front of the former Environment Commissioner Carlo Ripa di Meana (at the time he was the Italian Environment minister), its new ultra-low environmental impact plant in Antwerp. It is a model plant: the production process requires here only one sixth of the unit energy input of its competitors, waste water is recycled after treatment through a reed meadow. There is no air emission. Even the plant iself is built wit natural materials; pinewood for the structures, bricks (made from recycled saw-dust and mining waste) for the walls, linoleum and baked earth tiles for the floors, a grass roof for energy-saving insulation. Staff get a travel allowance that is highest if they cycle to work, lower if they car-pool, lowest if they drive in their own car. Ecover is a small firm (the new plant employs only 45 workers in an industry dominated by corporate giants such as Unilever, Henkel, Procter&Gamble. These three stories have a denominator in common, and that is the visibility of the firms involved. Environmentalists, both British and Italian, could never demonstrate calling for action on environmental problems in front of a small shop, run by a family enterprise. They need counterparts that are big and powerful enough to infuence, with their behaviour, the overall picture, and to take some of the blame for the existing problems. As for Ecover, its seems the story of a firm trying to become even more visible to "its" consumers; by building a model plant, and inviting environmentalist groups to visit it, it seems to be offering a sort of warrant as to its determination in tackling environmental problems.

4

These reflections suggest that information plays a crucial role in the firms' choice of their behaviour with respect to the enviroment. This conclusion is far from surprising, and traces of it can be found in the mainstream literature on environmental economics. Pearce et.al. [1989], for example, write that "for ecological consumerism to be effective, consumers must be informed as to the polluting potential of the products they buy". This intuition, however has not been developed: consumers as autonomous and capable of taking initiatives social actors are totally absent from environmental economics textbooks. The same is true of firms' strategies to capture their demand. Also, it is not totally clear what kind of information consumers really lack; Pearce et.al. seem to have in mind scientific knowledge, whereas the three stories just told deal with information pertaining more to the physical characteristics of the goods manufactured ond of their production processes. Developed by public economists and fiscal scientists, the theory of environmental policy has developed as the theory of the internalization, via fiscal policy, of the negative externalities (i.e. pollution) generated by economic activities [Baumol and Oates, 1989]. This approach has been recognized to be vulnerable to organizational failure objections [Veljanowsky, 1985]; despite this, it continues to dominate the field. The aim here is to suggest an approach based on the existence of information asymmetries on product markets. These asymmetries prevent consumers to effectively express their preferences for the "pollution content" of products in the same way they do for their quality. The behaviour of firms, environmentalists and "green consumers" in the face of information asymmetries, and the consequences of such behaviour in terms of policy, are the subject of this dissertation.

5

Chapter 1 Making Green Consumerism Work 1.1. Chapter contents This chapter looks at the information asymmetries hindering consumers' preferences for environmental friendliness to being exerted on the market place, and explores ways to get round them. Section 2 establishes a comparison between these asymmetries and those regarding product quality in markets for differentiated products. Section 3 introduces the notion of quality assessment modes, and suggests that of "trust" for environmental friendliness. Section 4 sets the notions of quality assessment mode and trust good in the literature on consumer theory. 1.2. Environmental friendliness as product quality The notion of environmental friendliness seems to bear a striking resemblance to that, familiar to industrial economists, of product quality. Both are quite compatible to the idea of a horizontal dimension in the product space (some consumers may prefer durable cars to fast cars; others may be much more concerned about the extinction of whales than about the ozone layer), but suggest a strong verticality (there are longer-lasting and less longlasting cars, and CFCs-free hairsprays do less damage to the ozone layer than non-CFCsfree ones). In other words, both these concepts imply that products can be ranked from "best" to "worst", and that this ranking is, at least in part, objective. This reflection casts a different light on the view, expressed by Pearce et.al. [1989] and apparently quite widespread, that green consumerism can't work because consumers lack information on the environmental friendliness of products. In fact, consumers lack information about product quality as well, which does not always prevent some industries from displaying a keen competition on quality grounds. Those industries have found a way round the information asymmetries to exploit the consumers' willingness to pay for highquality product. Enlightened politicians and the less anti-capitalist wing of the environmentalist movement have long been putting forward the vision of a greener, kinder capitalism whereby firms behave respectfully towards the environment. Such respect would stem from the environment, just like product quality, having become a ground for competition. Can the theory of information asymmetries on product quality provide an economic rationale for what has so far largely been wishful thinking? This chapter suggests it can. However, to do so it is necessary to tackle the problem of how consumers can assess environmental friendliness. To begin with, it is useful to distinguish between two different kinds of information that consumers need to carry out such an assessment. One is the scientific knowledge needed to understand the relationship between productive activities and the environment; the cost of acquiring this kind of information is, as Pearce et.al. point out, very high for the average

consumer. The other is what kind of productive activities each product needs to be manufactured: for example, that batteries contain or not mercury, or that bottles made of recycled plastic keep some waste out of landfills. The cost of acquiring information of this second kind is obviously very much lower than that of acquiring information of the first kind, although still (as will be explained) quite high for the individual consumer. The question then arises of what role is played by each of the two kinds in carrying out the assessment of environmental friendliness. Consider a similar problem, that of consumption of cholesterol-free food. It does not seem reasonable that all consumers choosing to eat cholesterol-free food are aware of the biochemistry of cholesterol, and of the complex ways in which diets differing in cholesterol content affect, over time and according to a stochastic relationship, human health. In a sense, they do not need to know: a public awareness has formed, in complex ways well worth investigating by sociologists, that cholesterol is best avoided. All consumers need to know in order to do so is the cholesterol content of food products. Similar processes seem to be at work behind green consumerism. The extraordinary commercial success of phosphate-free detergents is not easily ascribed to consumers having a thorough understanding of the effects of phosphates on aquatic environment. It certainly seems more likely that consumers accept a "popular wisdom" that phosphates are "bad" for the environment, and try to buy detergents that do not contain them if they can do so at a reasonable cost. The ways in which such beliefs form are not investigated in this work. The rest of this chapter is devoted to exploring the ways in which consumers may acquire information of the second kind. 1.3. Quality Assessment Modes Nelson [1970] has divided goods in search and experience goods. The former are products the quality of which can be assessed prior to purchase, like, say, clothes; the latter are those the quality of which can only be assessed by buying them, like meals at restaurants or electric appliances. Search and experience are, in this context, alternative quality assessment modes. Nelson runs a number of simple empirical tests the result of which suggest that information is passed on to the consumers in ways aimed at bridging the informational gap between himself and the purchase decision. So, advertising search goods tends to involve the transmission of "hard" knowledge: a picture of a dress tells consumers a lot of what they want to know about the dress. On the other hand, canned tuna is often advertised with marine scenes; anyway, a picture of the can is not particularly informative as to the taste of its content. The evidence suggests that consumer behaviour is significantly different across quality assessment modes. How is environmental friendliness assessed? Certainly not by search. Very little can be ascertained about it prior to purchase, by simply looking at the good. But not by experience, either: evaluating the environmental impact of a good involves, even taking as given a system of beliefs about how dangerous certain productive activities are, at least some knowledge of the product content in dangerous substances (e.g. CFCs in hairsprays), of the production process (e.g. in the case of organic food products) with 7

which it was manufactured, and of the method by means of which it will be disposed of when it becomes waste (it is less risky to buy batteries when a separate collection method for waste batteries exists). The cost of acquiring this information is quite high for individual consumers. One of the main purposes of this study is to suggest that, despite these information asymmetries, green consumerism can be made to work if the assessment of the environmental friendliness of goods is carried out by some better informed agent whom consumers trust. Environmentalist groups are an obvious candidate for the job, and there is some evidence that they are beginning to carry it out. Trust is, then, a third quality assessment mode. To preserve symmetry with Nelson's terminology, we propose to call trust goods those goods the assessment of whose quality takes place mainly by trust. 1.4. Trust Goods in Consumer Theory The purpose of this section is to clarify the notion of trust good and its relationship with consumer theory in the literature on industrial organization. Two issues seem to be most relevant; the "goods vs. characteristics" dilemma and the comparison between "trust goods" and "credence goods". Lancaster [1966] has suggested that consumers are not interested in goods at all. They see goods as bundles of characteristics, which are the real utility yielders. In other words, consumers choose their consumption bundles in order to obtain certain characteristics: a well-known example is that two light bulbs with a durability of six months each yield the same utility as one light bulb with a durability of one year. Consumers desire hours of lighting, not light bulbs in themselves. The general equilibrium tradition sees things from a different point of view: utility functions are entered by goods, which may differ from each other as subtly as necessary to explain consumer behaviour. A characteristics approach seems to have important advantages over a goods one when looking at quality assessment modes. The rationale for this is that, as a rule, the quality assessment of any one good takes place through a combination of modes [Tirole, 1988]. For example, the design of a car can be assessed by search; its durability, only by experience; the environmental friendliness of its production process, by trust. Being this the case, it is quite natural to start thinking in terms of search, experience and trust characteristics. One can think of several other trust characteristics besides environmental friendliness; "Americanness" under President Nixon's "buy American" campaign, or "fairness of labour use", when consumers do not wish to buy products manufactured exploiting, say, children's work, or "kasherness" of food products for orthodox Jews. It is worth noticing that consumers seem capable of providing themselves with the means of assessing other trust characteristics besides environmental friendliness; for example, the obvious candidates to carrying out the assessment of the fairness of labour use of a product are trade unions1 . This suggests that the notion of trust as a quality assessment mode, and the economic theory derived from it, may find a wider field of application than environmental economics; that of business ethics. Reasoning in terms of characteristics rather than goods has, however, an all-important disadvantage: it makes it much more difficult to do empirical research. The approach 8

suggested by Lancaster [1979] is to employ hedonic pricing procedures to extract the demand curves of characteristics from those of goods; the existence of indivisibilities in consumption2 introduces nonlinearities and further complications. In the original Nelson article the problem was solved by applying the notion of main quality assessment mode: when buying tuna fish, consumers are basically interested in its taste, an experience characteristic; on the other hand, jewels are largely bought for their design, a search characteristic. This allowed Nelson to term tuna and jewels, respectively, experience and search goods, and therefore to run simple tests on consumer behaviour when buying the one and the other. Since the durability of a car is not bought separately from its style, it would have been difficult to run those tests in terms of pure characteristics. Much of the literature this study draws from thinks in terms of goods, and the term "trust goods" will henceforth be employed here. The models we present, however, lend themselves to be discussed in terms of characteristics as well. It remains to be discussed the relationship between the notion of trust goods and that of credence goods, introduced by Darby and Karni [1973] shortly after the appearance of Nelson's article. The authors introduce this notion in the context of the "repair problem", i.e. the joint provision of diagnosis and services. How can the advantages of, say, the removal of an appendix be evaluated? The average consumer simply can't do it; on the other hand, his doctor may have (indeed, Darby and Karni contend that he has) an incentive to overstate the need for the operation to be done. Consumers have the choice of buying diagnosis from one doctor and their operation from another one, but this often involves additional monetary, time and transaction costs that must be compared with any expected reduction in fraud. An additional complication is introduced by the existence of a stochastic relationship between repair services and the flow of services from sophisticated durable goods, such as cars, or from human beings. The doctor may advise me to remove my appendix because, although I have no symptoms now, I am likely to get problems in the future. If I follow his advice, his diagnosis is unfalsifiable; even if I don't, and I turn out not to have any problems, I cannot judge his performance as a diagnosis maker: I may just be one of a small minority whose appendices resist deterioration under certain clinical conditions. The conclusions reached by Darby and Karni along this line of reasoning are that the markets for credence goods display an unusually high optimal amount of fraud. Since the government is subject to the very same informational limitations that affect consumers, one cannot make a convincing case from government intervention. It seems reasonable to distinguish credence goods from trust goods as defined above. , Credence goods are those goods that are actually never, or very imperfectly, assessed.3 Trust goods are those goods a fairly accurate assessment of which is carried out by a third agent; the results of this assessment process are then made available, at a reasonable (and often nonmonetary) cost to consumers. The reason why informational costs are low to single consumers is that they trust the assessing agent, so they don't need to go into the details of the assessment process as Darby and Karni's consumers would.

9

Chapter 2 Information Asymmetries And Market Structure 2.1. Chapter Contents This chapter is concerned with the industrial organization side of the theory of information asymmetries on consumer markets. It revolves around the notion of brand reputation as a signaling device for the persistence of quality. Section 2 provides an overview of the theory of competition through capital investment in brand name promotion ("reputation"). Section 3 defines the notion of reputation and outlines a role for it in the markets for trust goods; "green reputation" is introduced. Section 4 explores the relationship between competition through reputation and market structure. 2.2. Repeat Purchase, Quality Premia and Market Structure Information asymmetries imply moral hazard. If quality can not be assessed prior to purchase, manufacturers who sell experience goods have an incentive to undersupply quality; this incentive is strongest where consumption is most transient or one-shot, like for services directed to tourists or life insurance contracts. Industrial economists have found repeat purchase to be an effective mechanism to reduce moral hazard problems on experience goods markets. The idea is here that a deceptive firm can be punished by turning one's demand to its competitors if one is deceived. This point was developed by Klein and Leffler [1981]. Their model is built as follows. Consumers buy one unit of an experience good each period. They communicate costlessly and immediately; so, if one firm supplies less-than-contracted for quality to one consumer at time t, by time t+1 it has become a notorious cheater and loses all of its sales. The market is assumed to be perfectly competitive.There are no sunk costs. Two prices are defined for the good, corresponding to basic and high quality. Assume, for the moment, that the low and the high price are both perfect competition prices, although related to different cost curves. Since quality can only be assessed after purchase, the moral hazard problem for firms becomes a choice between three options. •

the first one is to sell the low quality product at the low price. By assumption, no profit is earned.

•

the second one is to sell the high quality product at the high price. Again, no profit is earned.

•

the third option is to cheat, sell the low quality product at the high price for one period, and exit immediately after. This third option does yield a positive profit (a one-period quasi rent), and therefore will be the one chosen by a rational firm. Again, moral hazard leads to the undersupply of quality.

Suppose, however, that there is a price, higher than the competitive high quality price, that, if it were the equilibrium price for high quality: •

would provide a stream of profit the present value of which is higher than that of the cheating one-period quasi rent;

•

would not dissipate the entire consumer surplus.

If such a price exists, it can characterize an equilibrium. Klein and Leffler contend that, under very general cost conditions, this should be the case. Notice that a rational consumer would not buy the high quality product at a price equal to its average cost; he knows that firms have no incentive to sell it at that price. The higher price, consisting of average cost plus a quality premium, has then the property of being a quality guaranteeing one. In the long run, the presence of extra profits to be earned triggers entry, and therefore increased competition. However, competition cannot occur on price: a reduction of price below the premium price would result in zero sales. By the assumption of perfect competition, the long-run equilibrium must be a zero-profit one. This implies a rentdissipating kind of competition; Klein and Leffler suggest that it occurs through sunk investments in firm-specific capital assets, such as the promotion of the firm's brand name. The competitive process will direct such investments towards those forms that provide the greatest direct service value to consumers. The "real" price of purchase of the experience good, i.e. the premium price net of the services yielded by the jointly supplied brand name assets, is thereby minimized. Net investment of this kind will go on until the average cost reaches the premium price. Notice that sunk costs are firm-specific, and therefore non-salvageable. This guarantees the unprofitability of the "cheat and exit" strategy described above. Bearing this in mind, the reason why the promotion of a firm's brand name yields services to consumers is fairly intuitive: it provides them with information about the firm's commitment to staying in the market and therefore being vulnerable to consumers' retaliation in the repeat purchase game. So, when consumers do not know the minimum quality-guaranteeing price, the magnitude of a firm's brand name capital investment, relative to its sales, provides a useful proxy for its motivation to produce the high quality product4 . 2.3. Signalling Quality: Brand and Green Reputation Klein and Leffler's article provides an elegant rationalization to an idea economists had had for at least twenty years: that brand reputations are information devices, used by firms to send signals to consumers5 . The purpose of this section is to clarify the notion of reputation, and to investigate the role it can play in the markets for trust goods. A good starting point is the following: single product firms, like those of Klein and Leffler's model, are rarely observed. Why, then, should reputation be attached to a brand name and not to a product name? The obvious answer is that the brand name has the advantage of synthesis: by denoting all products of a brand, it is a less costly way to convey information about product quality. This, however, has a far-reaching implication: for brand reputation to be informative about product quality, all products of the same brand must be of the same quality. This must be true both across a range of products at a given moment in time, and over time6 . 11

The role of time in reputation deserves particular attention. Klein and Leffler's static framework should not conceal that, in a repeat purchase game, the interval between purchases may be long enough to alter completely the range of products consumers can buy. Cars are a good example: only very rarely do consumers, when they sell their old car, buy a new specimen of the same model (it may have gone out of production to begin with). They tend rather to be loyal to the brand. For reputation to be effective as a quality signaling device in such a context, past quality must imply present quality7 . If this is true, it follows that reputations are slow to build. Spending ten million pounds today in brand name promotion is just not as effective as having spent half a million pounds a year over the last twenty years. Symmetrically, they are slow to destroy: once established, brand loyalty tends to persist even when the quality of single products has declined [Paba, 1989]. For these reasons, it is useful to think of reputation as a firm specific, intangible capital asset. In the markets for trust goods repeat purchase plays a different role, in that purchase does not imply, in itself, quality assessment. The latter is carried out independently of purchase, by the more informed agent (in the case of environmental friendliness, environmentalist groups); the repeat purchase mechanism provides consumers with the means of punishing manufacturers if the results of such an assessment are negative. The Klein-Leffler story can easily be told with regard to environmental friendliness. The long-run equilibrium involves sunk capital investments in promoting the "greenness" of the brand name, and this process generates, by the dynamics described above, a brand reputation for environmental friendliness: a green reputation, as we shall henceforth call it. 2.4. Reputation and Market Structure Paba [1986, 1989, 1991] explicitly links reputation effects to market structure. He contends that markets where information asymmetries are very important will generally be characterized by •

a high degree of concentration; consumers can only remember and compare with each other a limited number of brands. Reputation requires visibility.

•

stable market shares by brand (although not necessarily by firm: concentration can increase by takeovers);

•

long-run efficiency.

The two last results are due to the role of time in building and destroying brand reputations. This implies that a firm, threatened by a successful new product launched by a competitor, has time to respond, either by imitating or by developing an even better product. A keen (nonprice) competition is then compatible with the stability of market shares by brand. By this token, an oligopolist with a very good patent is prevented from becoming a monopolist. Empirical evidence from durable consumer goods' industries seems to fit very well this kind of model. It seems reasonable to apply these same market structure features to markets where information asymmetries of the "trust" kind are particularly important. However, it should 12

be noted that a notoriously good environmental performance can also be used as a way to overcome barriers to entry, thereby reducing concentration. This seems to have been the case with the detergent industry; with cost functions dominated by scale economies, new entry by "green" companies such as Ecover could take place at a profit in the early eighties only because of the existence of a quality premium on phosphate-free detergents. However, in the long run, it is reasonable to expect a further concentration process, the extent of which will depend on the share of consumers who will switch to "green" detergents.

13

Chapter 3 Green Reputation : A Simple Model 3.1. Chapter Contents This chapter presents a simple optimal control model of competition on environmental friendliness in a stationary environment. It is built on the idea that green reputation can be viewed as a capital asset. It is characterized by two parameters: its monetary value and its capability of generating sales. A firm's cleanup expenditure decision is the result of the maximisation, over time, of a value function comprising cash-flow profit and the capital gains from the change in value of the firm's green reputation. Section 2 introduces the firm's demand and its cost function. Section 3 sets up a modelling device to allow for environmental friendliness assessment. Section 4 computes and discusses first-order conditions. Section 5 computes and discusses the steady state equilibrium. Finally, an appendix presents briefly a different version of the model. 3.2. The Firm Consider a firm competing on a market where environmental friendliness is the main ground on which competition takes place. The manufactured products have a negative environmental impact only during the production phase of the life-cycle: it consists of emissions discharged into the environment. Pollution can be reduced by applying filters to the production line.The following assumptions are made: •

sales at time t+1 are a function of green reputation at time t alone. Therefore, for each period, they are given.

•

price p, the quality-guaranteeing price, is also given.

•

production costs are normalized to zero. The only costs are those of increasing the environmental friendliness of the production: they consist in current expenditure on abatement devices ("leasing filters").

•

the firm acts so as to maximize the present value of its stream of cash-flow profits plus the capital value of its assets, i.e. its green reputation.

Strategic interaction is ignored: each firm makes its own decisions in isolation. This is not consistent with Paba's emphasis on oligopolistic interaction, and investments in quality manufacturing and signalling through brand name promotion decided in response to competitor's moves8 .

15 3.3. Green Watch Assume there exist an environmentalist group, Green Watch, that has among its aims that of discovering firm's frauds on environmental friendliness. Green Watch behaves at follows: •

it checks a number of firms randomly chosen, each period. The probability for each firm to be checked during any one period, , is common knowledge.

•

the probability for any one firm to be checked in period t is independent of whether it was checked in period t-1 or not9 .

•

it evaluates the amount of their unit expenditure on environmental protection EP; on those grounds, it gives firms a "mark". Marks are given according to a function GW of EP, the shape of which is known, and are expressed in the same unit of measure as green reputation itself10 . The function GW is, realistically, assumed to display decreasing returns. GW is therefore expected to take a shape of the following kind:

FIGURE 1: The GW function

the assessment process affects the firm insofar it increases or decreases the value of its green reputation, and by that token the value of future sales. It behaves as an independent evaluation of the firm's green reputation: at the end of each period, if the green reputation has been found overestimated, a share 1− of its customers will accept Green Watch's evaluation, and consequently reduce their demand for the firm's product. The reason why the firm does not lose all its customers has been clarified in chapter two: green •

16 reputation commands brand loyalty. , the share of consumers who do not alter their opinion of the firm's environmental friendliness, can be interpreted as the degree of brand loyalty. 3.4. The Model. We are now in a position to write the three basic equations of the model. [1]

Π t = p ⋅ S (Rt )− Ct

Where Π denotes profit, S sales, p the quality-signalling price and C costs. [2]

Ct = pe ⋅ EPt ⋅ St

Where EP (for environmental protection) denotes the quantity of "leased filters" per unit produced and pe the lease price. Finally, [3]

Rt +1 =

Rt + (1 − )⋅GW (EP) +(1 − )Rt

where, as we have seen, is the probability of a Green Watch inspection in any given period, brand loyalty as defined above, GW the evaluation made by Green Watch as a function of EP. The first part of the RHS describes what happens if Green Watch does check the firm: the second one what happens if it doesn't. The maximization problem can be written as follows [Neher, 1990]: [4]

MAX V = ∑t = 0 Π t /( 1 + r) t + ∞

t +1

⋅ ∆Rt

Or, in its more familiar continuous-time form,

17

18 Clearly, EP acts as control variable and R as state variable. The expression for ∆R can be obtained by [3], subtracting Rt from each side: Rt +1 − Rt = ∆Rt +1 = [5]

Rt + (1 − )⋅GW (EP) +(1 − )Rt − Rt Rt +(1 − )⋅GW (EP) − Rt

∆Rt +1 = (1 − )⋅ GW (EP)− Rt

What follows deals with the continuous-time form, assuming continuity in R and piecewise continuity in EP. The current value Hamiltonian is [6]

H = Π + qRt

Where q ≡ e rt . Substituting [1], [2] and [5] into [6] yields: [7]

H =(p − pe ⋅ EP)⋅S (R )+ q (1 − )⋅ GW(EP)− R

FOCs can be derived from [7]. We start from the maximum principle (subscripts denote partial derivatives): H GW = − pe ⋅ S + q ⋅ (1 − )⋅ EP EP pe = q

[8]

⋅(1 − ) ⋅ GWEP S

Portfolio balance condition: H =− R

= rq − q

p − pe ⋅ EP ⋅S R − q ⋅(1 − )= rq − q [9]

q p − pe ⋅ EP ⋅S R = r + ⋅(1 − )− q q

Dynamic constraint: H =R q

19 [10]

⋅(1 − )⋅ GW (EP)− R = R

Equation [8] means that production will be "greened" up to the point where the marginal capital gain resulting from Green Watch's evaluation (filtered through brand loyalty and Green Watch's probability of checking this particular firm) equals the marginal cost. of doing so. pe/q represents the price of "leasing filters" in terms of reputation. This situation can be represented by the following diagram:

20 FIGURE 2: The maximum principle

ks in the diagram are equal to ⋅(1 − )/S , and k1
the lower the price of "leasing filters" in in relation to that of green reputation;

•

the higher the probability

•

the lower the degree

•

the smaller the firm is.

of being checked by Green Watch;

of brand loyalty;

The first three results are hardly surprising. The fourth one, which seems to contrast with the intuition that firms need to be visible to enjoy exploit their green reputation, and need to be large to be visible, is a byproduct of the specification of the model: Green Watch is interested in unit expenditure on environmental protection, whereas its contribution to building up green reputation is scale-independent. This means that the incentive to comply with Green Watch's standards increases as firm size decreases11 . This effect disappears if the EP on which Green Watch's evaluation is based is let to be a fixed cost rather than a variable one, as the appendix to this chapter demonstrates. Equation 9 has a structure similar to that of Hotelling's rule. It implies that, for portfolio balance, the rate of growth of the price of reputation must equal the rate of interest minus a

21 factor accounting for the sales- (and therefore, in our model, profit-) generating power of investments in green reputation. This factor, in its turn, is taken net of the factors limiting it, namely the uncertainty surrounding Green Watch's monitoring activity and brand loyalty. More analitycally, on an equilibrium path, q must grow faster: •

the higher the rate of interest r;

•

the lower the price-cost margin (p-pe. EP);

•

the lower the sales-generating power of green reputation SR;

•

the higher the probability

•

the lower the degree

of being checked by Green Watch;

of brand loyalty.

The analogy with Hotelling's rule can be carried one step further. r + (1 − ) can be thought of as a risk-adjusted rate of interest. In this interpretation, the (marginal) capital gain on reputation plus the marginal cash-flow profit gain deriving from a higher "stock" of green reputation must, for equilibrium, equal the risk-adjusted rate of interest. Finally, equation [10] simply recovers the dynamic constraint. 3.4. The model in steady state It is interesting to look at the long-run evolution of the system described by these equations. Recall first order conditions: pe S ( R) ⋅ (1 − ) GWEP

[8a]

q=

[9]

q p − pe ⋅ EP ⋅S R = r + ⋅(1 − )− q q

[10]

⋅ (1 − ) ⋅ GW( EP) − R = R

Equation [8a] is simply equation [8] solved by the price of reputation q. To depict the motions of the control variable, the environmental protection expenditure EP, and of the state variable, the green reputation R, begin by considering [10]. Firstly, it is straightforward to check that the steady state version of [10] has the same shape of the GW(EP) function: [10a] R°= 0

hence

GW(EP)= R

22 FIGURE 3: The R°=0 condition

Consider a point such as A in the graph. It is characterized by a higher R than the steady state requires; a glance at [10] shows that this implies R°<0. So, first order conditions require that R decrease when above the steady state level and vice versa. The motion of EP is less intuitive, and requires a little manipulation. Begin by differentiating [8a] with respect to time: [11]

q°=

pe S pe S ( R ) ⋅ GWEP′′ ⋅ R ⋅ R °− ⋅ ⋅ EP° 2 (1 − ) GWEP (1 − ) GWEP

From [11] and [9]: r + (1 − ) ⋅ q − p − pe ⋅ EP ⋅ S R =

pe S pe S ( R) ⋅ GWEP ′′ ⋅ R ⋅ R °− ⋅ ⋅ EP° 2 (1 − ) GWEP (1 − ) GWEP

Substitute for q from [8a]: r + (1 − ) ⋅

pe S ( R) pe S pe S ( R ) ⋅ GWEP ′′ ⋅ EP ° ⋅ − p − pe ⋅ EP ⋅S R = ⋅ R ⋅ R° − ⋅ 2 (1 − ) GWEP (1 − ) GWEP (1 − ) GWEP

Substitute for R° from [10] and simplify:

23 r + (1 − ) ⋅

pe S ( R) S pe S ( R) ⋅ GWEP ′′ ⋅ − p − pe ⋅ EP ⋅S R = pe ⋅ R ⋅ GW ( EP) − R − ⋅ ⋅ EP ° 2 (1 − ) GWEP GWEP (1 − ) GWEP

Multiply each term in the equation by r + (1 − ) ⋅

GWEP : SR

pe S ( R) pe S ( R ) ⋅GWEP ′′ ⋅ − p − pe ⋅ EP ⋅ GWEP = pe ⋅ GW ( EP) − R − ⋅ ⋅ ⋅ EP ° (1 − ) S R GWEP (1 − ) SR

Rearrange [12]

pe S (R ) ⋅GWEP pe S (R ) ′′ ⋅ ⋅ ⋅ EP°= pe ⋅ GW(EP)− R + p − pe ⋅ EP ⋅ GWEP − r + (1 − ) ⋅ ⋅ (1 − ) S R GWEP (1 − ) SR

[12] describes the motion over time of EP as a function of the absolute values of EP itself and R. Together with [10], it constitutes a system of two differential equations in two unknowns, EP° and R°, which, at each point in time, can be solved given the values of EP and and R and of the relevant parameters. At the expense of some loss of generality, it is possible to proceed further. Assume [13]

S ( R) = R

hence

SR = R

GW( EP) = EP1−

with

0 < <1

and

GWEP ′′ = − (1 − ) EP −

−1

and [14]

Hence GWEP = (1 − ) EP −

−1

Substitute in [12] to yield pe R pe R − ⋅ ⋅ ⋅ EP°= pe ⋅ EP1 − − R + p − pe ⋅ EP ⋅(1 − )⋅ EP− − r + (1 − ) ⋅ ⋅ EP (1 − ) (1 − )

Rearrange as • ⋅ EP °= pe ⋅ EP1 − +(1 − )⋅ p ⋅ EP− − pe ⋅ EP1− − r + (1 − ) ⋅

[15]

• ⋅ EP °= ⋅ pe ⋅ EP1− + (1 − )⋅ p ⋅ EP − − pe ⋅

Let

K ≡ pe ⋅

r + (1 − ) +1 (1 − )

pe ⋅ R − pe ⋅ R (1 − )

r + (1 − ) + 1⋅ R (1 − )

24 In steady state, since EP°=0, the LHS of [15] must also be zero. The steady state version of [15] is therefore reduced to [16]

K ⋅ R = ⋅ pe ⋅ EP1− + (1 − )⋅ p ⋅ EP −

To compute the slope of [16] take its total differential: K ⋅ dR = (1 − )⋅ ⋅ pe ⋅ EP− dEP − (1 − )⋅ ⋅ p ⋅ EP − −1 dEP K ⋅ dR = − ⋅(1 − )⋅ EP− [17]

−1

p − pe ⋅ EP dEP

dR (1 − )⋅ ⋅ EP − =− dEP K

−1

⋅ p − pe ⋅ EP

K is clearly positive; so is the numerator of the term multiplying the profit margin p − pe ⋅ EP . The sign of dR/dEP will then be the inverse of the sign of the profit margin term. dR < 0 ⇒ Π > 0 ⇒ EP < dEP dR = 0 ⇒ Π = 0 ⇒ EP = dEP dR > 0 ⇒ Π < 0 ⇒ EP > dEP

p pe p pe p pe

Of course, the last case does not make economic sense; there cannot exist a steady state in which the firm makes a loss. [16] must then be negatively sloped over the relevant range.

25 FIGURE 4: The EP°=0 condition

It is now possible to depict the motion of EP outside the steady state. The term between square brackets multiplying EP° in [15] is always negative; this implies that [15]

EP° > 0 ⇒ K ⋅ R − ⋅ pe ⋅ EP1− + (1 − )⋅ p ⋅ EP − > 0

Consider a point such as C in the diagram. It displays a lower R than the steady state would require. A glance at [17] shows that EP in C must grow to satisy first order conditions. The vice versa holds for a point such as D. Over the economically relevant range, the two curves [10a] and [16] will only intersect once, giving rise to a unique steady state equilibrium.

26 FIGURE 5: The steady state equilibrium and equilibrium paths of EP and R

The model predicts that, along "stable arms" equilibrium paths, firm starting from a low level of green reputation will begin the program with a high level of (unit) environmental protection expenditure. This will cause its green reputation to increase; as the equilibrium point is approached, the per period EP will decrease. The opposite happens when the firm starts off from a high level of green reputation. This will cause it to spend comparatively little in environmental protection, letting its green reputation lower to its equilibrium level. As the equilibrium level is approached, EP grows. Notice that Green Watch can always shift the steady state equilibrium point towards the right (towards an environmentally friendlier equilibrium) by "rising its standards", i.e. flattening and shifting to the right the R°=0 curve. However, this move is constrained by the requirement that the firm make a profit in the steady state. Appendix: a Fixed Costs Model An alternative way to think about environmental protection expenditure is as a fixed cost. Assume, for example, production is "greened" via research and development, which leads to an environment-saving innovation, like a new, easier to recycle plastic. The obvious way to think about environmental protection expenditure as a variable cost, as is done throughout chapter 3, is waste generation, which is linked to the amount of output produced. Equations in this version of the model will be marked with an f, standing for "fixed cost".

27 The cost function becomes: [2f]

Ct = pe ⋅ EPt t

So the current value Hamiltonian for the problem is [7f]

H = p ⋅ S(R )− pe ⋅ EP + q (1 − )⋅ GW(EP)− R

F.o.c.s differ in the maximum principle and in the portfolio balance condition, whereas the dynamic constraint is still equation [10] of the variable cost model. [8f]

pe = q

⋅(1 − )⋅GWEP

Equation [8f] does not display the scale diseconomy of equation [8]. [9f]

q p ⋅S R = r + ⋅(1 − )− q q

Equation [9f] is less restrictive than its variable cost counterpart, in the sense that it requires a slower growth of the price of reputation for equilibrium. This is because the marginal cash-flow profit of investment in reputation is higher in this model, as it amounts to the entire value of the marginal increase in sales. A steady state for this model can be characterized proceeding as in section 3.5, but is not as clear-cut as that of the variable cost model. Equation [12f] reads [12f]

−

pe ⋅GWEP pe 1 ′′ ⋅ ⋅ EP° = r + (1 − ) ⋅ ⋅ − p ⋅ GWEP (1 − ) GWEP ⋅ SR (1 − ) S R

Assuming the same functional forms for GW(EP) and S(R) as in [13] and [14], one gets [15f]

−

pe − 1 r + (1 − ) 1 ⋅ ⋅ ⋅ EP°= pe ⋅ − p(1 − )⋅EP − −1 (1 − ) EP R (1 − ) R −1

In the steady state [15f] reduces to [16f]

K′ ⋅ EP = R p(1 − )

Where K ′ ≡ pe

−1

r + (1 − ) (1 − )

The slope of [16f] is

28 [17f]

dR K ′ ⋅ ⋅ EP −1 = dEP p(1 − )⋅( − 1)⋅ R

−2

K' is clearly positive. In order for the RHS of [17f] to be negative, an additional restriction must be imposed that < 1.

Chapter 4 Information-Based Environmental Policy 4.1. Chapter Contents This chapter looks at the policy implication of the theory of green consumerism. Section 4.2 makes a case for government intervention to reduce the information asymmetries discussed in chapter 1. Section 4.3 proposes a range of "information-based" environmental policy instruments, classified according to the allocation of consumers' trust. Section 4.4, 4.5, and 4.6 discuss the policy instruments included in the range so far; they are, respectively, green consumerism, voluntary agreements and eco-labeling. Finally, section 4.7 discusses briefly the possibility that information-based environmental policy be used as a nontariff barrier to trade. 4.2. Environmental Policy by Circulating Information Environmental economics was born with the idea of internalizing environmental externalities using fiscal instruments [Pigou, 1938]. Although it has gone a long way from The Economics of Welfare, tackling more and more problems and developing new analytical tools, the state of the art in environmental policy seems to have kept faithful to this public finance matrix [Baumol and Oates, 1989]. Indeed, the prescription of "marketbased instruments", i.e. mechanism whereby the marginal external damage to the environment of human activities is equalled to the marginal cost of preventing it, seems to be the ubi consistam of the discipline12 . This approach to environmental policy has had a tremendous cultural impact on policy makers, and, where put into practice, has performed reasonably well, or even very well. However, one cannot help noticing that it rests on a simplistic, ultra-orthodox view of firm behaviour; firms are perfectly rational black boxes that equal marginal costs to marginal revenue. Consumers, by definition, are not involved in externalities until they are priced and passed on to consumer goods' prices; then they individually maximize their utility. Consumers as economic agents capable of taking collective action and possessing a certain degree of autonomy are simply absent from the model. Environmental economists can claim credit for raising a number of objections, but the response to these objections in terms of policy prescriptions has so far been quite weak. The main point of this chapter is that some of these objections can be dealt with by borrowing from more recent developments in firm and consumer theories. This will be argued by considering how well environmental policy based on the theory of green consumerism performs in the face of two of the main arguments against more traditional economic instruments; organizational failure and regulatory capture. Eco-taxes (à la Baumol and Oates, 1989) have been criticized on the basis that they are subject to organizational failure [Veljanowsky, 1983], especially if the marginal external 29

30 cost of the polluting activity happens to be low. If, within a firm, managers who pay taxes are physically and "organizationally" separated from engineers in charge of technical choice, the incentive to pollution abatement may get lost in the meanders of the firm's headquarters. More generally, in a bounded rationality framework a firm may simply not bother reducing its discharge of polluting substances into the environment if the charge they pay for it is low. The decision-making power of managers, itself a costly resource, may be better allocated to "strategic" activities. Regulatory capture is also a danger for the policy maker; on this ground, the green movement has often asserted its dissatisfaction about the introduction of eco-taxes. The danger consists in the bargaining process that precedes the introduction of any regulatory measure; industry can, and often does, blunt the edge of market-based instruments by lobbying and using occupational blackmail to lower the rate of eco-taxes. The talks about the introduction of a carbon tax at the 1992 Earth Summit make an illuminating example. Suppose, however, that the government, by a mix of measures including environmental education campaigns and giving easy access to the media to environmentalist, could raise the effectiveness of environmentalist groups in spotting and reporting environmentunfriendly firm behaviour. In terms of the model of chapter 3, this means •

increasing the opportunity cost of not investing in green reputation, by increasing the effect on sales of an increase in green reputation

•

increasing the probability of a check by Green Watch. The model predicts a consequent increase in the unit expenditure on environmental protection.

It is reasonable to expect firms with very high marginal costs of abatement to stay out of the "green" market; by this token, at least some degree of static efficiency is ensured. Dynamic efficiency follows from the arguments developed in chapter 2. This way of doing environmental policy enjoys two advantages over traditional marketbased instruments. Firstly, it seems reasonably robust to organizational failure objections. Environmental friendliness is here one of the grounds for competition; even in the presence of bounded rationality, it is plausible to expect firms to react to a change in consumer preferences towards it as they would if consumers changed, say, their taste for product design. Secondly, it seems less subject to regulatory capture than other available policy measures. Regulatory capture involves bargaining and a high degree of sharing the same values and cultural background between firms and regulators; what's more, it implies that, to some extent, the latter rely on the direct knowledge of the former to understand the problems they are trying to regulate. This is very much more likely to take place with government officials than with environmentalists13 . Of course, green consumerism as environmental policy also has two major disadvantages over eco-taxes. The first is that it makes it impossible (save by chance) to drive the economy to the optimal amount of pollution emission, as market-based instruments do. This may be a less serious limitation than it seems: environmental economists have known for a long time [Baumol and Oates, 1971] that it is virtually impossible to compute the position of the optimum. Eco-taxes are now being marketed as second-best efficient, leastcost policy instruments. 30

31 The second disadvantage is obvious: it only works for industries where the environment is, or can be made, an important differentiation factor. Notice that no mention is being made of what goes under the heading of "environmental education". The reason for this is that the understanding of environmental interdependencies belongs to the first of the two kinds of information described in section 1.2, with which this study is not concerned. However, it may be useful to keep some of the relevant issues in the background of what follows. Firstly, it should be noted that environmental education policies are being implemented. Italian schoolchildren in the majority of primary schools are taught that life forms are linked to each other in non obvious ways; they are taught how to plant trees and, in the most advanced cases, even that only indigenous species of trees should be planted in any given place. Most interestingly, an environmentally friendly detergent producer, Atlas, finances the education program of the Italian environmentalist group Legambiente [Cantoni, 1993]. Secondly, environmental education tends to shape preferences. Environmental awareness tends to be keener among the young than among their parents: exposure to the environmental issues from an early age seems to be one of the causes behind this pattern. It is not totally clear how the notion of optimality applies to a context of endogenously determined preferences: does teaching schoolchildren that extinction of species is "bad" lead to "too much" conservation? It may be so from the point of view of people who went to school before environmental education was introduced, but not from that of the schoolchildren themselves. Brusco [1993] has noted that teaching moral values means running the risk of a confessional state; for this reason, public schools should limit themselves to pointing out the inherent interdependency of species from one another. 4.3. Information-Based Environmental Policy Instruments: A Proposed Taxonomy Environmental policy by "making information circulate", then, appears to enjoy desirable properties. The purpose of this section is to discuss and evaluate the technical means by which information-based environmental policy, as we shall label it, can be done. We propose to distinguish information-based environmental policy instruments according to where the trust of the "green" consumers is allocated. By this criterion, it is possible to list three of them, each one characterized by a different allocation of consumers' trust. •

green consumerism itself; trust lies with firms

•

firm-government agreements; trust lies with the government

•

eco-labeling: trust lies with the eco-label awarding agency

Notice how environmental groups themselves are not mentioned among the possible allocations of trust. As is made clear by the model presented in chapter 3, they play a major role in the case of green consumerism: by providing an independent source of information about firms' environmental performances, they ultimately allow green reputations to be meaningful to consumers. Without Green Watch acting as a committment device, a firm's signaling "I am environment-friendly" would have no information content [Milgrom and Roberts, 1986]14 . The other two policy instruments can be implemented even without 31

32 strong and informed environmentalist groups at the price, as will increased risks of regulatory capture.

be discussed, of

4.4. Green Consumerism The desirable properties of green consumerism as the archetypal information-based policy tool have been discussed in section 4.2. We are here concerned with the technical means of putting it into practice. Government intervention on information asymmetries is by no means new. The consumers' movement, started off in the U.S.A. well before the Second World War, was actively encouraged in its development by in European governments in the late sixties [Shonfield, 1984]. Most active in this process were the Scandinavian countries and the United Kingdom. The measures put in place by those governments were chiefly aimed at reducing the time, transaction and monetary costs to consumers of claiming reimbursement for purchases whose quality was less than contracted for. Consumers' advisory offices were set up; Sweden, in a typically corporatist arrangement, formed a "consumers' tribunal" with no cohercitive power, but the recommendations of which no Swedish company would defy; Britain set up its Fair Trade Office, and for a while Small Claims Courts, much swifter and less costly to set in motion than the ordinary justice machine, functioned. All of these measures were inspired by the idea that information asymmetries are all to the firms' advantage; to counterbalance that, these tribunal-like structures should have a bias towards consumer protection. In the case of green consumerism, the availability of free or cheap laboratory facilities where environmentalists could bring sample of industrial effluents or waste to analyze would have, in countries with a reasonably strong green movement, an impressive impact. On the diffusion of information side, another possible policy measure is subsidizing green consumerist media. Tirole [1988] provides an argument for such an action15 . It is our belief that the history of the consumer movement has much to teach to governments interested in green consumerism. Further investigation into it is recommended. 4.5. Voluntary Agreements A possible path to building consumers' goodwill lies in asking governmental bodies to monitor one's environmental performance. Firms can sign agreements whereby they commit themselves to improving their environmental records, in return for the possibility to use the government as a guarantee of environmental friendliness. Such voluntary agreements (accordi di programma) are part of the Italian law, and are increasingly seen as a useful policy-making tool. The new waste disposal plan of the Lombardia region, for instance, includes a series of agreements with such major waste producers as hotel and supermarket chains, whereby the latter agree to source separate all their waste for recycling. These agreements are being much promoted on local media, and they are thought to help build up the involved companies' green reputation [Cantoni, 1993].

32

33 On a larger scale, in autumn 1991, the Italian car manufacturer Fiat signed another agreement with the Ministry of the Environment, committing itself to spending 23 billion lire (£10 million) on environmental improvements. The Italian government promised incentives to the use of catalytic converter equipped-cars, which were launched a few months later. This move by Fiat may be interpreted as an attempt to restore its environmental credibility by a company to which green activists are becoming a nuisance16 . It is very difficult to draw conclusions about the allocative efficiency of voluntary agreements. Static efficiency, which could stem from Coasian bargaining, is endangered by information asymmetries on firms' marginal abatement costs; once agreements are reached, if many firms are involved there is a risk of free riding. However, as has been noted, the bargaining process itself helps build a consensus between firms and the government, so that voluntary agreements end up "building trust" as a desirable side effect [Glachant, 1993]. Like green consumerism, such agreements are relatively robust to organizational failure objections; on the other hand, they are inherently subject to regulatory capture17 . 4.6. Eco-Labeling Eco-labeling involves an agency, often governmental, carrying out the environmental friendliness assessment and awarding a "green label" to the friendliest products. This policy tool, intuitively appealing, appears to be flawed by a number of practical problems. The progress of the European eco-labeling scheme, officially approved in December 1991, provides some insights about the nature of these problems. The first issue confronting eco-labeling schemes is that of defining the borders of product groups. It is a relevant one: often, products that are (imperfect) substitutes differ significantly in their degree of environmental friendliness. When carpets become waste, they leave behind synthetic fibres that must be burned, which is costly and can cause pollution; ceramic tiles, by contrast, are simply inert baked clay and pose only an aesthetic problem. Clearly, if flooring materials are grouped within an eco-labeling scheme, ecolabels will tend to be awarded to ceramic tiles and not to carpets. This will give rise to interindustry competition on environmental grounds. If, on the other hand, tiles and carpets are separated, the competition will be of the intra-industry kind: consumers will be able to tell, say, lead-free glazed tiles from ordinary tiles, but will have no information as to the environmental impacts of alternative flooring solutions. Assuming products can be sensibly grouped, the problem arises of selecting the criteria by which products can be assessed. The ideal instrument would be, of course, a cradle-tograve cost-benefit analysis of each product; since this is, obviously, a very costly solution, existing schemes have taken shortcuts, with some paradoxical results. Under the German "Blue Angel" scheme lawnmowers and vacuum cleaners, the only criterion for the awarding of the label was noisiness, which is obviously an unsatisfactory definition of environmental impact18 . Once found, the criteria have to be weighed against each other, and in the case of multiple criteria, a decision has to be made as to whether a product that falls short of the standard set for one criterion should be denied the eco-label, regardless of a better-than-average performance under the other criteria. This is known as the "hurdle vs. decathlon" problem.

33

34 Finally, a cutoff point has to be set to divide "green" products from other products. In the case of the European scheme, industry lobbying resulted in an unofficial 10-20% target of eco-labeled existing products; criteria will be set in such a way that 10 to 20% of existing products qualify for the eco-label. It can be reasonably concluded that eco-labeling suffers from an inherent difficulty to communicate in an exceedingly simple way (a logic variable like the presence or absence of a label on the package) to consumers complex information about the product's environmental impact. It also seems to be extremely vulnerable to regulatory capture19 . 4.7. Possible Consequences on Foreign Trade An unrestricted foreign trade and global environmental protection are, traditionally, conflicting policy objectives. Banning or taxing polluting production processes in one country may simply lead to their relocation to other countries, with less strict environmental policies. This move then only leads to a loss of competitiveness of the "environmentalist" country, with no reduction of the overall amount of pollution. Policy makers at the national level are then confronted with international prisoner's dilemmas [Gatsios and Seabright, 1989]; the political debate on many environmental issues, from the biodiversity convention to European Community carbon and energy tax, has been characterized by such a prisoner's dilemma structure. In principle, a tariff on environmentally unfriendly import could correct for the problem; but, as is well known, for the monitoring authority to distinguish between eco-tariffs and protectionist tariff may get very difficult if such a principle were to be accepted. In recognition of such problems, GATT does not permit tariffs on environmental grounds. So, when American environmentalists tried to ban Mexican dolphin unfriendly tuna fish, they were denied the right to do so. However, there is nothing stopping American environmentalists from boycotting Mexican tuna fish, or even lobbying their retail chains to cancel their contracts with Mexican fishing companies. A very similar operation was organized in 1992 against Norwegian products, to protest against the killing of 100 whales by Norwegian fishermen, exploiting loopholes in the International Whaling Convention. Campaigners organized a boycott of all Norwegian products, which resulted in the cancellation of several contracts. The sigle largest one was that between the American fast food restoration chain Burger King and the Norwegian fishing company Frioner, which used to supply Burger King its cod fingers. The move cost Frioner £5 million [Brown, 1993]. One hardly needs to point out that there is absolutely nothing GATT can do if consumers or retailers in one country refuse to buy foreign environmentally unfriendly products. In this sense, a well-led information based environmental policy may act as a nontariff barrier to trade, leading to a gain, rather than a loss, of competitiveness of the country in question. Similar forces seem to be at work in the case of the Packaging Ordinance in Germany. The German government introduced in 1991 the duty for retailers to withdraw the empty packages of the goods they sell and for manufacturers to recycle them or, alternatively, to set up a collection and recycling system of packaging materials. German industry has chosen the second option; collection, sorting and recycling costs are now pre-paid to a company called DSD (Duales System Deutschland); packages taking part in the circuit are 34

35 recognizable by a Green Dot printed on them. However, it is still possible for any one producer to stay out of the DSD circuit, as long as they agree to recycle their packages. The barrier to trade part of the story is that all German retail chains refuse to distribute products wihthout the Green Dot. A Green Dot must be applied for, and the package must have certain technical characteristics (i.e. must be easy to recycle) in order to obtain it. So, German controlled DSD regulates, an in important sense , access of foreign goods to the German market by setting standards that they must meet in order to compete [Southern, 1993]. One cannot help being reminded of the history of the powerful Deutsche Institut fuer Normung (DIN), which has admittedly been setting its standards in order to make the German market more difficult to access [Gatsios and Seabright, 1989].

35

Chapter 5 Conclusions And Indications For Further Research 5.1. Concluding Remarks The idea that information plays an important role in determining the environmental strategies of firms seems to be a promising one. If it is accepted that beliefs regarding the environmental friendliness of certain substances and production processes form among consumers, for green consumerism to work it is no longer necessary that each consumer can carry out an environmental impact analisys of each product he or she buys: it is enough that the content of the product and its production technology are known. Environmentalist groups are increasingly collecting and diffusing this kind of information in order to drive firms to an environmentally friendlier behaviour. In turn, this pattern of behaviour may bear consequences for market structure; a very important one is that firms with a reputation for environmental friendliness may enjoy an advantage over their competitors. This seems to imply at least the possibility for environmentally aware firms to be as profitable, if not more profitable, than their less careful competitors. To a certain extent, this intuition seems to lend itself to some kind of formalization. The model developed in chapter 3 captures some of the features of this process: greening production as an investment in "green reputation" is undoubtedly the most important of them. On the other hand, the model is wanting from many points of view: its most serious shortcoming is its inability to take oligopolistic interaction into account. If the theory of green consumerism is found to be useful to understand firm behaviour, and if the analitycal instruments to model it can be perfected, it follows that one of the ways to do environmental policy is by collecting and circulating information, and letting market forces drive firms to an environmentally friendlier behaviour. For a number of reasons that have to do with industry structure and firm visibility, it is unlikely that this model of intervention can be applied to all industries, but it does not seem unreasonable that it could play a role in some. 5.2. Directions for Further Research There is obviously a lot of work to be carried out if one wants the theory of green consumerism to develop into an useful tool for policy makers. However, the most urgent lines of research to add to it seem to four. The first one concerns the development of a model which takes into account oligopolistic competition. Paba [1989] makes his (unformalized, although consistent with his empirical data) prediction of long-run efficiency rest on market share competition between firms. The optimal control model presented in chapter 3 can be extended in this direction only at the expense of a sharp increase in its mathematical tractability. A more interesting approach, on which we have carried out some preliminary work, seems to be that of building a game in

36

37

which "green" firms compete for market share with one another, and at the same time the "green" product competes with its nongreen counterpart. Such a model would help understand under what conditions competition on environmental grounds is triggered; by contrast, the model in chapter 3 depicts a stationary environment, in which the green product is basically established. The second line of research concerns the refinement of the policy instruments range presented in chapter 4. A careful investigation would certainly bring to light more instruments, as well as implementation problems and informational requirements of the existing ones. A particularly promising approach seems to be that of reconstructing the history of consumer movements in the developed countries, and especially of the policy measures taken by European governments in the sixties to encourage their development. The third one concerns the "barrier to trade" features of information based environmental policy. If the reasons of international competitiveness can really be partially reconciled with those of environmental quality, new, politically acceptable ways of tackling old problems may be found. The strategies of environmentalist groups and their capability of building temporary alliances with firms, especially retail firms, should be studied carefully. Finally, some light should be cast over the cases in which solving information asymmetries does not, in itself, drive the system towards an efficient solution. The theory of industrial organization is full of perfect information models that display welfare losses of various kinds, due to the structure of incentives facing agents. An analisys of these cases could prove itself useful to the policy maker by providing him or her with a number of situations in which information-based environmental policy cannot be expected to work, and therefore a more conventional approach is recommended.

37

Bibliography Baumol, W., and W. Oates, 1989, The Theory of Environmental Policy, University Press, Cambridge

Cambridge

Brown, P., 1993, "Playing Football With The Whales", The Guardian, 1/5/1993 Brusco, S., 1993, personal communication Cantoni, S., 1993, personal communication Conrad, J., and C. Clark, 1987, Natural Resource Economics, Cambridge University Press Darby, M., and E. Karni, 1973, "Free Competition and the Optimal Amount of Fraud", Journal of Law and Economics Dosi, G., 1986, Technical Change and Industrial Transformation, Basingstoke

MacMillan,

Francia, P., and G. Ghessi, 1991, "Struttura e tecnologia dell'industria cartaria italiana", in E. Gerelli, Imballaggi: ambiente ed economia, Franco Angeli, Milano Frank, W., 1989, Passions Within Reason, Universe Books, New York (?) Fudenberg, D. and J. Tirole, 1986, Dynamic Modes of Oligopoly, Harwood, Chur Gatsios, K., and P. Seabright, 1990, "Regulation in the European Community", Oxford Review of Economic Policy Glachant, M., 1993, "Voluntary Agreements: A Confidence Building Tool to Achieve Efficiency", paper proposal to "Designing the Sustainable Enterprise", Second International Research Conference of the Greening of the Industry Network, Boston (Mc), USA Harris, C. and J. Vickers, 1985, "Perfect Equilibrium in a Model of a Patent Race", Review of Economic Studies Harris, C. and J. Vickers, 1987, "Racing with Uncertainty", Review of Economic Studies James, A.J., 1993, Sustainable Development: A Critique and an Alternative Approach, mimeo, London Kaldor, N., 1950, "The Economic Aspects of Advertising", Review of Economic Studies Kaldor, N., 1950-51, "The Economic Aspects of Advertising", Review of Economic Studies

38

39 Klein, B., and K. Leffler, 1981, "The Role of Market Forces in Assuring Contractual Performance", Journal of Political Economy Kreps, D. and A.M. Spence, 1984, "Modelling the Role of History in Industrial Organization and Competition", in G. Feiwel (ed.), Contemporary Issues in Modern Microeconomics, McMillan, London Lancaster, K., 1966, "A New Approach to Consumer Theory", Journal of Political Economy Lancaster, K., 1979, Variety, Equity and Efficiency, Columbia University Press, New York Milgrom, P., and J. Roberts, 1986, "Prices and Advertising Signals of Product Quality", Journal of Political Economy Neher, P., 1990, Natural Resource Economics, Cambridge University Press Nelson, P., 1970, "Information and Consumer Behaviour", Journal of Political Economy Nelson, P., 1974, "Advertising as Information, Journal of Political Economy Paba, S, 1986, "Brand-naming as an Entry Strategy in the European White Good Industry", Cambridge Journal of Economics Paba, S., 1989, Reputazione ed efficienza - Crescita e concentrazione nell'industria europea degli elettrodomestici bianchi, Il Mulino, Bologna Paba, S., 1991, "Brand Reputation, Efficiency and the Concentration Process: A Case Study", Cambridge Journal of Economics Pearce, D., A. Markandya, E. Barbier, 1989, Progetto per un'economia verde, Il Mulino, Bologna Pigou, A., 1938, The Economics of Welfare, Macmillan, London Rosen, S., 1974, "Hedonic Prices and Explicit Markets: Product Differentiation in Pure Competition", Journal of Political Economy Shonfield, A., 1984, In Defence of the Mixed Economy, Oxford University Press, Oxford Southern, S., 1993, personal communication Stigler, G., 1961, "The Economics of Information", Journal of Political Economy Tirole, J., 1988, The Theory of Industrial Organization, MIT Press, Cambridge (USA) Veljanowski, C.G., 1985, "The Market for Regulatory Enforcement", The Economic Journal, Conference Papers Waterman, J. and R. Peters, 1984, Alla ricerca dell'eccellenza, Sperling & Kupfer, Milano

40

1A

NBC Christmas show has recently asserted that Wal-Mart, America's largest clothing retailer, buys from Bangladeshi sweatshops where children as young as 9 are employed. This matter is causing vast embarassment to the company, and has already pushed tools and building materials retailer Home Depot to take immediate action to spare itself similar problems [Mc Cormick and Levinson, 1993]. The research work that led to the show had been carried out by a trade union leader. 2

"Two fiddles don't make a Stradivarius" [Rosen, 1974]

3

Darby and Karni's implicit assumption is that consumers don't object to being cheated, as long as the expected loss due to the fraud does not exceed the costs of preventing it. This could be criticized on the grounds of the recent game theoretical developments on human behaviour, whereby individuals with a reputation for retaliation at all costs are never wronged, and therefore never have to incur into the costs of retaliation [Frank, 1989]. 4

In the seventies, the Italian electric appliances manufacturer Zanussi built refrigerators and washing machines. Part of the production was sold to consumers with the firm's own brand; another part, consisting of physically identical machines, was commissioned by the German firm AEG, which commercialized it with its own brand. The AEG-branded appliances commanded, on the German market, a substantially higher price than their Zanussi-branded identical twins [Paba, 1989]. This can be explained in the light of the brand name promotion approach: AEG, through higher level of sunk costs, was signalling more credibly for high quality than Zanussi. 5

The embryo of this idea can actually be traced back to Kaldor's [1950-51] work about the transition of the American capitalism from a "wholesaler dominated" phase to a "manufacturer dominated" one. 6

Much of the literature about business management practices, with its emphasis on quality control and corporate identity, seems to be in line with this intuition. See for example Waterman and Peters [1981?] 7

"Reputation is a word that denotes the persistence of quality" [Stigler, 1961]

8

Some preliminary work has been carried out to extend the model presented in this chapter to the symmetric duopoly case. The "open-loop" [Fudenberg and Tirole, 1986] precommitment Nash equilibrium conditions could not be solved without specifying functional forms; the ideal choice of functional forms is, turns out to be quite a difficult problem. We have chosen not to follow this line of research for the time being. 9

This assumption is obviously unrealistic. Notorious polluters tend to stay in the crosshair of environmentalists: the Italian group Legambiente has even bought shares of some of the biggest, most polluting companies in the country (including the car manufacturer Fiat); it carries out monitoring of their environmental performance on a regular basis, and its representative turn up the each shareholders' meeting to demand a more careful

41

environmental policy. Italian media such events sexy, and give them good coverage, causing major embarassment to the companies involved if their environmental performances are particularly poor. 10

An evaluation of green reputation may be thought of as an assessment of the discounted flow of profits that will accrue to the firm from competing on environmental friendliness grounds. 11

In the detergent market, the Belgian firm Ecover, mentioned in the introduction, provides an interesting example of firm response to such an incentive structure. Its much bigger competitors would need an enormous financial effort to get the same environmentalist simpathy and media coverage that Ecover gets with a 45-employees plant. In this sense, greening production seems to display true scale diseconomies as well as profiting from the absence of sunk costs (noted in chapter 2). 12

Baumol and Oates [1989] supply a clear and exhaustive discussion of market-based instruments, which include taxes, subsidies, tradeable permits, and their various combinations. In what follows, the terms "market-based instruments" and "eco-taxes" will be used indifferently to signify the whole range of market-based instruments. 13

Interestingly, this suggests that the much-criticized anticapitalism of environmentalist leaders may play a positive role: it could be a committment device against collusion with either big business or government. As Frank [1989] has suggested, the best way to perform credibly an independent role is to be independent; lack of self-interest, then, brings about personal advantages. When we asked a Friends of the Earth spokesman if he didn't think that the mahogany campaign implied that big firms are good for the environment (for they obviuosly couldn't target small firms), he answered "Oh, no, they aren't. But we can use the bastards". It is straightforward that the answer translates as "yes"; our hypothesis is that admitting the positive role of big business in the greening of industry he would weaken the set of committment devices that make him a tough bargainer and a credible environmentalist leader. See Frank [1989] for a fascinating treatment of the commitment problem. 14

Coherently with Nelson [1974], Klein and Leffler [1981] and much of the tradition in economics, we assume that consumers are rational, and that their preferences cannot be shaped by advertising. 15

The idea is that a fraction of consumers have read Consumer Report and know in advance if the product is high quality or not. The (monopolistic) firm has to choose whether to cater only for uninformed consumers, supplying low quality, or to cater for all consumers, supplying high quality. It will decide for high quality if p − c1 ≥ (1 − )(p − c0) Where p is the price of the good, c0 and c1 denote the average cost of, respectively, low and high quality and is the share of informed consumers. Clearly, the higher the more likely to be satisfied this conditrion is. So, informed consumers should be subsidized for what they do.

42

16

The Fiat case also shows signs of heavy lobbying behind curtains. The company had long been opposing the introduction of compulsory catalytic converters on all cars in Italy, on the grounds that that would benefit European competitors and affect the trade balance. In 1991 it was ready; government measures triggered a mini-boom of catalysed automobile sales in the spring of 1992, and Fiat, who had been building up stocks, could offer customers shorter waiting lags than its competitors. Despite this, we believe that the agreement was also, and mainly, a signal to environmentally aware consumers. 17

Regulatory capture is the other side of the "trust building" coin. Glachant's "French bias" is obvious in stressing the benefits of businessmen being close to government officials and not its risks. 18

In other examples, shortcuts work better. Under the European scheme, dishwashers and washing machines are ranked on the basis of water consumption, energy consumption and detergent dispersion. Experts stated that these three parameters summarize the environmental impact of this group of products during the use phase of the life-cycle, which accounts for more than 90% of the total environmental impact. 19

Perhaps the most apparent sign of regulatory capture under the European scheme took place in the fields of detergents. In September 1992, twenty small producers, led by the "radical green" Belgian firm Ecover, left the European association of detergent manufactures and associated in the Environmental Detergent Manufacturers Association (EDMA). They contend that major producers are unduly influencing the development of ecolabelling criteria, biasing them towards packaging and turning the attention away from the products themselves..