The Microeconomics Of Environmental Innovation In The Packaging Industry (1994)

June 1994

THE MICROECONOMICS OF ENVIRONMENTAL INNOVATION IN THE EUROPEAN PACKAGING INDUSTRY Alberto Cottica Nomisma - Eco&Eco

Paper prepared for the Fifth Annual Conference of the European Association of Environmental and Resource Economists - Dublin, 22-24 June 1994

The research work that made this essay possible was carried under the scientific supervision of Sebastiano Brusco. Data input was shared among myself, Federica Tagliazucchi, Giuseppe Di Lena and Donata Maccelli. The econometric estimation rested entirely on Raffaele Miniaci's broad shoulders; German interviews were carried out by Sabine Geissler. I am deeply indebted to Jim Skea for discussion, to Paolo Bertossi for encouragement and helpful comments, and to David Ulph for asking the right questions. I also wish to thank all industry experts and companies' spokespersons who accepted to help me along the way. I wish to mention John Beckett and Ruth Steinholtz at Cadbury Beverages Europe, Alberto Bruna at Henkel, Dieter Bürkle and Jean-Jacques Couchoud at Elf Atochem, Giulio Cainelli and Fabrizia Forni at ERVET, Professor Vittorio Capecchi at University of Bologna, Milena Cucconi at Farcon, Robert Dangerfield and Brian Waygood at British Steel Tinplate, Elie Eliasco at Pechiney, Carlo Guidetti at Tetra Pak, Adriano Landi at Barilla, Jennifer Lovell at Shell Chemicals, Carlo Montalbetti at COMIECO, Marcello Pieroni and Dr Iascone at Istituto Italiano dell'Imballaggio, Frau Rafalski at Tengelmann, Len Randall at Marks&Spencer, Dr Rudi at VIAG, Françoise Schell at Prisunic, Francesco Sutti at Saffapack, Roberto Tagliaferri at IMA R&D, Peter White at Procter&Gamble, and above all the exceedingly patient Paolo Simonazzi at IMA. Correspondence to: Alberto Cottica, Nomisma-Eco&Eco, Strada Maggiore 29, 40125 >Bologna, Italy Tel. (+39) 051 64 83 309 - Fax (+39) 051 225 352 - E-Mail [email protected]

2 1. 1.

2.

Introduction

INTRODUCING THE DATABASE OF PACKAGING INNOVATIONS IN EUROPE 2.

Setting the stage: size and structure of the packaging filiére

3.

A database of innovations in the European packaging filiére: data sources and their reliability

4.

A database of innovations in the European packaging filiére: available information

THE SOURCES AND PROCEEDINGS OF ENVIRONMENTAL INNOVATION IN PACKAGING: SOME ANECDOTAL EVIDENCE 5.

Interview format

6.

Strategies for reducing the environmental impact of packaging: source reduction by elimination of overpacking

7.

Strategies for reducing the environmental impact of packaging: source reduction by lightweighting

8.

Strategies for reducing the environmental impact of packaging: re-using

9.

Strategies for reducing the environmental impact of packaging: using recycled material

10.

Strategies for reducing the environmental impact of packaging: materials shift

11.

Strategies for reducing the environmental impact of packaging: developing recycling technologies

12.

A set of hypothesis on the microeconomics of environmental innovation in packaging

3

3.

THE

DETERMINANTS OF ENVIRONMENTAL PACKAGING: A DISCRETE -CHOICE MODEL

INNOVATION

IN

13.

The data

14.

The model

15.

Results

16.

Interpreting results: technology and demand effects on the innovative process

17.

Interpreting results: the role of firm size and position within the filiére

18.

Interpreting results: packaging innovation as a superproduct quest

19.

Interpreting results: private optimality of environmental innovation

20.

Interpreting results: the role of regulation

21.

Conclusions

REFERENCES

4 1.

Introduction

The majority of European large grocery goods companies have now environmental policy statements on packaging. Reading through them, one has the impression that packaging is undergoing a process of change, aimed at the reduction of its environmental impact. The word "sustainability" is used in almost all of these statements. But is this impression correct? Does the European packaging industry innovate to reduce the environmental impact of packaging? How does it do it? Is the process generating "green" innovations any different from that generating innovations otherwise oriented? Which sort of stimuli set "green" innovative processes in motion? How do innovating activities respond to environmental regulation? These questions share a microeconomic focus. This paper reports the results of an empirical investigation into the microeconomics of environmental innovation in the European packaging industry. It does so by a rather unusual research methodology, consisting of a two-stage investigation of the innovation generating process in the packaging industry, and of the role played by environmental concerns in it. The first stage consists in round of talks with industry experts and firms' spokespersons in Germany, the UK, France and Italy, that highlighted innovation strategies and generated hypothesis regarding the microeconomics of environmental innovation, i.e. what explains the choice of a particular firm of steering its technical progress on what Giovanni Dosi would call an environmentally friendly technological trajectory. The second stage consists in the setting up of a database of innovative packings on which these hypothesis were tested. Chapter 1 provides background information on the European packaging filiére and data sources; Chapter 2 reports on the innovation strategies adopted by different groups of players within the filiére and spells out a set of microeconomic hypothesis; Chapter 3 presents an econometric model to test them.

1. 2.

INTRODUCING THE DATABASE OF PACKAGING INNOVATIONS IN EUROPE

Setting the stage: size and structure of the packaging filiére

There is no such thing as a packaging industry. Packaging is rather a filiére, i.e. a number of firms, belonging to different industries, linked to each other by customer supplier relationships. The metal, chemical, glass, paper, mechanics and electronics industries all draw a part of their turnover from packing goods. For this reason, it is no use turning to non-speciaize statistical sources to find out how many firms are actually involved in the process. We distinguish eight different filiére partitions, henceforth called "sectors". The RAW AND FINISHED MATERIALS FOR PACKAGING sector (sector 1) includes, together with raw materials manufacturers stricto sensu, firms which treat such materials into shapes that can enter a production line. This is especially important with plastics, where things like bars and tubes, coupled sheets and PET pre-forms are manufactured. These firms are called "converters" in the industry's technical jargon. Firms in sector 1 employ technology that range from "mass production" to fine chemistry. The FINISHED PACKINGS AND COMPONENTS sector (sector 2) includes manufacturers of items such as boxes, bags, bottles, aerosol containers, but also edge reinforcements, pouring spouts and labels (pallets manufacturers are counted in sector 7). Plastic films and plastic sheets are also included, which is somewhat odd because they are manufactured by means of conversion techniques. So, converters are split between this sector and the raw materials one. It is important to notice that the classification is done accordingly to the markets firms cater for, not to the technology they employ. so, for example, Rocco Bormioli, the largest glass manufacturers in Italy, is recorded as a finished packings producer because it has vertically integrated forward to supply bottles, rather than glass, to packaging users. The range of technologies used here is therefore quite wide, spanning from chemistry to fine mechanics. The MACHINERY FOR THE MANUFACTURE OF PACKING AND PACKAGING PRODUCTS sector (sector 3) includes manufacturers of machines that perform operations such as expanded polystyrene pressing, diecutting, injection moulding. This sector employs a highly flexible fine mechanics technology. The PACKING AND PACKAGING MACHINERY sector (sector 4) includes manufacturers of machines that perform the operations of putting a product in its pack. Feeding and filling machines, forming machines, closing machines, palletizers are examples of the range of products. The technology employed is fine mechanics again, with an important contribution of electronics in the latest years: in the newest models, electronics may account for as much as 40% of the selling price [Capecchi, 1993]. The AUXILIARY MACHINES AND EQUIPMENT sector (sector 5) includes manufacturers of such miscellaneous equipment as shaping machines for polyurethane foam, washing machines for containers, labelling machines, cutters. Fine mechanics seems to be again the prevailing technology, but not the only one employed.

6 The ACCESSORIES AND TESTING EQUIPMENT sector (sector 6) include producers of biologic, chemical, mechanical, physical and optical testing equipment as well as items like meters, sterilizing systems, pumps and motors. Technologies employed vary accordingly, as can be said of sectors 7 and 8. The HANDLING AND STORAGE EQUIPMENT sector (sector 7) includes warehouse equipment manufacturers. Products range from pallets and box-pallets, to forklifts, to conveyors and even shelving. The SERVICES sector (sector 8) includes designers and engineering firms with a specialization in packaging. It also includes firms that perform packing and packaging operations on other people's products. The customer-supplier relationships among sectors are highlighted in Figure 1. Figure 1. also presents some key figures for each sector: number of firms, number of employees, and the Herfindahl-Hirschmann concentration index1. The figures refer to Italy, the country for which statistical information abut the packaging filiére is most complete of the four. Industry experts confirm that the following statements regarding concentration can be generalized to Europe: 1.

the raw materials sector is highly concentrated, dominated by multinationals

2. the packaging technology sectors (3-7) are more concentrated elsewhere than in Italy, but still rather dispersed 3.

1HH

the finished packings and components sector is very dispersed.

is computed under the assumption that the market share per employee is a constant within sectors. This assumption is at least dubious, and is sure not to hold for sectors 1 (whose concentration is underestimated) and 2 (whose concentration is overestimated) because of the vertical integration problems highlighted in their definitions.

7

8

It is important to keep in mind that most firms in the filiére are multiproduct. We have, however, assigned each of them to only one sector on the basis of the number of products belonging to each sector manufactured. The flow chart in Figure 1., therefore, should not be taken too literally. Among the four countries, France, Italy and the UK publish yearly Packaging Directories, which keep track of most, if not all, firms involved in the filiére. The Italian packaging filiére (users are not included in the definition of filiére) consists of about 2,200 firms; the British one of about 2,600 firms; the French one, of about 3,000 firms. These figures are liable to be underestimated, since very small firms, or firms to whom the production of packaging is marginal, may just not bother to make an entry into the Directories. In fact, a different estimate [III, 1992] Quotes a figure of about 3,500 firms for Italy. It is much more difficult to estimate the filiére turnover in the presence of vertical disintegration. An IFEC estimate based on calculations done by the European Institutes of Packaging reports the production of packs to be worth around 76 bln ECU in 1992 [Morino, 1992]. The four countries account for over three fourths of this figure. The IFEC estimate does not take into account our sector 4, which normally sells directly to packaging users. According to a COPAMA estimate, it is worth about a further 6 bln ECU in 1990 [Reis, 1992]. The entire filiére production should therefore amount to around 82 bln ECU, as shown in Table 1. TABLE 1. PRODUCTION OF PACKS AND PACKAGING AND PACKING MACHINES IN EUROPE.

Packs - 1992 Machines - 1990 Germany France Italy UK Rest of Europe

16,73 14,45 12,93 12,17 19,77

2,59 0,50 1,64 0,30 0,89

Europe

76,03

5,92

Billion ECU. 11/3/1994

Sources: IFEC,

Copama.

Exchange

rate

of

We have found the total number of employees in Europe to be somewhere around 550,000, plus a further 20,000 in the "auxiliary " sectors 5, 6 and 7. Of the 550,000, about 450,000 operate in the "vertically integrated finished packings and components sector" (sectors 2, 3, 1 and 8). The remaining 50,000-plus operate in sector 4. This is summarized and broken down by country in Table 2.

9 TABLE 2. N UMBER OF EMPLOYEES IN SECTORS 2, 3, 1 AND 8 AND IN SECTOR 4 IN G ERMANY, F RANCE, THE UK AND ITALY

Employees

Packs

Machines

Total

Germany France Italy UK

143.701 124.106 111.042 104.510

38.149 7.310 24.212 4.371

181.850 131.416 135.254 108.881

Total

483.359

74.041

557.401

Source: own estimate on ETAS, IFEC and Copama data

There is a further complication to introduce, which will not be analysed further but must be kept in the background. The discussion has so far been in terms of packaging filiére, but in fact different techniques are wielded by different firms to treat different raw materials into packings; therefore, it would be more rigorous to speak of glass, plastics, paper, metals filiéres. Firms tend to specialize in materials: in general, a firm making glass bottles will not have any cardboard box in its range, and it won't even have plastic bottles. The exceptions to this rule are the very large packaging firms (Carnaud Metalbox, VIAG, Pechiney). Figure 2 and 3 show how the European packaging production, expressed in weight and in value, is distributed among the different raw materials. FIGURE 2. P RODUCTION OF PACKAGING MATERIALS IN EUROPE (BY WEIGHT)

8%

20% Metals Cardboard

29%

wood Plastics 4%

Source: IFEC, III

39%

Glass

10 FIGURE 3. P RODUCTION OF PACKAGING MATERIALS IN EUROPE (BY VALUE)

1% 36%

Metals Cardboard

44%

wood Plastics Glass 1%

18%

Source: IFEC, III

3.

A database of innovations in the European packaging filiére: data sources and their reliability

One of the key issues here is the relationship between environmental regulation and technical progress. As a vast body of literature documents, it is by no means easy to measure the latter. However, industrial economists have tried to do it by asking a panel of experts to build lists of innovations, to each of which a string of variables could be attached. The result is a database of innovations which lends itself to statistical treatment. This has provided important insight in the economics of innovation [Townsend et. al. 1981, Scherer 1982]. This work draws on that tradition. In order to monitor technical progress in the packaging industry, a database of packaging innovations in Europe in the last 15 years was set up. Innovative packings competition in the four countries are used as data sources. Interviews with industry experts and firms' representatives have shown European competition to be fairly representative of the industry's innovative activity. All of these countries hold competitions, organized by the national packaging institutes, for innovative packaging solutions. In addition to these five competitions (there are two in the UK), there is a fourth one held, at a European level, by the European Packaging Federation. It is worth it to recall the main characteristics of each of these data sources. ITALY: OSCAR DELL'IMBALLAGGIO - Oscar dell'Imballaggio is the oldest of the European packaging competitions: the first edition was held in 1956 (meaning that the entries were accepted in 1956 and the prizes were awarded in 1957) and knew almost

11 no discontinuities to the present day. It is organized yearly by Istituto Italiano per l'Imballaggio. Entry is restricted to Italian firms (that is, to Italian innovators, whereas users do not have to be Italian). Data have been obtained from 1981 on. FRANCE: IFEC PRIZE - The Institute Francais pour l'Emballage et le Conditionnement organizes a yearly competition similar to Oscar dell'imballaggio, and nearly as old as its Italian counterpart. Its outstanding features are the large number of prizes awarded each year (40-50) and the good quality of IFEC's files, from which data can be obtained directly. Data have been obtained since 1978, with a few discontinuities. UNITED KINGDOM: STARPACK. Organized yearly by the Institute of Packaging, Starpack is modelled on Oscar dell'Imballaggio. One major difference is that there are three ranks of awards: Gold, Silver and Bronze Stars. We have adopted the policy of entering all innovations, regardless of their rank. Data have been obtained since 1978, with a few discontinuities. UNITED KINGDOM: INNOVATORS OF THE YEAR. The Institute of Packaging has been running this competition since 1989. It is structured in a way roughly similar to Oscar dell'imballaggio, with three differences. The first one is that there is only one winner for each of the "categories" (initially three, now five); the second one is that the entry is open to foreign firms as well; the third one is that, since 1990, a "Packaging and the environment" category was created. Data regarding all four editions have been gathered. GERMANY: VERPACKUNGSWETTBEWERB. The RKW organizes a competition structured in a way similar to the Oscar - IFEC model. It issues a large number of awards per edition (usually about 40), but it is held only every three years. Data are available for all editions since 1978. EUROPE: EUROSTAR - Eurostar was established by the European Packaging Federation only one year after the launch of Oscar dell'Imballaggio, and explicitly linked to the national competitions: the winners of the former enter automatically the latter. To avoid double counting, innovations from Germany, France, the UK and Italy countries have been excluded from the data set. Data are unusually hard to obtain, and have been gathered for a limited number of editions; this is due to the fact that the EPF only exists on paper: national associations take turns in organizing Eurostar. There seems to be an ample consensus that such competitions are representative of the innovative activities going on in the packaging filiére. They are important enough for firms to enter, but not important enough for them to try to bias their outcome. Individual freelance designers, and even industrial design students, have been known to make winning entries. Also, since rules typically require that packs be already traded on the mass market before they can enter (except for the prototypes, which is were individual designers come in), clever-but-economically-unviable ideas are screened out. The witnesses the research group has talked to maintain it is reasonably safe to treat award-winning packs as a proxy of innovation on packaging (but not, for example, in recycling techniques). 4.

A database of innovations in the European packaging filiére: available information

Each of the about 1,400 innovations awarded with one of these prizes was coded into a string of data. Available information concerns:

12 *

The year of entry

*

The name of the innovator and of the user of the innovation, and their countries of location (for innovators, the latter is usually bounded by regulations internal to each competition: for example, a company has to be Italian to enter Oscar dell'Imballaggio. Multinationals get round the problem by entering competitions with their local subsidiaries. When there is no user, innovations are treated as prototypes.

*

The sector (in the sense given above) to which the innovator belongs within the filiére. The coding is as in section 1.

*

The sector in which the pack is to be used, coded with the relevant EUROSTAT three-digits code

*

The number of employees of both innovators and users. Firms were grouped in eight classes: up to 10 employees, from 11 to 25, from 26 to 50, from 51 to 100, from 101 to 250, from 251 to 500, from 501 to 1,000 and over 1,000.

*

The function served by the pack. This may be presentation, transport or both.

*

The reasons, given by the jury, for the awarding of the prize. They are quite easily grouped into nine categories. Use of new materials Packing previously unpacked products Cost reduction Distributor friendliness User friendliness Product protection Aesthetics Low environmental impact Others

The jury can, and often does, give more than one reason for finding the prizewinning pack innovative. In these cases, all of them were reported, and an effort was made to extrapolate the main one. VIII. The number and kind of the materials used to manufacture the pack (labels and inks not included). The possibilities are wood, aluminium, steel and other metals, paper and cardboard, polyethylene (henceforth PE), PET, PVC, polystyrene, other plastics and composites (e.g. tetrapak). IX.

The total number of recorded innovations presented by the same industrial group entering this particular innovation. The choice of looking at groups, rather than single firms, is due to the centralization of some R&D facilities and to the frequent innovative spillovers between firms within the same group.

13

2.

THE

5.

Interview format

SOURCES AND PROCEEDINGS OF ENVIRONMENTAL INNOVATION IN PACKAGING : SOME ANECDOTAL EVIDENCE

The reduction of the environmental impact of packaging can happen by means of very different technical innovation strategies. This section illustrates them, highlighting their similarities and differences; its purpose is to generate a range of hypotheses about the microeconomics of environmental innovation, to be tested for through the use of econometric techniques. It consists of anecdotal evidence, gathered by a series of discussions held with industry experts and companies spokespersons in the four countries, which were carried out mostly in the period January-March 1994. These talks were aimed at finding evidence on the state-of-the-art of environmental innovation in packaging rather than sketching a picture of the general trend; therefore, we concentrated on companies known for being particularly sensitive to environmental concerns. Most of them are members on environment-oriented trade associations, such as ERRA or INCEPT. Respondents were asked to confirm or challenge, supporting their response with examples, the following two statements, generated by earlier statistical treatment of the data on Italian innovations alone: 1.

environmental innovation, like packaging innovation in general, is largely done by manufacturers of finished packings and components. These, however, work in very close connection with their customers, who identify problems and priorities; it is almost as if packings manufacturers were acting as R&D facilities for packaging fillers.

2.

environmental concerns and regulatory recycling targets for packaging are not inducing any significant shift from materials that are difficult to recycle (plastics, composites) to materials that are easily recyclable (glass, metals, paper).

The interviews resulted in a clearer identification of the research strategies available to reduce the environmental impact of packaging through innovation, and of the complex relationship between choice of innovation strategy, technical constraints and market structure variables. A common feature of all strategies is that they identify a more or less environmentally-friendly waste management option to be applied to the innovative pack. In what follows, we divide our anecdotal evidence on state-of-the-art environmental packaging innovation according to the options proposed in the EU Packaging Directive: source reduction, re-using and recycling. Somewhat loosely, we have grouped incineration with energy recovery with recycling, a position enthusiastically shared by respondents. It is important to underline that respondents rejected strongly the idea that any waste management option is inherently environmentally superior to the others: plastics and composites producers and fillers, for example, point to the superior energy efficiency of plastics packaging, lighter to transport, and maintain that recycling is, generally speaking, an inefficient option. Metals producers, on the other hand, prefer to talk about high recycling rates and not about energy consumption; paper producers highlight the fact that they are tapping renewable resources, and that the forestry industry contributes to fight global warming. Although some of the more research-oriented companies use life-cycle models to evaluate the overall environmental impact of

14 packings, it is true that no clear benchmark has been established to tell regulators and consumers what is environmentally friendly and what isn't. 6.

Strategies for reducing the environmental impact of packaging: source reduction by elimination of overpacking

Source reduction can take place in either of two ways. One is the elimination of what industry experts are beginning to call overpacking, layers of packaging which, while adding to the service content of the pack, are not strictly necessary to its delivering a satisfactory performance. Overpacking is, of course, a totally subjective concept. One example of elimination of overpacking comes from the experience of the Italian food producer Barilla. Its successful minicakes, a relatively expensive and high-margin range of products (Dolcetti), used to be conditioned one by one in small rigid paper baskets, which could be stacked inside a plastic sack. The paper baskets system ensures a better protection of the product, and it makes it easier to extract one minicake from the pack. Nevertheless, they were eliminated: now minicakes are simply conditioned into the sack. Barilla integrates the R&D on packaging into an 11-strong research unit, and designed the innovation itself. About 12% of the research unit's time is devoted to the reduction of the environmental impact of packaging. Of the four under investigation, Germany is probably the country where overpacking is most counterproductive in marketing terms. What's more, the implementation of the Dual System has imposed an additional cost on all packs earlier than elsewhere. The result of this is a richer array of innovations in the sense of overpacking abatement. It is worth it to report four examples from different industries. *

Bottles of wine from Asbach & Co. Weinbrennerei are no longer wrapped in transparent polypropylene foil, which gives the bottle a more glossy appearance. This saves about 40 tons of foil a year.

*

The cosmetics producer Croldino Schneider manufactures a hand-washing cream which comes in plastic tubes. The tube used to come into a cardboard box for presentation and space-fractionating purposes, which has now been removed. About two tons of cardboard a year have been saved.

*

Schöller Lebensmittel & Co. produces ice cream on an industrial basis. Ice cream is conditioned in polystyrene thermal boxes, which used then to be packaged in cardboard boxes for presentation purposes. The cardboard boxes have now gone, and the presentation function is taken care of by paper labels stuck directly on the polystyrene box.

*

Hans Warholtz Konserven's sardines are conditioned into tinplate steel cans; they too used to come in presentation cardboard boxes. Now the information about the product is printed on the tin, and the cardboard box has been removed.

Blister packaging is also increasingly being regarded as overpacking. Blisters are PVC thermally shaped packs, often glued to a cardboard base for presentation purposes. Consumers, and German consumers in particular in particular, tend to dislike this sort of pack, and they are now being substituted for. Beiersdorf, for example, manufactures an office glue stick called Tesa, which used to come in a blister (about five times the size of the stick itself), with an eyelet, used to hang the pack to displays in stationery stores. In 1992, the blister was eliminated altogether, and a PET hook fitted to the cap of the stick. This way, the stick can still be displayed while saving the PVC. The manufacturer calculates the material saving in about 4 tons per year. A slightly different version of this same idea won a VPW in 1993.

15 An extreme version of the same operation was done by the German bicycle component manufacturer Prophete. Prophete sells, among other things, bicycle bells, which used to come in a blister as well . The brand is printed on the bells themselves, so the blister was eliminated altogether. A small cardboard tag carrying the bar-code is now the only form of packaging attached to Prophete bells. 7.

Strategies for reducing the environmental impact of packaging: source reduction by lightweighting

Another path to source reduction is what goes under the name of lightweighting. It consists in redesigning the pack so that it delivers the same performance with less material. From a merely technical point of view, lightweighting is more interesting than elimination of overpacking, because it involves "real" re-engineering of the pack, and therefore a comparison of R&D costs and expected benefits. The raw materials side of lightweighting is called downgauging; this means producing a grade of a polymer which will yield good properties even when extruded to a very thin layer. This innovation strategy is, of course, a no-regret one; a lighter pack costs less money than a heavier one, regardless of what it does to the environment (in fact, if the environmental problems caused by packaging were thought to be waste generation only, it could successfully be contended that lightweighting does not do much good to the environment, because it does not necessarily reduce the volume of packaging waste). This encouraged lightweighting even before the environment became a core issue in the political agenda of European governments. Figure 4 depicts the decrease in body weight of a representative 0.44 l steel beverage can. It is immediate to see that the 30% reduction in weight did not come in a single wave, but by means of several incremental innovations dating back to 1979. Environmental awareness of consumers has sped up the process; the introduction in various forms, of packaging levies throughout Europe are expected to speed it up further. FIGURE 4 - WEIGHT OF A 0.44 L STEEL BEVERAGE CAN OVER TIME

45 40 35 30

93

92

91

90

89

88

87

86

85

84

83

82

81

80

79

25 Source: British Steel Tinplate

It can certainly be contended that lightweighting is taking place in pretty much all consumer good industries and with packs of all packaging materials. This shows in Table 4.

16 TABLE 4 - M ATERIAL SAVING THROUGH LIGHTWEIGHTING 19701990 Pack

Material

Wine bottle 0.75 l Beer bottle 0.25 l Can 4/4 Heavy duty sack Shopping carrier Yoghurt glass Detergent bottle 2 l Shrinkwrap film for pallets

Glass Glass Steel Paper PE PS HDPE PE

1970 (g)

1990 (g) Variation

450 210 69 247 23 6.5 120 1400

350 130 56 215 6.5 3.5 67 350

-22% -38% -19% -13% -70% -45% -45% -75%

Source: various sources, reported from Elf Atochem

An example of lightweighting from the soft drinks industry comes from the Japanese firm Hosokawa, that sells drinks and diet integrators for athletes. These drinks come in flexible packs: a "specialized supplier" sort of relationship between the Japanese user and two Italian societies, Guala Pack (a packaging manufacturer) and Safta (a conditioning technology producer), has led to the development of CheerPack, a flexible container for liquid whose main feature is that of being confortable to carry (it adapts its shape) and easily reclosable. These are useful features to athletes, who take a sip at a time while, for example, taking part in a bicycle race. CheerPack weighs about 65% less than earlier soft drinks flexible packs. It won an Oscar dell'imballaggio in 1990. Detergent producers have taken this concept one step further, and lightweighted their product as well as their packs. Procter&Gamble is probably the single company that pursues lightweighting most vigorously; it is worth recalling the Ariel Ultra story. Ariel is a laundry detergent. In 1989, P&G introduced Ariel Ultra, the first compact detergent in Europe. The formulation had been changed; the detergent could now be a effective with a smaller amount of powder, which was placed directly inside the washing-machine drum. As a result of this, Ultra, like all compacts, uses less packaging per load washed than non-compacts. P&G calculates that, on Ariel Ultra alone, more than 12,500 tons of packaging material have been saved throughout Europe. The move was quickly imitated by competitors, and compact detergents enjoy now a share ranging from 20 to 60%, depending on the country. It is estimated that the overall savings on packaging materials in Europe since the introduction of compacts amount to 800,000 tonnes approximately. Notice that this involved no packaging innovation at all. The second step in the lightweighting of Ariel was the introduction, in 1992, of a refill, which comes in a minimal package similar to a sugar bag. Once home, the consumer pours the detergent in the carton (which lasts 4-5 refills on average) and disposes of the refill bag. This saves a further 50% of packaging material with respect to the compact carton. Again, only fine tuning of the pack was needed, so innovative activities were limited to a minimum; again, the move was quickly imitated by all major European detergents producer. With a stronger innovative effort, Henkel produced a lightweight pack for liquid detergents, essentially a flexible pack with a cardboard reinforcement to hold it standing. The refill philosophy is being enthusiastically adopted by detergent manufacturers. Refills of liquid detergents, however, posed a technical problem, that of finding an ultra-light packaging solution that was also waterproof. P&G's Dash 3 Ultra is an example. In this case, the answer was found in laminate stand-up pouches (the stand-

17 up feature is necessary to optimize exploitation of shelf space in supermarkets), a 20year-old French invention that had hitherto found very limited application. The expertize for manufacturing pouches from laminate was held by a French converter called Soplaril, a subsidiary of the plastic giant Elf Atochem. P&G asked Soplaril to develop the pouch; Soplaril, then, worked together with Elf Atochem for the fine tuning of the laminate, almost a textbook example of specialized supplier relationship. Soplaril is now enjoying a rapid growth due to the rediscovery of the stand-up pouch on environmental grounds. The savings of packaging material are, in this case, even more significant: a pouch weighs 70% less than a plastic bottle. 8.

Strategies for reducing the environmental impact of packaging: reusing

In principle, containers can be used several times, thus lowering the amount of packaging waste produced per unit of product packaged. In fact, quite a lot of transport packaging (pallets) and some primary packaging (typically, beverage bottles) are recovered for re-use on a regular basis. Switching to re-usable containers, however, can be very difficult, because it involves setting up a product-specific recovery system. Prize-winning innovations of re-usable containers in the database are all transport packs but one. Despite such difficulties, re-use of primary packaging is being contemplated by the beverage industry. Plastic manufacturers, and notably ICI, are undertaking highly structured and very expensive R&D projects about what goes under the name of "flavouring" to make plastic bottles re-usable. The technical problem with this is twofold. Firstly, bottles are washed with very hot water before re-use, and most plastics will melt or degrade. Secondly, since both plastics and flavours of soft drinks are oil-based, beverage and bottle interact chemically, and the latter retains the taste of the former. In fact, contact with some plastics can alter substantially the taste of some beverages; to prevent these problems, the beverage industry works with a flavourmaterial compatibility matrix. Glass bottles, on the other hand, are perfectly reusable because of a relatively high melting point and chemical inertness. A first success was obtained by Continental PET, who produced a re-usable PET bottle for Coca-Cola: however, it won't work with less strong-tasting drinks, like water. 9.

Strategies for reducing the environmental impact of packaging: using recycled material

Another possible strategy to reduce the environmental impact of packaging is to incorporate recycled material in it. This creates a market for secondary raw materials, thus diverting some waste from the waste stream. In some cases, use of recycled material in packaging does not stem from environmental considerations at all; it is simply a cost-saving feature of the production process. This is, for example, the case of steel cans, which embody 25% of scrap steel, or of most cardboard packaging (typically, detergent cartons). Substitution of recycled material for virgin material, on the other hand, poses several technical problems, due to poorer performance and relatively bad looks of the former. For this reason, reduction of environmental impact by this innovation strategy is not as widely pursued as, for example, lightweighting. Nevertheless, some companies, especially, but not only, in the detergent industry, have gone down this route. The innovation database records several examples. One part of recycled PE can be mixed with three of virgin PE to yield a material with reasonable mechanical properties, reasonable enough to condition liquid detergents. However, recycled PE has a disagreeable grey-green colour, disliked by marketing

18 men; furthermore, the colour is subject to random variations. Two routes have been attempted by Procter&Gamble to solve these problems. The Viakal bottle, winner of an Italian Oscar in 1992, incorporates a coloured master batch to camouflage the greygreen: then, a shrinkwrap label with good graphics was added, in order to cover the colour variations. A better result, from an aesthetic point of view, is guaranteed by the development of a coextrusion technology, that allows the blow-moulding of a three-layer bottle. The middle layer is made of recycled PE; the inner and the outer one, of virgin PE. Again, the weight of the former is about 25% of the total weight of the bottle. This innovation won another Italian Oscar, in 1990. Interestingly, the R&D that led to the invention of the coextrusion technology was done jointly by P&G and several of its suppliers; the company launched the idea and proposed converters to solve the problem. This way, a share of the R&D costs was dumped on to the supplier; the deal was that, had they been successful, P&G would sign a large order at a relatively advantageous price for the new bottle. 10.

Strategies for reducing the environmental impact of packaging: materials shift

Of particular interest are those innovations that imply a change in the raw materials from which the packing is made. This is the only innovation strategy that is not aimed to any one waste management option; it is generally, but not always, directed to increasing the recyclability of the pack. "Recyclability" is, in the packaging industry, a fairly abstract concept; which has nothing to do with the existence of collection, sorting and recycling systems. For example, single material plastic packs are marketed as easily recyclable; what that means is that, if recycling facilities and a separate collection system for that particular plastic existed, the pack could all go into the same polypropylene or PET bin. The reason why we are interested in looking at this particular kind of innovation is that packaging regulation, especially in Germany, is now rearranging the set of incentives to use some raw materials rather than other. The principle is to collect a sort of levy on packs, that is then used to pay collection, sorting and recycling costs. Table 5 reports the set of prices that packaging users selling their products in Germany have to pay to participate in the Duales System. TABLE 5 - DSD TARIFFS PER KILO OF RAW MATERIAL

Material Glass Paper and cardboard Tin Aluminium Plastics Cardboard composites Other composites Natural materials Source: DSD, March 1994

DM/Kg 0,15 0,40 0,56 1,50 2,95 1,69 2,10 0,20

19 The Duales System introduces an incentive to move out of plastics and composites. In the case of some polymers, like PE, the DSD levy amounts to about 100% of the price of the virgin material, so the incentive can be quite substantial2. Under these conditions, one would expect to find a trend to replace plastic packaging with glass, paper and metal packaging. Some of the firms interviewed have reported developing or using "green" innovative packings that embody some material substitution. A well-marked trend concerns the move out of PVC, perceived as environmentally unfriendly3.This has gone so far that Tengelmann, Germany's largest retailer, sells nothing in PVC. In the beverage industry, for example, PVC was almost totally replaced by HDPE in bottle caps, and by PET in bottles for non-carbonated drinks. The food industry is also moving out of PVC: Marks&Spencer reported replacing its PVC trays for prepared meals (a pack that has to be microwaveable) with foams, whereas Barilla removed PVC trays from its line of minicakes. The latter is a quite interesting example of how innovations tend to have consequences all along the packaging filiére; Barilla's minicakes used to sit on PVC trays. The company's environmental policy is now to move out of all plastics that contain elements other than hydrogen, carbon and oxygen, so as to minimize the environmental impact of incineration, so PVC had to be taken out. . The choice was to use cardboard instead. This involved no innovation in the packaging manufacturing technology, because a cardboard tray is very easy to make with a simple diecuttingshaping-gluing process. However, it involved adding a step to Barilla's conditioning lines: PVC trays are bought from a converter, put on a conveyor belt by a machine, and another machine places the minicake on them. On the contrary, cardboard trays are shaped and glued in Barilla's plant, so the conditioning line had to be modified in order to support a shaping and gluing continuous process. Barilla's packaging experts designed the tray, and a medium-sized packaging technology firm in Bologna designed and realized the improvement to the conditioning lines. This move takes out of the waste stream 150 mln PVC trays a year. Barilla is an example of a company making extensive use of packaging material shift, aimed at either recyclability enhancing, incinerability enhancing or both. Other innovations of this kind are the substitution of the PE, aluminium and paper composite formerly used in biscuit packs with a two-layers PE-aluminium composite (with the percentage of paper rising from 60 to 80 per cent) and the replacement of polypropylene-aluminium composites with metallized polypropylene in all packs. PVC is at the centre of another example of all-filiére innovation which involves material shifts. In the pharmaceutical industry, pills, capsules and the like are usually conditioned in blisters. The "traditional" blister is made of a PVC transparent shell, glued to an aluminium foil; pressing on the PVC makes the pill tear the aluminium and drop out the back of the packing. In the 80s, German pharmaceutical companies decided to move out of PVC, which can release toxic substances when incinerated. Unfortunately, PVC was the only cheap polymer with good barrier properties for which a thermal shaping technology existed. It was decided that polypropylene could be the closest substitute for PVC. Bayer started a joint R&D project with the Italian packaging technology company IMA; IMA technicians built a prototype polypropylene blister-making machine, initially very

2

It must be noted, however, that the levies are expressed by weight, which mitigates the disadvantage for low-weight plastics. 3The

move out of PVC has reached innovation as well. A Hausman test comparing innovative packs invented before 1992 with those invented in 1992 show that the latter use significantly less PVC.

20 similar to the standard PVC ones, and started feeding it polypropylene. Some of the problems they met with could be solved by modifying the machine; for example, it turned out that pre-heating polypropylene before thermally shaping it improved greatly the performance of the machine. Other problems were solved by means of fine-tuning the grade of the polymer. The process was a trial-and-error one, where both fine mechanics and fine chemistry technologies were used to get round obstacles, and fed into each other. The results of the process were a new blister-making machine that could work with polypropylene and a new grade of propylene to feed to it. In order to increase the recyclability of blisters, German companies are now working on the allpropylene blister, which is already employed by some manufacturers. It is a kind of blister which replaced the aluminium foil with an easy-to-tear polypropylene foil. Another example of recyclability-improving material shift is EcoTop, a steel can top developed by British Steel Tinplate. Cans are made of two parts: bodies, which undergo a blow-moulding process, and tops. At the moment, about a half of the bodies are made of steel, the other half of aluminium; on the contrary, tops are all aluminium made because of technical difficulties in producing steel tops with a satisfactory performance. EcoTop has no ring to pull, but two "buttons" to push down and fond inside the can, one to pour the drink, the other for air to flow in as the drink is pulled. It doesn't perform as smoothly as traditional tops, but it can be marketed as very environmentally friendly because of the lower energy intensity of steel with respect to aluminium and because of the very easy recyclability of all-steel cans, made possible by the new top. Steel can be pulled out of the waste stream by magnetic extraction and sent back to the furnace for re-melting. British Steel feels the environment is now the single most important ground for competition to producers of raw materials for packaging. 11.

Strategies for reducing the environmental impact of packaging: developing recycling technologies

Some materials, like plastics and composites, are difficult to recycle; of others, recyclability, while already good, can still be improved. In order to reduce the environmental impact of packaging, then, some firms work on the development of better, cheaper recyling technologies. It is mostly raw material manufacturers, who have the necessary expertise and feel the green movement's pressure more than firms elsewhere in the filiére, engaging themselves in such efforts. Recycling has become a very important public relation issue for highly visible multinational companies, who fund pilot schemes and produce literature. Especially active in this field are plastics producers: APME, the association of plastic manufacturers in Europe, lists 109 plastic recycling schemes going on. Nevertheless, technical progress in this direction has not so far been impressive. Shell claims to be doing quite a lot of R&D on what goes under the name of "feedstock recycling". The idea is to work waste plastics into an oily feedstream suitable for cofeeding to existing petrochemical or refinery processes, such as gasification, distillation or hydrogenation into more refined processes. This option features a higher public acceptability than waste-to-energy (the additional plants would be seen as part of existing refineries), and a better energy balance, as hydrocarbon molecules are not destroyed. Moreover, feedstock equipment will be easy to integrate into existing refineries or petrochemical plants, thereby minimizing the capital cost of the operation and exploiting economies of scale in full. However, such plants are large and very far from each other; this implies such schemes would have high transportation costs. Fully developed feedstock recycling technology is not yet available. Elf Atochem is heavily involved in Valorplast, the French plastic recycling consortium. They follow a line of research to improve plastics sorting technologies; Valorplast will soon be starting a sorting plant which employs an Italian X-ray scanning technology to

21 separate PVC from PET, and tests are being run on a different line which separates PE as well. British Steel Tinplate declared that steel producers are constantly working for the improvement of their recycling technologies. To them, it is not a public relation issue; scrap steel is a strategic resource for the industry, since all grades of steel employ a relatively high proportion of scrap and some, like stainless steel, are made of 99% recycled material. A tin-separating treatment to recycle the layer of tin that covers scrap tinplate steel was also developed. Tetra Pak is engaged in several R&D projects to recycle cardboard composites. It is trying to develop technologies that recycle waste composite into items such as pallet spacers, thermal insulating material and even shoe sole reinforcements, as well as investigating repulping techniques. However, it is dubious that these products can find a market in the short term. Finally, Marks and Spencer funded the development of a technology for the recycling of PVC. It is not being used, because the company finds it is too expensive. 12.

A set of hypothesis on the microeconomics of environmental innovation in packaging

The interviews show that the packaging filiére is simultaneously pursuing several strategies in order to reduce the environmental impact of packaging. Often, different strategies are pursued within the same company, as Table 6 shows. TABLE 6

-

EXAMPLES

OF

ENVIRONMENTAL

INNOVATION

STRATEGIES Strategy

Companies

Elimination of overpacking

Tengelmann, Barilla, German companies

Lightweighting

Henkel, P&G, Tetra Pak, Barilla, M&S, Cadbury, British Steel

Re-using

Coca Cola, ICI

Using recycled material

P&G, Henkel

Material shift

IMA, Barilla, M&S, Henkel, British Steel, P&G, Cadbury, Tengelmann

Developing recycling techniques

Tetra Pak, Shell, British Steel, Elf Atochem

However different in concept and technicalities, these strategies share some similarities. Firstly, with the partial exception of the development of recycling techniques, environmental innovation in packaging tends to be very incremental. This reflects the existence of a number of technical constraints underpinning this as well as other sectors. For example, if the beverage industry were to replace cans with glass bottles, transportation costs would increase; this would make the existing scale-intensive beverage plants inefficient, and would call for a new industry configuration, characterized by a larger number of smaller plants serving regional or national markets. The Belgian non-returnable beverage container tax is having precisely such effects. However, within the rigidities imposed by the technical properties of materials and the packaging needs of different products, the environmental impact of packaging attracts quite a lot of attention. Many routes to environmental improvements are being investigated at the same time, and the technologies employed are flexible enough to yield a relatively high degree of innovativeness. This is reflected by the unexpectedly high number of innovations in the database that make some kind of environmental claim, 110 out of 1407.

22 Secondly, there is a very high degree of "vertical" cooperation in the innovative process. Often, the history of an environmentally friendly innovative pack begins with its filler imagining it and asking to its supplier of packings to make it for him. The latter, then, sets out to solve the technical problems implied by the manufacturing of the new pack. Sometimes, this implies cooperating with the raw material manufacturer to fine-tune the material to suit the need; sometimes it implies modifying the conditioning technology, a task that is generally carried out by packaging technology firms. In the long run, suppliers and customers learn to work together towards a common goal; this is reflected in the time span, typically very long, of business relationships in this area. Quite often, indeed, the researcher investigating these business relationships find themselves looking at examples of quasi-vertical integration à la Richardson; in the beverage industry, for example, the can factory sits next door to the beverage factory, and the two are linked by a system of conveyor belts that carry 6,000 empty cans a minute from the former to the latter. In the detergent industry, there are examples of a Japanese-style relationship with converters blow-moulding bottles for liquid detergents; the converter rents a shed inside the detergent producer's plant, so as to abate transportation and financial costs and to keep a real-time quality control on the blowmoulding process. Thirdly, the environment has a marketing appeal. This shows in the marketing policy of both raw material producers and packaging users who happen to have environmental claims to make. British Steel is an obvious example; low collection costs and low (relatively to aluminium) energy content of steel are being banked on by the company, which is trying to stop the loss of market share to aluminium. Several companies take environmental communication seriously enough to fund education programs. This seems to be at least compatible with a demand-pull theory of environmental innovation in packaging. In fact, some companies maintain their environmental policy is totally consumer-driven, and well ahead of government regulation. Taken together, the common features of environmental innovation in packaging seem to draw a picture familiar to industrial economists. Its incremental nature seem to fit well Giovanni Dosi's [1984] theory of technical change. He characterizes technology in terms of technological paradigms or technological research programs. In analogy with Kuhn's [1975] scientific paradigm, a technological paradigm is essentially a body of scientific and technical knowledge used to satisfy a need. "Normal" innovative activity focuses on improving only some of the characteristics displayed by a certain technology, which Dosi calls a technological trajectory. Economic variables play a focusing role, calling the researchers' attention on some characteristics, ignoring the other. For example, the oil shocks of the 70s set the internal combustion engine paradigm on an energy saving trajectory; a low fuel consumption was recognized to "good" even before the oil shocks, but it wasn't important enough to actually work on. Nowadays, energy efficiency improving is part of the standard research program of engine designers. The rise of green consumerism and environmental regulation do not seem to be inducing a change of paradigm in the packaging industry. On the other hand, they seem to be quite successful in steering the existing paradigms on a whole range of environmental impact reducing trajectories, which were described as "innovation strategies" in this chapter. In passing, one may note that incremental innovation is much more easily market-driven than paradigm shifts. The notion of technological trajectory, then, fits in well with the marketing appeal of environmental issues described above. Keith Pavitt's [1984] notion of "specialized supplier" relationship between innovator and innovation user in certain industries also springs to mind when investigating the packaging filiére. It is our conviction that vertical disintegration plays an important role in making the technology so fluid, liable to be steered smoothly on to "green" technological trajectories. On the one hand, it minimizes sunk costs; when Barilla wanted to replace PVC trays with paper trays it simply cancelled the order to a plastic

23 converter (who continues, however, to supply it with polypropylene sacks) and placed a new one on its supplier of paper packaging. This happens all the time, and opens up roads for nonprice competition between "packaging paradigms" (in this case, plastic trays or paper trays), and therefore packings producers across paradigms. On the other hand, as it is well known [Williamson, 1975], the market alone has problems carrying "weak" signals like those that lead to an innovation; co-ordination is often needed. The packaging filiére solves this problem with various forms of what was called "quasivertical integration" above. The Dosi and Pavitt contributions, taken together, seem to provide a reasonable, if highly stylized, picture of environmental innovation in the European packaging filiére. A further, relevant question on which theory may help cast light is that of the "specialization" vs. "superproduct" nature of innovation. In a series of papers, dedicated to R&D races in multidimensionally differentiated markets, David Ulph [1993, Ulph and Owen 1993] argues that, under very general R&D cost assumption, and given that imitation is easy and technological lags can be caught up with , firms will aim for a "superproduct", superior to its competitor in all respects. Both assumptions are fairly reasonable to make with reference to the packaging filiére. The evidence is here less clear: certainly some leading firms seem to be doing packaging innovation both along the environmentally friendly and along other technological trajectories, which seems to support Ulph's views. Accepting such a picture of environmentally friendly technical progress has far-reaching policy implications. The rest of this study is devoted to an econometric test of the hypothesis that this picture is true.

24

3.

THE

13.

The data

DETERMINANTS OF ENVIRONMENTAL PACKAGING : A DISCRETE - CHOICE MODEL

INNOVATION

IN

As pointed out in chapter 1, the research team collected 1407 packaging innovation awarded with prizes in the countries being con sidered or at the European level. Annexe 1 contains the main descriptive statistics; this section contains a brief outline of the characteristics of this population of innovative packs. About 100-120 prizes were awarded each year, from 1978 to 1993 (relative to entries dated from 1977 to 1992). The number of prizewinning packs making environmental claim has been constantly on the rise, from 1-2 in the late 70s to around 15 in the late 80s, reaching a maximum of 30 in 1992. Medium-sized and large firms are the largest contributors to innovations.Within the filiére, almost all innovators belong to the finished packings and component industry. The second most innovative sector is the raw materials one which, however, has produced no environmental innovations. Among the industries using innovative packs, the most important ones are obviously the grocery goods ones; the food industry is the single largest contributor, followed by the detergent and the beverage industry. The detergent industry is the single largest contributor to environmental innovation. The materials most frequently used are plastics (PE being the most popular one), paper and cardboard, steel, glass and aluminium, in this order. The materials pattern of environmental innovation is roughly the same as that of non-environmental innovation. 14.

The model

The world of packaging innovation is a multidimensional one. A pack has several characteristics that innovators can choose to improve, from its ability to protect the product to its aesthetic appeal; in Dosi's terms, innovators can choose to move along several technological trajectories. In the innovation database, this shows in the multiplicity of reasons given by juries for the awarding of their prizes. It seems appropriate to model an innovation decision with discrete-choice techniques. A dependent discrete-choice variable (CHOICE) was constructed in the following way. Value 1 stands for the decision to innovate on the reduction of the environmental impact on the pack. CHOICE takes value 1 whenever the main or one accessory motivation for the awarding of the prize is environmental friendliness. There is here an element of asymmetry: environmental innovations are defined by means of both main and secondary reasons for the awarding of the prize; other innovations are grouped according to the main reason only, once environmental ones have been taken out. This was done for technical reasons: the reduction of environmental impact is the main claim to a packaging award for only 35 packs, too low a figure to run a multilogit estimation. Value 2 is more complex to define. Loosely, we could say it stands for the decision to minimize the pack's production cost. CHOICE takes value 2 whenever the main reason for the awarding of the prize is "cost reduction", "use of new materials" (generally introduced on cost effectiveness grounds), or other minor reasons.

25 Value 3 stands for the decision to minimize the logistical and distribution costs. CHOICE takes value 3 whenever the main reason for the awarding of the prize is "distributor friendliness" or "product protection"; these innovations are aimed at the optimization of exposition space in supermarkets, or of truck space, or at reducing the chances of damaging products during transport. Value 4 stands for the decision to enhance the pack's user friendliness. CHOICE takes value 4 whenever the main reason for the awarding of the prize is "user friendliness"; easy-open cans, no-spill lids and so on. Value 5 stands for the decision to improve the pack's attractiveness to the consumer. CHOICE takes value 5 whenever the main reason for the awarding of the prize is "aesthetics". Table 7 shows innovations grouped by main reasons for the awarding of the prize and by values assigned to CHOICE. TABLE 7 - I NNOVATIONS BY MAIN REASON FOR AWARDING THE PRIZE AND VALUE OF CHOICE Low environmental impact Cost reduction Use of new materials Others Distributor friendliness Packing previously unpacked products Product protection User friendliness Aesthetics TOTAL

35 148 44 17 233 20 329 280 238

CHOICE=1

110

CHOICE=2

209

CHOICE=3

538

CHOICE=4 CHOICE=5

260 227

1344

TOTAL

1344

Asymmetries apart, CHOICE has a potentially serious shortcoming as an innovationmodelling device; it implicitly assumes that technical trajectories are mutually exclusive. The reality is quite different, as the discussion in chapter 2 should have pointed out, and often R&D efforts lead to packs that are better than the ones they replace in more than one characteristics. This suggests that a binary choice approach might be more appropriate than a multiple choice one to model innovation. Nevertheless, estimation with a dichotomous version of CHOICE (ENV=1 iff CHOICE=1, 0 otherwise) does not yield significantly different results; furthermore, multiple choice models allow comparisons between groups of innovations; for example, the string of parameter characterizing "green" innovation can be tested for statistically significant difference from "cost reduction" innovation. The rest of this chapter is therefore devoted to the presentation and interpretation of multiple choice models. The estimation technique employed is the multinomial logit. The model is of the kind prob(1) =

ex

1

1 + ex 1 + e x 3 + e x 4 + e x

5

where prob(1) stands for the probability that CHOICE=1 and 1 for the set of coefficients attached to the vector x of explanatory variables when CHOICE=1. CHOICE=2 is chosen as the base alternative.

26 Regressors are divided into four groups. The first one, from ITALY to UK refers obviously to the country of the innovator; the second one, FOOD to SERVICES, refers to the industries using innovative packs; the third one, from ALUMINIUM to PVC + OTHER PL. to the materials used; the fourth one, composed of PRESENTATION and BOTH, to the function performed. All these variables are dummies; NMATERIALS, the number of materials used, INNUMBER, the total number of innovation done by the innovative firm (at the group level), and TIME (a variable set to 1 in year 1978, 2 in 1979 and so on) are continuous variables. Table 8 summarizes the meaning of setting all dummies of the same groups, to zero, i.e. which countries, industries, materials and function are incorporated within the constant term. TABLE 8 - COMPOSITION OF THE CONSTANT TERM IN COUNTRY Country of the innovative firm

All but D, F, UK, I, NL

Industry in which the pack is used

Agriculture, clothes and textile, petrochemical, pharmaceutical, hardware, mechanics, automobiles, miscellaneous

Materials

Paper and cardboard, wood

Functions

Transport

It is very important to keep in mind that parameters do not measure absolute innovativeness, but the influence on CHOICE given that one is doing an innovation anyway. 15.

Results

This section is devoted to the presentation of the multiple choice model section 13 has recognized the need for. Estimates for parameters and tests are simply presented; any attempt of interpretation is left to the following sections. In fact, this section can be skipped altogether by readers who do not find econometrics exciting. Table 9 summarizes the values of the parameter computed under the specification of the model we call COUNTRY, for reasons that will soon become clear. Regressors regarding firm size and position within the filiére are dropped; this allows the model to run on 1337 observations, thus recovering the possibility of adding regressors regarding the country of the innovator. Coefficients that are significant at the 90% significance level are marked by an asterisk; coefficients that are significant at the 95% significance level by two asterisks. This notation will be held on to throughout the rest of the paper.

27

TABLE 9 - THE COUNTRY MODEL

Regressor Italy France Germany UK food dairy beverages chemicals detergents services aluminium glass metals PE PS PET composites PVC+other pl. presentation both innumber nmaterials time constant term Log-likelihood

Number of obs chi2(96) Prob > chi2 Pseudo R2

1337 757,90 0 0.1941

DIST. COSTS

US. FRIEND.

AESTHETICS

(CHOICE=1) (CHOICE=3) 0,7483295 -0,067276 -0,4272889 -0,0613709 -0,5115607 -0,7840004 -0,3592673 -0,125181 0,3252054 0,0368444 -0,0847592 -0,5940911 0,6441551 0,2622063 0,7326784 0,1146932 0,4910869 -0,802351 0,6246732 -0,2307875 -0,0051769 0,4538976 0,3268512 **-0,9911177 -0,8442857 -0,34878814 0,5925449 0,2071943 -0,7922398 0,0124827 1,220964 -0,4134379 -0,2095655 *-0,5885556 -0,1362668 -0,0363005 0,2819869 **-0,7249768 -0,5283439 0,1734307 0,0015192 0,0064908 0,2514121 **0,5265409 **0,1922828 **0,466248 **-3,209503 0,4550341

(CHOICE=4) -0,0646024 0,026313 -0,4568147 0,8358023 -0,1630353 -0,7371387 -0,0267376 -0,1008564 -0,3792353 -0,2513948 **1,284311 0,2047462 0,5825683 **0,817034 -0,1646754 0,6907529 -0,4312786 0,6411559 **1,483126 **1,167096 0,0005184 0,0966031 0,0268966 **-1,57022

(CHOICE=5) -0,9767713 -0,3177113 **-2,25486 0,6556234 **0,687657 0,2767336 **1,077684 -0,2011145 **1,336627 -0,4593833 -0,5619221 0,5015573 -0,2806713 -0,3975288 -0,57086 -0,9832869 **-1,25944 *-0,529067 **4,806585 **3,400975 -0,0087633 **0,823733 *0,0459294 **-5,26455

Base case

CHOICE=2

ENVIRONMENT

-1573,1132

Two separate sets of tests were run on the parameters characterizing COUNTRY. The first one concerns the joint significance of all parameters for each value of CHOICE. In particular, four F-tests were run on the null hypothesis that i

=

1

Their results are summarized in Table 10.

i = 2,3,4,5

28

TABLE 10 - COUNTRY: JOINT SIGNIFICANCE OF PARAMETERS DIFFERENCE FROM ENVIRONMENTAL INNOVATION (CHOICE=1) OF INNOVATION OF OTHER KINDS

= 3 = 4 = 5 = 2

1

KO** KO** KO** KO**

1 1 1

The second set of tests concern the joint significance of groups of parameters across all different values of CHOICE. The null hypothesis is that, for each group j of parameters 1 j

=

2 j

=

3 j

=

4 j

=

5 j

=0

Table 11 summarizes the results. TABLE 11 - COUNTRY: TESTING FOR ZERO VALUE OF GROUPS OF PARAMETERS ACROSS VALUES OF CHOICE Country

KO**

Industry in which the pack is used

KO**

Materials

KO **

Functions

KO **

NMATERIALS

KO **

INNUMBER

OK

TIME

KO**

This model of technological trajectory choice in packaging innovation can be used to test the appropriateness of the Dosi-Pavitt-Ulph framework spelt out in chapter 2 to describe technical change in this filiére. In order to do so, that framework must be translated into testable hypothesis; such a translation, and a test of the translated hypothesis, is the subject of the following sections. There is, however, one exception. That innovation occurs along a technological trajectory rather than through a paradigm shift is obviously not picked up by the database described in chapter 1, and therefore it cannot be tested for. This statement, however, has quite solid nonstatistical evidence standing for it: it is hard not to notice a paradigm shift when one occurs. This point will not be dealt any further with.

29 16.

Interpreting results: technology and demand effects on the innovative process

One of the hypothesis generated by the nonstatistical evidence presented in chapter 2 is that environmental innovation should display demand-pull, as opposed to technologypush, characteristics. If this were true, we would expect that: 1.

environmental innovation, when compared to production cost reduction innovation, should be more concentrated on presentation packaging than on transport packaging. Consumers hardly ever see transport packaging, so the marketing appeal of an environmentally friendly shrinkwrap film for pallets is likely to be close to zero. In this respect, environmental innovation should behave more or less like aesthetic improvements. In terms of the model, this means testing that the coefficient estimates on PRESENTATION and BOTH when CHOICE=1 are positive and significant. A further, useful test is to compare the values of the coefficient estimates of environmental innovation (CHOICE=1) are similar to those of aesthetic innovation (CHOICE=5), which is very likely to be market-pull in nature.

2.

environmental innovation, when compared to production cost reduction innovation, should be more concentrated on grocery good industries. In those industries, as it is well known, packs play a strategic role in the purchase decision. In terms of the model, this means testing that the coefficient estimates on variables representing grocery goods industries are positive and significant.This is, however, a much weaker test than the former, because the role of environmental issues in determining purchase patterns vary quite a lot across grocery good markets. One feels the need for a deeper investigation in each industry before accepting econometric results as proof of such a hypothesis.

The results do not seem to provide strong support for a consumer-driven theory of environmental innovation in packaging. A glance at table 9 shows that the coefficient on PRESENTATION packaging is indeed positive, but not statistically significant, whereas the coefficient on the dummy representing packings that serve both a presentation and a transport purpose (BOTH) is negative (recall that the base is transport packaging). The parallel we tried to draw with aesthetic improvements is rejected by the data; when CHOICE=4 the coefficients on PRESENTATION and BOTH are positive, high and significant at the 99% significance level. Grocery good industries do not seem to be specializing in environmental innovation either. Coefficients on FOOD, BEVERAGES and DETERGENTS are positive as expected, but the one on DAIRY is unexpectedly negative. None is significant at the 90% (nor, indeed, at the 80%) level . Again, aesthetic innovation behaves differently; all coefficients on grocery goods industries are positive, and all but DAIRY are significant at the 95% level. Coefficients on other industries are negative. This regularity seems to point to an important role played by pack design in determining the purchase decision in supermarkets; obviously the pack's environmental friendliness is not as effective in this respect. If demand-pull factors do not seem to influence environmental innovation any differently than production cost reduction innovation, it is worth it to test for technology-push effects. In the innovation database, the raw materials variables can be interpreted as innovation opportunity variables; comparing their coefficients when CHOICE=1 should give an idea of the extent to which some materials specialize in green innovation more than others. With respect to the constant term, which incorporates paper, cardboard and wood, the coefficient estimates on GLASS, PE and PET are positive; those on ALUMINIUM,

30 other METALS, PS, COMPOSITES and PVC+OTHER PLastics are negative. None is statistically significant at the 90% level, although the coefficient on PET, a plastic that has a reputation for being environmentally friendly, is significant at the 85% level. The conclusion is that the environmental innovation opportunities offered by packaging materials do not seem to be radically different from their production cost reduction opportunity. These results do not, by themselves, allow the researcher to label environmental innovation in packaging as demand-pull, nor as technology-push. In fact, the whole database only allows to draw comparisons between innovations along different trajectories, whereas innovation economics has concentrated mostly on the problem of whether to innovate. Data do, however, yield some insight into the matter. Nonstatistical evidence for the demand-pull nature of at least some environmental innovation is quite robust, and nowhere contradicted by data. Nonsignificance of coefficient estimates on industry variables certainly does not, by itself, deny it. The detergent industry is an example of an industry where firms anticipate environmental regulation in order to compete for market share. The fact that the coefficient estimate on DETERGENTS is not highly significant in the model should not come as a surprise: in a market where price elasticity of demand is high, innovation that abates production costs also provides innovative firms with a weapon for market share competition. What the model does, then, is simply comparing two different trajectories along which firms face the same sort of incentives. The nonsignificance of the coefficient estimate on PRESENTATION cannot deny that some environmental innovation is demand-pull either; what it does say is that there is at least some environmental innovation, that done on transport packaging, that isn't demand-pull. This, together with the significance of the coefficient estimate on TEMPO, can be interpreted as an indirect proof of the existence of a regulation effect. On the other hand, it must be kept in mind that most environmental innovation, as most packaging innovation in general, is done on presentation packaging; transport packaging (the "certainly not demand-pull" component) only accounts for 12% of total green innovation. 17.

Interpreting results: the role of firm size and position within the filiére

The degree of intra-filiére "specialized supplier" cooperation in running innovative activities is another variable that is not directly picked up by the data. On the grounds of rather robust nonstatistical evidence, it is safe to state that such degree is high. An interesting question, then, is whether this relationship holds in environmental innovation as well as in innovation in general. In specialized supplier relationship the innovator is a smaller firm than her customer, and she is positioned just upstream of the latter in the filiére. In econometric terms, this means testing for the significance of coefficient estimates on variables representing the position within the filiére when CHOICE=1. Unfortunately, we cannot do it with the COUNTRY model, which includes no such variables among its regressors. It has proved much more difficult than expected to attach a number of employees to firms in the packaging filiére. The Italian Packaging directory is the only source to report this piece of information for all firms listed; the number of employees of the UK top 300 firms is also reported, in a separate directory. However, it was not possible, within the time and budget limits of this paper, to complete French and German records, as well as the majority of British ones. Similar problems, especially in the case of German firms, were met to ascertain the position within the filiére of innovative firms. In order to assess the impact of firm size and position within the filiére variables on CHOICE, we have built a different specification

31 of the model, and fed to it the 384 observations for which both the number of employees and the position within the filiére is known. Since most of these describe either Italian or British innovations, it is not possible to include dummy explanatory variables for the countries. Table 12 presents the results. TABLE 12 - THE SIZE MODEL Number of obs chi2(96) Prob > chi2 Pseudo R2 Regressor

ENVIRONMENT

food dairy beverages chemicals detergents services aluminium glass metals PE PS PET composites PVC+other pl. presentation both innumber nmaterials time emp 100-250 emp > 250 packs machines user constant term

(CHOICE=1) 1,609978 1,733797 *1,852976 1,828322 1,52033 **2,890611 1,820642 1,281806 0,4572033 **2,580927 0,7494971 1,849628 0,213209 0,6068378 0,868859 -0,7117254 -0,0341353 -0,4577608 **0,1305372 0,3348567 -0,2734245 -0,277146 0,7757324 -0,1672445 **-3,413482

Log-likelihood

-423,53442

384 282,27 0 0.2499

DIST. COSTS US.FRIEND. AESTHETICS (CHOICE=3) -2,247743 -0,5298077 -0,1019609 -0,51955181 **-1,620469 -0,1797892 0,4074697 -1,041754 -0,1064408 0,0090434 0,0003121 -0,1550363 -0,1140569 -0,5703742 -0,5210054 0,2369865 -0,0087868 **0,8627923 0,0401625 0,3638188 0,0453242 -0,9251709 -0,957263 -0,4589985 0,132658

(CHOICE=4) 0,0999075 -1,859202 -0,5671713 -0,3674815 -0,4611052 0,6017028 1,619598 0,129624 0,5322105 0,9999107 1,103055 -0,2274608 -0,1778426 0,7382443 **2,019392 **1,939191 -0,032014 0,08258 0,0713587 0,781951 1,027991 -0,6889066 -1,582167 -0,4136654 *-2,287443

(CHOICE=5) 1,089407 1,194231 0,919873 -0,2370626 1,11425 -0,7675196 -1,826852 0,496791 0,0419431 -0,1352663 0,6220359 -1,597753 -0,7624792 -0,6557542 **3,351634 **2,33546 -0,015778 **1,041066 0,0711953 0,7843987 0,2203095 -1,062358 -1,107866 -0,9300377 **-4,11333

Base case

CHOICE=2

Regressors are now divided into five groups. The first one, from FOOD to SERVICES, refers to the industries using innovative packs; the second one, from ALUMINIUM to PVC + OTHER PL. to the materials used; the third one, composed of PRESENTATION and BOTH, to the function performed; the fourth one, composed of EMP 100-250 and EMP > 250, to the number of employees; the fifth one, from PACKS to USER, refer to the position of the innovator within the filiére, as defined in chapter 1. As with COUNTRY, all these variables are dummies; NMATERIALS, the number of materials used, and INNUMBER, the number of innovation done by the same firm (at the group level) as the innovative firm, are continuous variables.

32 Table 13 summarizes the meaning of setting all dummies of the same groups to zero, i.e. which industries, materials, function, firm size and position within the filiére are incorporated within the constant term.

TABLE 13 - COMPOSITION OF THE CONSTANT TERM IN SIZE Industry in which the pack is used

Agriculture, clothes and textile, petrochemical, pharmaceutical, hardware, mechanics, automobiles, miscellaneous

Materials

Paper and cardboard, wood

Functions

Transport

Number of employees

1-100

Position within the filiére

Raw materials, services

Two separate sets of tests were run on the parameters characterizing SIZE. The first one concerns the joint significance of all parameters for each value of CHOICE. In particular, four F-tests were run on the null hypothesis that i

=

i = 2,3,4,5

1

Where i is the vector of all parameters attached to variables when CHOICE=i . Their results are summarized in Table 14. TABLE 14 - SIZE: JOINT SIGNIFICANCE OF PARAMETERS DIFFERENCE FROM ENVIRONMENTAL INNOVATION (CHOICE=1) OF INNOVATION OF OTHER KINDS

= 2 = 3 = 4 = 5

1 1 1 1

OK KO** OK KO*

What Table 14 is saying is that environmental innovation cannot be distinguished from "cost reduction and miscellaneous" innovation (CHOICE=5, row 1) and "user friendliness" innovation (CHOICE=3, row 3) on the basis of the explanatory variables included in the model. On the other hand, innovation of the "distributor friendliness" kind (CHOICE=2, row 2) and of the "aesthetics" kind (CHOICE=4, row 4) are both significantly different from environmental innovation. The second set of tests concern the joint significance of groups of parameters across all different values of CHOICE. The null hypothesis is that, for each group j of parameters

33 1 j

=

2 j

=

3 j

=

4 j

=

5 j

=0

Table 15 summarizes the results. TABLE 15 - SIZE: TESTING FOR ZERO VALUE OF GROUPS OF PARAMETERS ACROSS VALUES OF CHOICE Industry in which the pack is used

OK

Materials

KO *

Functions

KO **

Number of employees

OK

Position within the filiére

OK

NMATERIALS

KO **

INNUMBER

OK

TIME

OK

A glance at table 12 shows that none of the coefficient estimates on firm size (EMP 100-250 for firms with 100 to 250 employees and EMP > 250 for firms with more than 250 employees) or position within the filiére (PACKS for sector 2, MACHINES for sectors 4 and 5, USER for packaging users) is significant. Table 15 confirms that this is true across all values of CHOICE, and not only for environmental innovation. These results, combined with the nonquantitative evidence presented above, allow to draw quite strong a conclusion; the packaging filiére tackles and solves all innovation problems in the same way. The specialized supplier approach to innovation, as explained in chapter 2, is to build long-term alliances of packaging users and packaging manufacturers, where the latter solve technical problems chosen by the former. Interviews suggest that, as environmental concerns became more important, the very same user-supplier teams that had successfully innovated along other technological trajectories applied their expertise to the new problem. Data are fully compatible with the existence of a specialized supplier relationship: well above 50% of both environmental and non-environmental innovation is done by firms with less than 500 employees; about 15% by firms with less than 100. Figure 5 reports the distribution by class of employees of the innovator referred to environmental and non-environmental innovations.

34

FIGURE

5 - NON-ENVIRONMENTAL AND ENVIRONMENTAL INNOVATIONS BY NUMBER OF EMPLOYEES OF THE INNOVATIVE FIRM .

50,00 45,00 40,00 35,00 30,00

NON-ENV

25,00 ENV

20,00 15,00 10,00 5,00 0,00 1-25

26-100

101-500

> 500

A reasonable interpretation for the coefficient estimates and figure 5 is that the economies of scale in innovative activities are very similar in environmental innovation and non-environmental innovations, and that they are not so strong as to prevent small firms to do some innovation. 18.

Interpreting results: packaging innovation as a superproduct quest

The results discussed so far suggest that the profiles of the typical environmental and non-environmental innovators coincide. Variables representing the industry using innovations, materials used, innovator size, innovator position within the packaging filiére do not seem to make any significant contribution to explaining the innovator's decision to do an environmental innovation rather than an innovation of another kind. The issue of whether R&D competition in the packaging filiére resembles more a quest for specialization or one for a superproduct, however, has not been addressed. Tables 16 and 17 present a list of the most innovative firms, respectively in nonenvironmental and in environmental innovation. They show that some leading firms in innovation in general also lead in environmental innovation.

35

TABLE 16 - NON-ENVIRONMENTALLY MOST INNOVATIVE FIRMS BY NUMBER OF ENVIRONMENTAL AND NON-ENVIRONMENTAL INNOVATIONS, 1978-1992 Top non-env innovators Carnaud Metalbox SOCAR Saint-Gobain Reed Corrugated 4P CEBAL Rochette Europa Carton Zewawell Beghin-Say BSN Imprimerie Rey Gustav Stabernack Lembacel

Env

Non-Env

3 0 2 0 3 1 2 3 0 0 4 0 0 0

47 45 33 26 18 18 17 12 12 12 11 10 10 10

TABLE 17 - ENVIRONMENTALLY MOST INNOVATIVE FIRMS BY NUMBER OF ENVIRONMENTAL AND NON-ENVIRONMENTAL INNOVATIONS, 1978-1992 Top env innovators BSN Unilever Carnaud Metalbox 4P Europa Carton Saint-Gobain Rochette Carl Edelmann SCA Cartotecnica Poligrafica Bianchi Guala Mauser Werke Aviocart ISEA

Env

Non-Env

4 4 3 3 3 2 2 2 2 2 2 2 2 2

11 0 47 18 12 33 17 5 5 3 3 2 0 0

Table 16 includes firms with 10 or more non-environmental innovations; table 17 all firms with 2 or more environmental ones. Of the 14 top environmental innovators, 6 firms are also top non-environmental innovators; 3 have only done environmental innovations; the remaining five, while doing some non-environmental innovation, seem to have devoted an unusually high share of their innovative effort to reducing the environmental impact of their packs. It can safely be concluded that at least some firms are aiming at what Ulph [1993, Ulph and Owen 1993] calls full dominance in multidimensionally differentiated markets. The outcome of a series of such multidimensional R&D races depends on the structure of

36 R&D costs. In general, constant or increasing returns to R&D tend to be associated with the persistence of full dominance, with one firm leading in all characteristics forever; decreasing returns allow for a wider range of outcomes, including one in which different firms "specialize" and take the lead in one each of the R&D dimensions; the existence of specialized innovative firms in the packaging filiére may be linked to a decreasing R&D cost curve. It can be concluded that Ulphs's model seems to provide a rationale for the strikingly similar features of environmental and non-environmental innovation in the European packaging filiére. We have tried to show that packaging producers and packaging users have developed an effective approach to the solution, by way of vertical teamwork, technical problems. This approach has led to the development, over the years, of a considerable command of the technology (in the broad sense of technological paradigm) involved; this expertise is now being put to work on the solution of the environmental problems associated with packaging. 19.

Interpreting innovation

results:

private

optimality

of

environmental

It is worth it to look back at one side-result of table 14 in the light of the discussion on environmental innovation strategies. What the first row of table 14 is saying is that environmental innovation, as picked up in our databes, is microeconomically indistinguishable from cost-reduction innovation. In fact, two of the innovation strategies described in chapter 2, elimination of overpacking and lightweighting, are , as a rule, cost-reducing as well as environmental impact reducing. Lightweighting is possibly the innovation strategies most enthusiastically supported by the industry experts and companies spokespersons the research team has talked to. So, table 14 suggests that the typical environmental innovation is no-regret; it saves money as well as natural capital. Even assuming out operations of window-dressing cost-saving innovation as "green", one can safely conclude that most environmental innovation is privately optimal as well as being beneficial for the environment. 20.

Interpreting results: the role of regulation

It is very difficult to measure directly the impact of environmental regulation on CHOICE. A first difficulty is that the new wave of packaging regulation, that led to the setup of national recycling systems in Germany and France, is very recent, whereas the most recent data are those relative to 1992. TIME should pick up a tightening of environmental regulation in packaging, but it is likely to capture other phenomena, like the rise of green consumerism, as well. So, even though the coefficient on TIME is positive and statistically significant when CHOICE=1, it cannot automatically be concluded that regulation played an important role in inducing environmental innovation in the period considered. It would be of great interest to check this conclusion in five years time, after new regulation promoting packaging recycling in Germany, France and the UK has had time to affect R&D departments' work. In Germany, the country were packaging-specific regulation is oldest, there are signs of a shift away from plastics on to more easily recyclable materials; a shift, indeed, that can be thought of as a general equilibrium adjustment to the new DSD tariff policy reducing the degree of mutualization of recycling costs across materials. As pointed out by Glachant [1994], regulation on packaging in Europe is happening through the setting up of consortia which collect, via a system of levies, money to fund separate collection and recycling of packaging waste. The degree of mutualization of both collection and - above all - recycling costs seems to be a key variable in understanding the ways in which regulation affects innovative activities. In a few years time, it will be possible to regress the number of innovation

37 made of any one material in Germany on the DSD tariff on that particular material; the sign and significance of the coefficient estimate will yield a great deal of information about the effectiveness of this particular kind of environmental policy in stimulating environmental innovation. 21.

Conclusions

The environment has become an important issue for packaging, which contributes greatly to the generation of domestic waste. The appearance of a share of consumers willing to buy low-environmental impact goods on the one hand, and high political visibility on the other hand are the two main sources of incentives to industry to innovate their packaging along an environmentally friendly technological trajectory. Despite the enlightened wing of industry's claims that "pollution prevention pays", the latter source of incentives seems to have been more important than the former, except perhaps in Germany. Environmental innovation happens by applying to the reduction of the environmental impact of packaging the same across-filiére innovative teams that have done (and still do) non-environmental innovation in the past. By virtue of a large population of innovative firms, of a relative easiness of imitation, and of an extremely flexible technological paradigm, a relatively large number of environmental innovations were launched on the European market in the period considered. However, despite some successes, there is a feeling, widespread among politicians and environmentalists, that these problems are not being tackled aggressively enough. Industry experts remark that there is no commonly accepted way to assess the overall environmental impact of packaging, and that there is not even consensus on the most environmentally friendly packaging waste options. This contributes greatly to divert the innovative effort from environmental innovation, except for "no-regret" innovation strategies like lightweighting. It is not by chance that our models find a close resemblance between innovation on production cost reduction and environmental innovation. Environmental policy can play a key role as, in Rosenberg's words, a focusing device; it can give (and it often has) the industry's technological expertise a direction. Some firms, for example, are already modelling their packaging policy on the "ladder" of waste management option adopted by the EU Packaging Directive draft, before the Directive is even approved. The debate on the environmental policy on packaging in Europe, especially when it comes to levying collection and reprocessing fees, has often been in terms of static allocative efficiency; one wonders whether the "signalling impact" of environmental regulation on the directions of technical problems shouldn't be considered more carefully. The "new wave" of environmental policy making in Europe, leading to the setting up of consortia for recycling packaging waste, is too recent to have influenced the data prersented above. However, applying the same research methodology used here to innovative packings presented after these consortia started to operate will enable researchers to evaluate the impact of this kind of environmental policy on innovative activities.

38

REFERENCES Alaimo, A. and V. Capecchi, 1992, "L'industria delle macchine automatiche a Bologna: un caso di specializzazione flessibile", in P. P. D'Attorre and V. Zamagni, 1992, Distretti, imprese, classe operaia, Franco Angeli, Milano Capecchi, V., 1993, personal communication Dosi, G., 1984, Technical Change and Industrial Transformation, MacMillan, Basingstoke (UK) Dosi, G., 1988, "Sources, Procedures, and Microeconomic Effects of Innovation", Journal of Economic Literature ETAS, 1992, Repertorio professionale dell'imballaggio, Etas Periodici, Milano Glachant, M., 1994, Adoption and implementation of a voluntary agreement: the case of packaging recycling consortia, ERIC paper, Paris Kuhn, T., 1963, The Structure of Scientific Revolutions, Chicago Press

Chicago, University of

Morino, M., 1992, "Imballaggio: il fatturato '92 supererà i 22mila miliardi", in Il Sole 24 Ore, n. 145 Pavitt, K., 1984, "Sectoral Patterns of Technical Change: Towards a Taxonomy and a Theory", Research Policy Rosenberg, N, 1976, Perspectives on Technology, Cambridge, Cambridge University Press Scherer, F., 1982, "Demand Pull and Technological Invention: Schmookler Revisited", Journal of Industrial Economics Townsend, J., F. Henwood, G. Thomas, K. Pavitt and S. Wyatt, 1981, Innovations in Britain since 1945, Occasional Paper n. 16 (SPRU, University of Sussex) Ulph, D., 1993, "Racing to Introduce an Environmentally Friendly Product", paper presented at the Workshop on Environmental Policy, Oslo 17-18 September 1993 Ulph, D., and R. Owen, 1993, Racing in Two Dimensions, mimeo Williamson, O., 1975, Markets and Hierarchies: Analisys and Antitrust Implications, New York, The Free Press