---

III. Regulating Product Safety

---

10. Licensing Biologics in Europe and the United States

Martine Kraus

We got approval in the U.S. and we thought: If we can get it approved in the U.S., we can get it approved everywhere; the U.S. has the highest standard. That was a great mistake.

REGULATORY AFFAIRS OFFICER, MULTINATIONAL PHARMACEUTICAL COMPANY

I. INTRODUCTION

The conventional wisdom, based on a number of cross-national comparative studies, is that American regulatory processes are much more legalistic, adversarial, and costly than those of European nations, where more informal, negotiated policy making and implementation prevail and recourse to courts is rare.^[1] In the 1970s, for example, the U.S. Food and Drug Administration (FDA) was criticized for its legalistic conservatism, which, in comparison with other countries, led to higher costs and longer delays in the introduction of beneficial new drugs.^[2] FDA enforcement of Good Manufacturing Practices (GMP) regulations was described as legalistic and inflexible.^[3] cross-national comparative studies also typically indicated that American regulatory processes were much more complex and organizationally fragmented than their European counterparts.

During the 1980s and 1990s, however, political and bureaucratic pressures arose for regulatory reform and for cross-national harmonization of regulatory controls. The European Union (EU) adopted a multistate approval process for pharmaceuticals. The FDA announced its intent to harmonize new drug testing standards used in the United States with those in other countries,^[4] and to carry out joint reviews of important new therapies with the Canadian Health Protection Branch. Regulatory and company officials from the EU, Japan, and the United States founded the International Conference on Harmonization (ICH) to discuss the harmonization of pharmaceutical regulation and the mutual recognition of data among the three trading blocks.^[5] In the United States, the FDA has been subjected to powerful political pressures to accelerate its regulatory procedures. In 1992 the U.S. Congress enacted a user fee system tied to performance goals for the

---

FDA.^[6] An FDA advisory committee urged quick approval of a new class of promising AIDS drugs,^[7] and the Center for Biologics Evaluation and Research (CBER) hired three hundred new reviewers and introduced reviewer deadlines to hasten the approval process.^[8] Conversely, regulation at the transnational level in Europe has been criticized for becoming more legalistic than regulation at the national level. In the 1990s, as companies specializing in genetically engineered health products grew much more rapidly in the United States than in Europe, the disparity was widely attributed to the more favorable regulatory system in America.

In consequence, it is not entirely clear whether the national differences in regulatory style identified ten or fifteen years ago still persist in this policy area. To provide some perspective on this issue, this study investigates the regulation of biotechnology-derived “biological” products. Biologics are products made from or with the aid of living organisms, as opposed to pharmaceuticals produced through chemical synthesis. Biologics include human blood products or blood-derived products, vaccines, peptides, and carbohydrate products produced by cell culture. In all economically advanced democracies, the manufacture of biologics is closely regulated to guard against the risk of deadly viral and bacteriological contamination. Regulatory authorities in both the United States and the EU require manufacturers to obtain licenses both for their manufacturing facilities and for each new product, based on detailed descriptions and analyses of their facilities and quality control measures. This chapter compares the efforts of Company Y, a multinational manufacturer, to win regulatory approval from agencies in the EU and the United States for the manufacture of a medically valuable genetically engineered biological product.

II. RESEARCH METHOD

The research was based primarily on numerous, lengthy, in-person interviews with Company Y's director and vice president of regulatory affairs. These officials are responsible for obtaining regulatory approval of Company Y's products and manufacturing facilities in the United States and Europe. We discussed in detail the regulatory approval process for one of Company Y's genetically engineered blood products, because it was said to illustrate the characteristics of both regulatory systems. I asked the officials to describe the similarities and differences Company Y experienced with regard to the substantive requirements for regulatory approval, the amount of time it took to obtain approval, changes that had to be made to satisfy EU and U.S. regulators, and the cost to the company of the compliance and approval process. Company Y officials allowed me to review the voluminous documents submitted as part of the approval process and the responses they received from the regulating agencies, after which I was able

---

to ask follow-up questions, clarify any confusion, and confirm details discussed in interviews.

I also asked the Company Y officials for examples of either legalistic or cooperative interactions with both sets of regulators and reviewed FDA and European regulatory officials' responses to Company Y's submissions. Finally, I asked Company Y officials about their impressions of whether either system demanded a higher margin of safety or entailed higher “accountability costs” or “opportunity costs.” Company Y officials, who have had ongoing interactions with both sets of regulators, did not appear to have an “ax to grind” with either system; they are generally supportive of this form of product safety regulation, since the stakes for making a faulty product are extremely high. Thus I found their responses knowledgeable and credible.

To test the generalizability of Company Y's experiences, I conducted a series of telephone interviews with similarly situated officials in other multinational pharmaceutical companies.^[9] I presented these officials with a summary of the findings with respect to Company Y and asked them to comment, based on their own company's experiences. The results of this process, as well as my own subsequent experience working in the pharmaceutical industry and informal interactions at conferences and other gatherings of professionals in this field, give me confidence that the findings from Company Y's experience are accurate and typical.^[10]

III. SUMMARY OF FINDINGS

Based on Company Y's experience, in the 1990s the U.S. and European approaches to the regulation of biological products continued to diverge in several areas:

1. Despite increasing progress toward centralization, the European system, viewed as comprising both member states and EU agencies, was more institutionally complex and fragmented than the American system.
2. In certain respects, the European regime was as legalistic and more stringent than the American one. Although FDA regulatory interactions with companies had grown somewhat more flexible, both regulatory systems remained highly prescriptive and prone to regulatory unreasonableness. Both regulatory regimes insisted on certain regulatory requirements that Company Y's quality control officials regarded as unnecessary under the particular circumstances and costly to comply with. The European regulatory review process, however, was more analytically searching and less mechanical than the American one.

---

― 316 ―
3. There remained some differences in substantive requirements concerning manufacturing safety and reliability: the FDA focuses more on manufacturing facilities, the European authorities on process control analysis. Clinical study requirements still differ between Europe and the United States, as well as among member countries of the EU.
4. Both regulatory review systems imposed substantial delays in marketing new biologic products. A six-month wait for regulatory approval, once the product was ready for manufacture and marketing, was estimated to cost the company $48 million in revenues, while delaying benefits to patients eager for the product. In the case at hand, the application approval time by the EU Committee for Proprietary Medicinal Products (CPMP) (approximately twenty months) was substantially shorter than by the FDA (approximately thirty-one months). However, if we add delays stemming from individual country approval in Europe (an additional five months), overall differences in review times are less striking. Once manufacturing has begun, moreover, regulatory reviews of proposed process changes are much slower in Europe than in the United States.
5. Although both regulatory systems entail costly and burdensome transaction and opportunity costs, neither was clearly more cost-effective. The faster process change review in the United States was offset to an unpredictable degree by a higher rate of turnover and inexperience among regulatory reviewers at the FDA. In the case at hand, nonetheless, the cost of compliance was slightly higher in Europe than in the United States, because more tests were required and authorities insisted on tighter tolerances in manufacturing processes.

Generalizations from the particular case study must be qualified, however. The case studied reflected Company Y's first submission of a genetically modified product to the newly created EU authorities. Today, Company Y officials say, company and government officials alike would be more experienced, and the process might be significantly faster and less troublesome. Second, since the approval of the described product, the EU “concertation” procedure has been superseded by a more centralized procedure, which promises to streamline the process.^[11] More generally, the 1990s have been characterized by increased convergence of the two regulatory regimes. Product application review times between Europe and the United States are clearly converging and by 1999 had been reduced to less than 12 months, as agencies in both places strive to be more efficient.^[12] A recent EU directive promises convergence of the review times for process changes as well. For specified biotechnology and synthetic biological products, the FDA is moving away from an exclusive focus on the manufacturing facilities and toward a more European-style process control analysis.^[13] Conversely,

---

European authorities have increased inspections of U.S. manufacturing facilities. Finally, EU guidelines on clinical trials and the International Conference on Harmonization's focus on preclinical and clinical trial requirements have led to rapid convergence in the area of clinical studies.^[14]

IV. THE PRODUCT AND ITS CONTROL

All industrial democracies have established regulatory regimes to control the testing, manufacture, and marketing of pharmaceutical, biological, and medical device products. The harmful side effects of products such as thalidomide demonstrated that consumers and health care professionals cannot fully evaluate the safety of products that they use on a daily basis, and that liability law may not always provide sufficient additional incentives for manufacturers to test for and maintain the highest standards of quality control.^[15] Government regulation attempts to fill that gap for therapeutics by requiring extensive premarket research and testing concerning the safety, quality, and efficacy of new pharmaceutical and biological products, as well as controls on sanitary conditions and quality assurance during manufacture.

Governments impose special regulations on the manufacture and approval of biologics because, compared with pharmaceuticals, they are chemically less stable, more heat-and shear-sensitive, difficult to assay, and more subject to contamination. Contaminated intravenous medicines can rapidly spread disease and even cause death. In the early 1980s, many French patients became infected with the AIDS virus from contaminated blood banks. As recently as 1995, 350 people worldwide contracted hepatitis from a biological drug treatment for immune deficiency.^[16] Thus, in the United States, the regulation of biologics manufacture is more stringent than for pharmaceuticals; special attention is paid to the cleanliness of the production area, the operation of the equipment, and the control of the manufacturing process.

Genetically engineered blood products, employing recombinant DNA (rDNA), are less reliant on provisions of donor blood and provide a solution to the problem of viral contamination. The novel techniques, however, raise new questions. Do rDNA-derived products deviate from or contain contaminants not normally present in their conventional equivalents? In response, many countries have imposed special regulatory procedures for genetically engineered biologics and drugs. In 1985 the FDA issued Points to Consider in the Production and Testing of New Drugs and Biologicals Produced by Recombinant DNA Technology, which described the information needed to evaluate the expression system, host cells, and manufacturing procedures for products prepared by rDNA techniques.^[17] In 1987 the EU established a new mandatory evaluation procedure for all high-technology medicinal

---

products derived from biotechnology. In 1990 the EU issued two directives on the containment and release of genetically modified organisms. More recently, based on over ten years of experience with biotechnology, the FDA issued a new policy for specified biotechnology and synthetic biological products that can be clearly determined and controlled.^[18]

Development of a new chemical entity (NCE) or biological agent, including approval by the regulatory authorities, takes on average seven to ten years and entails a cost of $100 to $250 million. The regulatory process includes initial preclinical studies in animals, three phases of clinical studies in humans, and the submission of the license application to the regulatory authorities. While carrying out its clinical studies, the company operates a pilot manufacturing plant to provide material for the trials. Regulatory authorities closely monitor the company's methodology and test results, demanding proven results in each phase before allowing the company to move on to the next level of investigation. The regulators also request proof of proper operation of the pilot manufacturing plant.

A. Company Y's Recombinant Coagulation Factor

The product is a new antihemophilic factor, administered to hemophiliacs as an important cofactor for the coagulation of human blood. Until recently, the antihemophilic factor was produced from human plasma. Millions of liters of freshly donated blood plasma had to be obtained and processed every year. The new genetically engineered product, however, relies primarily on a constant supply of the cell culture from the manufacturer's cell bank and an ongoing fermentation process.

Manufacture entails three main steps—fermentation, purification, and finishing—each of which requires extremely high levels of purity and process control. The fermenter is operated, under sterile conditions, for up to six months, during which the culture fluid containing the recombinant coagulation factor is removed and fresh culture fluid is added. To assure the exact reproducibility of the recombinant protein, a variety of culture parameters are monitored and controlled.^[19]Purification entails a number of steps that increase the concentration of the coagulation factor, eliminate impurities, and remove or inactivate any viruses that may have been inadvertently introduced into the manufacturing process.^[20] The highly purified coagulation factor is then stabilized for longer shelf life by adding (plasmaderived) albumin. Regulatory authorities are interested in the capacity of the purification process to achieve the desired “clearance” (level of impurities), the manufacturer's justification for that level of clearance, and the consistency with which that level can be maintained in long-term production. All control tests, as well as the data from all preclinical and clinical

---

studies, must be described and analyzed in the company's product license application.

B. Why Regulation?

Manufacturers of biologics are part of an industry that has strong incentives for strict quality control. Biologics are easily contaminated, with serious consequences for the patient. No company can afford a widely publicized injury to a patient; the adverse reputational effects of putting an unsafe product on the market may hurt all of a company's products and create shortages for patients. In addition, the industry operates under strict liability law and is susceptible to multiple million dollar lawsuits. The industry competes not only on the basis of price but also on the basis of quality, as the products are paid for primarily by a vast array of health insurance arrangements. Finally, any scandal in the new biotechnology industry will affect the entire industry. In that sense, manufacturers of biological and particularly genetically engineered products are hostages of each other; a public health disaster or panic caused by quality control weakness in one company could conceivably destroy trust in the product and ruin the market for all producers.^[21] Hence, even aside from product regulation, corporate quality assurance officials in the manufacturing companies are granted powers to prevent shipment and to demand further testing or process changes.

Nevertheless, government and company officials alike believe that government regulation and product license applications are not redundant. Even if most firms are very conscientious, companies with unreliable safety controls may end up hurting a consumer, and thus the reputation of the entire industry. In addition, company officials point out that regulatory agencies review many applications and are in a position to compare, recommend, or require the most effective quality assurance procedures, pushing the industry toward higher standards. Similarly, cGMP (current Good Manufacturing Practices) compliance inspectors provide an extra set of eyes that help prevent and identify deficiencies in company manufacturing controls. This is particularly important when introducing new technologies with yet unknown risks. Finally, company officials claim that regulation increases consumer confidence and legitimizes the end product.^[22]

V. COMPARING U.S. AND EUROPEAN REGULATIONS A. Clinical Trials

In the United States, the FDA's approval decision for a new drug is based largely on data derived from clinical studies. To test a new chemical

---

compound or biological agent in humans, a company has to file with the FDA an Investigational New Drug (IND) application that includes (1) the results of all preclinical studies and an analysis of their implications for human pharmacology; (2) an analysis of the drug's composition, and the manufacturing and quality control procedures used in its production; (3) the clinical protocols describing the sponsor's plans for the initial-stage clinical studies and the qualifications of the investigators. Generating the data required by the FDA to show that a drug is reasonably safe for use in initial, small-scale studies takes on average 3.5 years.^[23]

If the company does not hear from authorities within thirty days, it can legally start its Phase 1 clinical trial as described in the IND. The FDA can put the trial on “clinical hold” until the company has provided further information. Clinical holds are common; thus, in the 1980s many companies conducted their Phase 1 clinical trials in Europe. This is less true today, following the efforts of the International Conference on Harmonization (ICH), but many U.S.-based companies still find it much faster to take their U.S.-discovered drugs straight to Europe and shift drug development back to the United States after completion of the first trials in humans.^[24]

In Europe, a company has to meet the regulatory requirements of the particular European country where it intends to conduct its trial. The EU has made some progress toward the harmonization of clinical trial requirements.^[25] Nevertheless, differences between member states persist regarding the need for government approval before initiating clinical trials; the status of ethics committees; their structure, function, and legality; the submission requirements; and agency review times prior to trial initiation.^[26] Even after the trials have been completed, member states differ in their insistence on inspection of investigational sites for Good Clinical Practice (GCP) compliance.^[27]

Phase 2 and 3 clinical trials entail successively larger and better controlled testing protocols, and they are much more expensive. To facilitate the worldwide introduction of new products, companies like to initiate parallel Phase 2 and 3 clinical trials in the United States and Europe, tapping into a U.S. and pan-European network of physicians and health organizations.

B. Manufacturing and Marketing

In order to manufacture and sell a new biological product in the United States, a company must apply for a product license and obtain an establishment license; in Europe, a company must apply for a marketing authorization and obtain a manufacturing permit. In both places the process is burdensome, costly, and at times slow. In America, the federal FDA's CBER^[28] monopolizes premarketing regulation of biological products, preempting

---

regulation by the states. CBER's decisions are not reviewable by other agencies and are rarely reviewed by the courts, since applicants do not want further delays and no advocacy groups routinely oppose the granting of licenses.

In Europe, before the establishment of the EU CPMP and the European Medicines Control Agency (EMEA),^[29] companies needed a separate product license in every European country in which they sought to sell their product. This was succeeded by a sequence of multistep procedures involving national and EU authorities, described more fully later. In Europe, as in the United States, the permit process rarely involves court proceedings or judicial review, although in Germany advocacy groups on occasion have challenged manufacturing permits for genetically engineered products in the courts, delaying the approval process.^[30]

1. COMPANY Y'SU.S. APPLICATION

The company's Product License Application (PLA) contained reports of all prior investigations and other information pertinent to an evaluation of whether the product, as manufactured, will be safe and effective. Applications are often between 50,000 and 250,000 pages, or hundreds of volumes. Company Y's original application for the genetically engineered biologic filled only eight volumes because the company was already marketing an FDA-approved non–genetically engineered coagulation factor and was able to cross-reference a former application. The company's Establishment License Application (ELA), describing its manufacturing facilities for the new product in question, was also eight volumes long. This was in addition to “its standing ELA”—an application of more than twenty volumes that provides detailed and technical descriptions of all the company's manufacturing facilities (including heating, lighting, ventilation, and air conditioning); equipment (fermenters, valves, and pipes, including cleanliness and maintenance routines); and utilities (water, gas, electricity, and compressed air).

Significant changes in the manufacture of a licensed product require prior approval by the FDA. Minor changes to the process or equipment may require only notification of the FDA. Major equipment changes are considered significant process changes and may require the company to file not only an ELA Amendment but also a product license supplement (which may include preclinical testing in animals and clinical testing to prove equivalence). Process changes are not uncommon. Company Y makes on average one to two minor process changes per year, and one major process change every five years.

---

2. COMPANY Y'S EUROPEAN APPLICATION

Whereas in the United States only the FDA is involved, in Europe the product application process involves both national and community authorities. Five types of regulatory procedures have been used: (1) the original national procedure whereby companies applied separately to each national agency; (2) the multistate procedure (from 1975 to 1995), whereby a company applied in one member state and then sought approval in other

---

member states based on the same application; (3) the concertation procedure (from 1987 to 1995), whereby an EU committee evaluated the application based on prior review by a national agency that assumed the role of “rapporteur”; and (4) since 1995, either (a) the centralized procedure, whereby an EU agency reviews the application and its decision is binding on the member states, or (b) the decentralized or mutual recognition procedure, wherein product approval in one country may be extended to other member states.^[31] The EU concertation and centralized procedures, along with the FDA procedure, are outlined in Table 10.1.

From 1987 to 1995 the concertation procedure was required for all biotechnology derived medicinal products. Since 1995, all such products must be submitted according to the new centralized procedure, while the mutual recognition procedure is available for all other products. In all cases, after a company obtains a product license, it must obtain a separate manufacturing permit from the state or national authority of the country of manufacture.

In 1990, Company Y applied for a product license for its genetically engineered coagulation factor, using the concertation procedure. For the product license, the company asked the relevant regulatory body of one member state (in this case, Germany) to assume the role of “rapporteur,” representing it at the EU's CPMP. After an initial product evaluation, the German regulatory authorities submitted the application to the CPMP, and to all member states for comments. Once comments had been addressed by the company or the rapporteur, it was up to the CPMP to recommend the product for approval, reject the application, or request additional information. After CPMP approval, the company had to obtain a marketing authorization from every country in which it intended to sell its product.^[32] Finally, to manufacture the product in Europe, the company would need a manufacturing permit from a local, state, or national authority. Partly for that reason, it chose to manufacture only in the United States and export the product to Europe.

VI. COMPARING OPPORTUNITY COSTS: THE LENGTH OF THE REGULATORY COURSE

The first injection of a recombinant coagulation factor was given to a patient as early as March 1987.^[33] Most of Company Y's product development work was carried out before 1988. Building on ten years of experience with its nonrecombinant coagulation product and after numerous tests of its pilot processes, the company determined in 1989 that its product was safe, effective, and ready to market. At that time, its projected market was two hundred million units per year at eighty cents a unit. Each month of delay

---

in the processing of its product and facility license application, therefore, could amount to $13.3 million in lost sales, as well as delays in the availability of the product to hemophiliacs in the United States and Europe.

cross-national analyses have indicated that European countries tend to approve new drugs more rapidly than the United States.^[34] Certain studies suggest that the CPMP is faster than the FDA in its review of rDNA-derived product applications.^[35] This is confirmed by Company Y's experience. More regulatory stops did not mean that the regulatory process for the blood coagulation product took longer in the EU than in the United States. The actual review time of the CPMP was seventeen months, as compared with thirty-one months at the FDA. Moreover, most EU member states granted swift marketing authorization following the CPMP approval. Company Y had to wait more than six months for marketing authorization in Italy, however, and still longer in Belgium.

A. The United States

In September 1989, Company Y filed its product license application with the CBER office of the FDA. Over the next several years, the company received four sets of questions from the FDA, the first two sets between December 1990 and February 1991, more than a year after filing its application. During that year, however, company officials were in contact with the reviewers and were kept up to date on the review status of their dossier. The first two sets of questions related very generally to the manufacturing process and the characterization of cell lines.^[36] Between April 1992 and June 1992, the company received a third set of questions addressing specific aspects of product manufacture such as a particular virus removal step of the purification process. The questions came from four different reviewers and were forwarded to the company independently; leading the company to believe that there was no “single person [at FDA] who coordinated the review consistently.”

In November 1992, three years after the original submission, FDA field officers carried out a preapproval inspection of Company Y's manufacturing facilities. Between December 1992 and February 1993, the company received a fourth set of questions, numerous and detailed, to which it responded by fax. The questions focused on details of the Physician Insert and package label, the use of parts per million instead of micrograms per liter for individual parameters, comments on abbreviations, and even punctuation. On February 25, 1993, Company Y was granted approval.

During the application process the company experienced cross-contamination problems that led it to put the approval process on hold for approximately ten months. Subtracting those ten months from the overall review times, the FDA approved the blood coagulation factor in 31 months,

---

just under the average review time (35.5 months) for important new biological products at that time.^[37]

Until 1992, CBER did not have time limits imposed on its reviews of the product license and establishment license applications. In the Prescription Drug User Fee Act of 1992, however, the FDA made a series of commitments to reduce review times in exchange for additional funding, provided largely through user fees paid by industry.^[38] Recent FDA research indicates that, as of 1995, agency approval times were comparable to, if not shorter than, those in other countries. Company experiences described at the 1996 annual meeting of the Drug Information Association confirm such findings.^[39] In November 1995 the President's Office issued a report, Reinventing the Regulation of Drugs Made from Biotechnology, announcing reforms that will “cut drug development time by months.” The most prominent of those reforms was the elimination of the requirement of an establishment license application (in addition to a product license) for specified new biotechnology and biological products.

B. Europe

In May 1990, Company Y filed an application with the German health authorities (Bundesgesundheitsamt, or BGA) based on the one filed with the FDA seven months earlier. The company chose the BGA as its rapporteur because of previous experience with the German agency. In November 1990, however, the BGA returned the application, stating that “the package was unacceptable,” that the dossier was incomplete, and that additional characterization of the process was necessary. Initially discouraged, and still in the laborious process of gaining FDA approval, Company Y decided not to pursue the European application until the fall of 1991.

From November 1991 to October 1992, the company's director for regulatory affairs interacted frequently with the BGA while redesigning the application to fit the European focus. Company Y also generated new data regarding the manufacturing process. Although it was costly and timeconsuming to meet the European requirements, the German reviewer cooperated in helping company officials to understand and meet the “new” requirements. Indeed, Company Y officials came to prefer the European to the FDA process, partly because it included not only the voluminous core documentation but also a high-level summary (called Expert Report), a tabular summary, and cross references.

In October 1992, a year after it really began preparing its EU application, the company officially filed for the second time with the BGA. Six months later, following validation by the German authorities, Company Y submitted the dossier to the appropriate authorities in all EU countries. In September 1993 the BGA submitted its favorable evaluation of the

---

application and forwarded its “assessment report” to all other member states. British regulators announced their intent to inspect Company Y's facilities in the United States.

In October 1993, a year after filing, the company received a first official list of 272 questions from all European authorities, compiled and coordinated by the rapporteur country. This is about the same length of time as the FDA's first response. Company Y responded to all questions within one month. In December 1993 a second set of nine questions was submitted to the company, followed by a third set in February 1994. In March 1994 the company's representatives made a presentation in Brussels to the CPMP Biotechnology Working Party, which remained concerned about three or four issues. On the night before the meeting, BGA representatives met with company officials to discuss and rehearse the presentation. According to Company Y's director for regulatory affairs, rather than creating an extra layer of bureaucracy, the rapporteur provided “a second set of eyes” and guided the company through the complex review process. The Biotechnology Working Party recommended the product to the CPMP, which within thirty days recommended the product for approval to the commission and the member states. The period from the second submission to the rapporteur country and final product approval by the CPMP was clearly shorter than in the United States.

Then, however, the company had to pursue marketing authorization in several member states, which were not automatically bound by CPMP approval under the concertation procedure. The Netherlands granted the authorization within two weeks of the CPMP's recommendation, Germany within one month, the United Kingdom and France within two months, and Luxembourg, Greece, Ireland, and Spain shortly thereafter. Italy, however, did not grant approval until September 1995, almost a year and a half after the CPMP recommendation, and as of March 1996, Company Y was still awaiting authorization in Belgium. Austrian authorities approved sale but stipulated that a sample of every batch of albumin used in the manufacture of the product sold in Austria had to be submitted to Austrian regulators. Because of such member-state variation, some companies apply for marketing authorization only in countries that represent a significant market.^[40] Company Y applied for marketing authorization in all twelve member states.

Under the centralized procedure enacted in 1995, a company is automatically granted marketing authorization in all EU member states following recommendation by the CPMP and approval by the Commission. Differences between member states like those described here must be resolved during the EU approval process. Product pricing and reimbursement, however, will remain within the realm of the national authorities and require companies to continue to negotiate on a country-by-country basis.

---

The length of Company Y's regulatory course in Europe can be calculated several ways. First, if one includes (1) Company Y's first unsuccessful submission (May 1990 to November 1990), but excludes (2) the twelve months in which Company Y put its European application on hold, and then includes (3) the period of redesign of the application (November 1991 to October 1992), (4) the EU review and approval process (November 1992 to April 1994), and (5) approval by most member states (April 1994 to June 1994), the approval time adds up to thirty-seven months, compared with thirty-one months at the FDA. Second, it would not be unreasonable to exclude the twelve-month redesign period, during which the company engaged in some consultation with the German regulatory authorities. Consultations with regulatory authorities prior to submission are common in Europe and the United States, and company officials generally do not include that period in calculating the overall review timeline. In that calculation, the EU approval time adds up to only twenty-five months.

Third, Company Y officials regard both the initial unsuccessful submission and the redesign period as products of their unfamiliarity with European requirements and of European authorities' lack of familiarity with the company's earlier plasma-derived product. The FDA, in contrast, was familiar with Company Y and its earlier product, while Company Y was experienced with the FDA requirements. Thus one might count only the period in which the European regulators and member states actually considered the application—steps 4 and 5. In this calculation, the regulatory course took only twenty months, eleven months less than the FDA's thirtyone. Company Y officials expect the European regulatory regime to be faster with subsequent submissions. On the other hand, the FDA has recently reduced review times to twelve months, as required by the federal User Fee Act for funding the administrative process, so currently it is difficult to determine in which jurisdiction regulatory review imposes larger opportunity costs.

The situation may be different for firms that choose to manufacture biotechnology derived products in Europe. The relative efficiency of the EU product approval system does not take into account costs and delays stemming from nation-level manufacturing permits (not necessary in Company Y's case), which certain cases in Germany and Switzerland suggest can be significant for biotechnology products that elicit local or advocacy group opposition.^[41]

C. Process Change Reviews

Europe and the United States also differ in review times for new manufacturing process changes (with respect to previously approved products or

---

facilities). In applying to the FDA for a minor process change, companies must give the agency a thirty-day notice. In Europe, authorities require an initial thirty days for all member states to acknowledge receipt of the application, followed by ninety days to grant CPMP approval, and another thirty days to gain approval from member states—a total of at least four to five months. In both the United States and the EU, major process changes take much longer and may require new preclinical and clinical studies.

The delays associated with process changes, even “minor” changes, can be very costly. At Company Y's sales level for the product in question during 1997, a thirty-day delay in the United States could cost $8 million in forgone revenue, compared with $32 million from a four-month delay in the EU. If the market grows threefold, as expected, the same process change in the year 2000 could cost the company $96 million in Europe, as opposed to $24 million in the United States. However, under some circumstances, the company can build up a large inventory before initiating the process change, or it can create a parallel new production process, continuing the old one while the agencies review the new one. But that is not always possible, for sometimes the process change is required because of problems in the existing process. In such cases the company may be seriously affected by lengthy review times.

VII. STYLES OF REGULATION

National “styles of regulation” refer to agencies' regulatory approach, their propensity toward legalistic or flexible modes of enforcement, and the character of the interaction between regulators and industry officials. American regulatory agencies have often been characterized as more inclined toward a legalistic or even an adversarial style of regulatory enforcement than their European counterparts.^[42] This section provides some perspective on differences in styles of regulation concerning product licensing for biologics in Europe and the United States.

A. The Regulatory Approach: Technical versus Process Orientation

In processing Company Y's product and facilities license application, the U.S. FDA focused primarily on manufacturing facilities and equipment. As Company Y's regulatory affairs official put it, the FDA said essentially; “Tell us how you run your operations on a daily basis and what operations you have in place to monitor the consistency of the process. Include details regarding utilities, equipment, facilities, and procedures for testing the product that you are manufacturing.”

As noted earlier, Company Y's U.S. application provides eight volumes of information on the manufacturing establishment for the specific product,

---

complementing the standing establishment license of more than twenty volumes. Multicolored floor plans indicate levels of sterility for each area, where and how people circulate, how contamination is controlled, and how buildings and manufacturing areas are protected from cross-contamination. Of the fourteen pages of the FDA's Draft Application for License of Recombinant DNA Derived Biological Products, nine full pages pose manufacturing-related questions, while just one page poses questions regarding the nonclinical and clinical sections, labeling, samples, protocols, records, and environmental impact.^[43] The Chemistry, Manufacturing and Controls (CMC) section of the product license application (one volume) must include information pertaining to the validation and reproducibility of the process design. Through “technical validation,” it describes each manufacturing step, the piece of equipment and its operations, what is achieved by that particular step, and data from several batches demonstrating that the step meets design criteria.

The European regulator, according to Company Y's officials, has a different approach. He says, in effect, “Tell me why you design your process the way you do, which parameters are critical, and how you control those parameters to ensure consistency.” Rather than demanding descriptions of specific technical aspects of the manufacturing facilities and equipment, CPMP officials in the EU demand and review the producer's own in-depth analysis of potential risks and justifications for the control measures adopted. Thus, the EU's CPMP was stricter than the FDA in demanding justifications of process parameters. For example, when Company Y claimed that it could assure the removal of 80 to 90 percent of a contaminant, the European regulators required proof that the company actually had operated at both 80 percent and 90 percent. The FDA had been satisfied by the company's statement of the range of operation. The European regulators, unlike the FDA, also required an explanation for why Company Y operated at the chosen quality control limits and whether it had knowledge of the consequences of operating above or below those limits. Similarly, for the EU application, Company Y had to supply trend charts concerning critical product characteristics that demonstrated consistency within a relatively tight range, rather than the wider variation permitted by the FDA.

While the U.S.-style technical validation considers all elements of the manufacturing process to be equally important, and sets broad limits across the board, the European approach monitors fewer elements more carefully. Thus, to show consistent operation to EU authorities, Company Y had to differentiate between more important parameters (known as “governing parameters”) and less important parameters, and had to establish a statistical relation between the variation in the governing parameters and effective reproduction of the process. Thus, in the long run, company officials

---

said, the EU's emphasis on crucial risk analysis and justification of all steps and specifications may ultimately be beneficial to the company.

The FDA focus on manufacturing facilities and equipment, it should be emphasized, also entails a considerable attention to the manufacturing process. Whereas FDA regulators may be somewhat less fanatical about the consistency of the manufacturing process than their European counterparts, they were interested in whether the rDNA coagulation factor consistently compares to the conventional coagulation factor. Conversely, European regulators have become increasingly interested in manufacturing facilities. Nonetheless, the European focus on the manufacturing process, process validation, and consistency, versus the American focus on the manufacturing facilities, equipment, and technical validation, compelled Company Y to provide two different sets of test data for the same product.

B. Regulatory Stringency, Flexibility, and Unreasonableness

Premarketing approval, like other ex ante regulatory controls, put the agencies in a powerful position vis-á-vis the company. Failure to meet regulatory demands results in delay or potential denial of the product license. Hence, companies commonly comply even if they disagree with the agency's standards and acceptance limits, or view its demands as overly stringent or unreasonable.^[44] Regulatory agencies, however, vary in responsiveness to a regulated company's argument that a particular requirement is unreasonable under the particular circumstances.

European and American regulatory regimes for new biological product applications differ substantially with respect to regulatory “acceptance limits,” that is, in the range of product characteristics within which a manufacturer must operate. Given the inherent variability of biologics and biological assays, relatively broad acceptance limits reduce distribution problems, since unnecessarily tight limits make lot failures much more common. If a lot fails to meet a limit, such as the required salt level, the company's Quality Assurance (QA) Group, in conjunction with clinical studies experts, must then determine whether this failure is critical to the overall efficacy, safety, and quality of the product. Depending on the type and number of failures per lot, the QA review may take as long as two months. As a result, the product will not be delivered on time. A lot failure, in the case of the product in question, can amount to a cost of as much as $10 million. Moreover, if delays occur frequently, the company will lose market share to its competitors.

The European regulators, according to Company Y officials, required tighter acceptance limits for some parameters and enforced them more legalistically. In the United States, the company has been allowed to operate within somewhat broader, self-defined acceptance limits for in-process and

---

final container controls. For example, the FDA expressed concern about a host cell contaminant that remains in the final container at the end of the manufacturing process. In 1989, when Company Y had limited experience with the manufacturing process, it proposed to the FDA CBER a level of 500 picograms of host cell contaminant per unit. The agency considered that limit high, considering the company's actual process experience, and Company Y reduced the limit to 100. After the FDA had completed its review of all preclinical, clinical, and manufacturing data, it concluded that Company Y actually operated well within the limit of 100 picograms and suggested that the limit be tightened. The company proposed a limit of 50, which was accepted, and in practice generally operated at 30.

In Europe, regulators from the start required a minimum level of 10 picograms per unit and refused to engage in any discussion on the subject. Company Y submitted data showing that the product was perfectly safe while operating the process at 30 picograms. The EU regulators, however, continued to press for a limit of 10. When asked about the motivation for such a stringent requirement, the CPMP replied that the 10-picogram limit reflected FDA standards. In fact, the FDA had briefly held the industry to such standards in 1991, but it had since reversed this requirement in light of additional scientific data. Nevertheless, EU officials did not relent. Company Y then carried out sophisticated pilot studies that achieved less than 10 picograms per unit and gained product approval from the EU agency. At times, however, the company has been unable to meet these stringent European requirements in full-scale operation. The resulting lot failures create extra work for the QA personnel, make it difficult to plan manufacture and distribution, and may make it difficult consistently to meet the European demand, risking loss of market share to the company's competitors.

The more stringent European approach sometimes has had useful secondary effects. When Company Y found that from time to time it had difficulties meeting the EU's 10-picogram limit, manufacturing personnel (following the more searching European approach) started wondering about the origin of variations in the contaminant in the final container. After a reevaluation of the purification system, Company Y instigated a process change that led to an improvement in the overall manufacturing system.

On the other hand, Company Y officials have found some European requirements unreasonably stringent. In 1994, Germany unilaterally tightened its requirements for the removal of viral particles in biologics. This forced the company to withdraw a plasma-derived product from the German market, develop an additional virus removal step, and eventually reregister the product with the German authority—although company officials do not believe that the additional process control and cost add

---

significantly to the safety or efficacy of the product. Similarly, the EU's CPMP required that Company Y map the gene of its recombinant product before use in manufacture and after fermentation, using a very difficult technique that remains highly contested in the industry. Company experts argued that the required tests would not provide any relevant information. The CPMP insisted on the performance of the test, arguing that Company Y's competitors had employed that test in the context of an earlier application. Ultimately, Company Y performed what from its experience were tests of no value, at a cost of $50,000.

Not that the FDA is always reasonable. Since 1989, the company had been operating a large-scale fermenter for development purposes. In 1994 it decided to add the fermenter to the product license. To evaluate the safety of the product, the company studied in-process results, tested the product in the final container, and carried out preclinical studies. The data showed that the product from the large-scale fermenter was identical to that from the existing fermenter. In spite of such clear scientific evidence, the FDA demanded that Company Y carry out new clinical studies with the product from the large fermenter. The company questioned the FDA's request on scientific grounds, and the agency allowed it to carry out its clinical studies after approval had been granted, but the studies were costly and, to company officials, wholly unnecessary.

In a more troublesome case discussed by Company Y officials, the FDA decreed in 1992 that manufacturers that pooled donated blood plasma from many donors to make their product should exclude any plasma containing antibodies to hepatitis C. Several experienced companies were reluctant to follow the decision because experience showed that the products on the market did not transmit the virus. Hence they saw no value in adding another step to the purification process, and several companies argued that by removing the antibodies, one increases rather than reduces the risk of contamination. One company (not Company Y) did introduce the FDA's antibody clearance step into its blood-screening process, but in 1995 it received reports that its product had infected hundreds of patients with hepatitis C. It pulled the product from the market. Tests revealed that the antibody-free lots were the culprits.^[45] The FDA argued that the company should have introduced an additional virus removal step. As a result of the FDA's decree, the company faced litigation from infected patients across the world and had to withdraw its product from the market, add an additional virus removal step, carry out animal and clinical studies, and resubmit a new application, with all fees and delays.

Overall, therefore, it seems that both FDA and EU regulators are prone to overregulation. The ethos of the regulatory process provides few incentives for regulators not to err on the side of caution. The keys to avoidance of excessive stringency lie (1) in the development and continued maintenance

---

of expertise in the regulatory agencies, so that regulators develop confidence in their knowledge of risks and in the level of expertise and commitment of particular regulated companies; and (2) in the quality of the interaction between regulators and the regulated. As the next section indicates, Company Y's experience provides some basis for comparing American and European regulators on these dimensions.

C. Interaction with Regulators

Since the agency must make discretionary judgments in its implementation of rules and regulations, the requirements imposed on companies are likely to vary according to the “dialogue” between regulators and manufacturers. In the ideal case, for example, if clinical or other studies do not show the expected results, a company that has a good rapport with an expert regulator may benefit from the agency's scientific and technical advice, and avoid triggering a high-profile and costly “clinical hold.”

Earlier studies have shown that company-agency interaction tends to be less formal and legalistic in Europe than in the United States. In the current case study, this is only partially true. Once engaged in the review of its EU product license application, Company Y experienced fairly informal interaction with the rapporteur country reviewer. Because of the time difference between Europe and the United States, the reviewer allowed calls in the evening at his home number. At times the same informality also existed between the company and the FDA regulators. However, while Company Y's informal rapport with the European rapporteur grew over the years, company relations with the FDA changed with the arrival of a new group of reviewers. Because of their inexperience, the new reviewers had a more formal, legalistic, and at times stringent regulatory style. Since reviewer turnover is common at the FDA, the risk of legalistic interaction seems to be higher in the United States.

Overall, in Company Y's experience, the EU's system of review operates more through expert dialogue, in contrast with the FDA's more mechanical regulation-applying approach. The European agencies require that the company submit “Expert Reports” summarizing the pharmaceutical, pharmacotoxicological and clinical parts of their application and to designate an expert with formal qualifications and practical experience in the subject matter. Only when the national reviewers discern inconsistencies or analytic weaknesses in the Expert Report do they pursue a detailed statistical analysis of test results and manufacturing procedures. The U.S. CBER, in contrast, typically carries out a broad range of batch release tests at its in-house laboratories and systematically reviews company data at its statistical division.

Yet in some respects, the European regime is more formal. In contrast

---

to the FDA, European authorities require that all applications be submitted in full; they do not allow parts of the data to be submitted after the initiation of the review process, largely because they must consult regulators from many national agencies. Pursuant to the 1997 FDA Modernization Act, a more gradual review process allows for a step-by-step submission for certain products, permitting a company to initiate application procedures before all the data has been generated.

Furthermore, according to Company Y, at least some European inspections are more legalistic than FDA inspections. Under an agreement of mutual recognition of good manufacturing practices, the European authorities rely almost entirely on the FDA to assure the reliability of the facilities for products manufactured in America and exported to Europe. Whereas the FDA inspects foreign facilities producing for the American market with some frequency, European inspectors rarely visit manufacturing facilities in the United States.^[46] In the case of Company Y, however, its U.S. facilities were inspected by British authorities, notwithstanding the mutual recognition agreement. The British officials conducted “a very harsh inspection” according to Company Y's regulatory affairs officer, demanding to see everything and requiring immediate corrective action when they discovered anything they regarded as problematic. FDA inspections, though frequent, are often very cursory. In the words of one industry official:

I think they [the FDA] operate a lot on appearances. I have not seen an inspection yet where they really find something and ask a question about a document and really take it to a conclusion. I don't know how popular these inspections are within the agency, if people consider these to be drudgery or not. It is almost as if they want to get in and get out as fast as they can.

Industry officials suggest several reasons for the contrast in inspection styles. First, British inspectors are on average much better trained than the FDA inspectors, and inspections abroad are carried out by the agency's more senior members. To be a facility inspector in Britain is considered a rather prestigious government career, whereas in the FDA, reviewers as well as inspectors often view their time in government as a stepping-stone to a “more promising” career in the private sector.^[47] Hence, FDA inspectors and reviewers are often new on the job and inexperienced with the companies. This can lead to either excessive legalism or excessive leniency. Company Y has experienced both.

Regulatory behavior, in Company Y's experience, varies according to the officials' specific knowledge. When the company submitted its U.S. product license application in 1989, the FDA reviewers were familiar with the company's history, earlier nonrecombinant product, and regulatory affairs personnel.

---

The reviewers, in the opinion of Company Y officials, were not excessively stringent. As discussed earlier, Company Y has experienced what it considers regulatory unreasonableness at the FDA, but these incidents were often tied to a period of frequent reviewer turnover.

The German reviewers, in contrast, had little history with either the company or its earlier product. They were very concerned about proper procedures and documentation. They held Company Y to stringent standards of data analysis and required clear explanations for every manufacturing step. The company considered some of these requirements excessively stringent and unreasonable. Five years later, however, the same German and CPMP regulators are still in place, and Company Y's official says,

I think that our relationships with Europe are now better than with the United States. Exactly because you have new reviewers [at the FDA] who are not familiar with the product. They are conservative. In Germany, they are knowledgeable about the product. They have been involved with us for over five years, and so, when we make a change, they have all the documentation, they know where to look, and, as a result of their involvement with this product, they are credible to the member states.

VIII. CONCLUSION

This study of the regulatory process for a new biological product does not provide consistent support for the conventional wisdom that American regulatory processes are more legalistic, adversarial, and costly than those of European nations. The European multistep approval process, as compared with the U.S. one-step process, was more complex, difficult, and costly. Although Company Y, once it figured out the EU system, gained product approval in Europe in eighteen to twenty months, compared with thirtyone months at the FDA, subsequent manufacturing process changes are faster and therefore less costly in the United States than in Europe. Moreover, the U.S. FDA now consistently reviews license applications within twelve months as the agency strives to meet its performance goals established under the User Fee Act. Europe is now attempting to align its review times for applications as well as process changes for biotechnology-derived products with those of the FDA. Therefore, differences in review times and time to market no longer constitute an obstacle for companies pursuing simultaneous submissions to European and U.S. authorities.

In their interactions with Company Y, European regulatory officials were in some respects more stringent and more legalistic than their U.S. counterparts, although both systems were considered legalistic and unreasonable on some issues. Legalistic inflexibility and its costs appear to be associated with lack of familiarity with the company and its product on the part

---

of the regulatory reviewer. In that regard, the risk of regulatory inflexibility seems in general to be higher in the United States because of greater turnover of regulatory personnel.

The U.S. FDA's regulatory approach entailed a more mechanical focus on manufacturing facilities, as contrasted with a more searching and useful European-style process control analysis. Yet the FDA recently has sought to emulate the EU method, while European regulators now put more emphasis on the manufacturing facilities, including inspections of U.S. facilities for products marketed in the EU.

Based on this study, both approval systems are costly and burdensome. Neither system is clearly more cost-effective, as savings in one area are undermined by costs in another. In this case, the cost of compliance probably was slightly higher in Europe than in the United States because more tests were required and authorities insisted on tighter tolerances. Yet, in light of the cross-national harmonization of product specifications and more global product planning in the industry, such differences may be eliminated in the future. While differences in regulatory requirements, procedures, and styles persist, and while European regulatory agency officials insist that U.S. approval is not automatically a ticket to the European market, regulatory convergence and harmonization are clearly under way.

NOTES

---

---

---

11. New Chemical Notification Laws in Japan, the United States, and the European Union

Lori A. Johnson, Tatsuya Fujie, andMarius Aalders

I. INTRODUCTION

Recently, local fishermen removed the net that for over twenty-three years had enclosed Minamata Bay in southwestern Japan. The net was used to prevent the migration of fish in and out of the bay because methylmercury chloride, discharged from chemical plants in the 1950s, accumulated in the marine life and poisoned 2,261 people in this region. During the 1960s and 1970s, many widely used chemical compounds—from DDT to kepone to chlorofluorocarbons (CFCs)—were judged, belatedly, to be harmful to human health, to ecosystems, and even to the global environment. Meanwhile, chemical companies continued to invent and market new, socially useful chemical substances, and to do so at a high rate. In all economically developed democracies, therefore, government policy makers were called upon to create laws and regulatory systems that would detect and prevent the potential health and environmental risks that might be associated with new chemicals—but without blocking the continued production of socially useful chemical innovations. The challenge was to devise a regulatory “net,” not unlike the one in Minamata Bay, that would evaluate chemical products, restrict the harmful ones, and allow the beneficial and safe ones through. Of particular concern was the lack of advance information about the environmental or health effects of the large number of new chemicals introduced each year.^[1]

Accordingly, many economically advanced countries adopted regulatory statutes requiring chemical companies to conduct tests for hazardous characteristics before manufacturing or marketing new chemical substances. Government regulators were instructed to assess the adequacy of the testing and to prescribe other precautions (e.g., concerning packaging and

---

warnings) before permitting manufacture or marketing—much as premarket testing and governmental clearance had been required for new pharmaceutical products. In 1974 Japan enacted Kashin-hou, which established legal controls for industrial chemicals, including premarket notice and classification of newly manufactured and imported substances, as well as some existing substances. The United States enacted the Toxic Substances Control Act (TSCA) in 1976. In 1979 the European Economic Community, following the lead of several European countries such as France and Switzerland, adopted the Sixth Amendment to the existing directive on the classification, packaging, and labeling of dangerous substances.

Notwithstanding these convergent trends in law, one would expect significant differences to remain in the ways these laws are designed, implemented, and enforced. Many studies indicate that U.S. regulatory and liability law tends to be designed and implemented in a more adversarial, legalistic, and punitive way than comparable legal regimes in Japan and Europe.^[2] A study conducted by Ronald Brickman, Sheila Jasanoff, and Thomas Ilgen compares the development of several chemical regulation schemes in Europe and the United States.^[3] Although there is no attempt to systematically compare the impact or effectiveness of the policies, the authors do note that “it appears that U.S. legislative and regulatory policies have created a potential for higher costs than are likely to be imposed on industry in Europe … [and] with respect to the administrative costs of complying with regulations … there is little question that U.S. industry has to spend considerably greater sums than its European counterparts.”^[4]

This case study attempts to test these generalizations with respect to the new chemical notification and clearance laws of Japan, the United States, and the European Union (EU). We draw on the experience of multinational chemical companies to examine the operational differences and relative burdens among premanufacture and premarket clearance laws, as actually implemented. Our findings partially contradict the conventional wisdom about comparative regulatory style. According to chemical company officials with substantial cross-national regulatory experience, initial compliance with new chemical notification and clearance laws is less costly, more expeditious, and less burdensome in the United States than in Japan and the EU. On the other hand, regulatory officials in the United States, as might be expected, are uniquely legalistic in their response to noncompliance; U.S. officials are much quicker to impose monetary fines, even for violations that are not directly harmful, and they impose much larger fines. Overall, however, the American regulatory system is regarded as more conducive to chemical product innovation than the Japanese and EU systems. Yet we have uncovered no evidence, moreover, that this has led to less protection of the public in the United States.

---

II. RESEARCH METHOD

The research effort began by arranging interviews with regulatory compliance officials in several multinational chemical companies, selected because they conduct parallel business operations in the United States and either Japan, the European Community, or both. The methodology is best described as a “snowball” sample, beginning with a few informed chemical company officials with cross-national experience. As we talked with these officials, they suggested other persons with broad comparative experience, both within their corporation and in other multinational corporations, who might be helpful in our research. We were particularly interested in talking with officials who have participated in repeat interactions with the different regulatory regimes and discussing the impact of those differences on their corporate practices. Although the interviews are not representative in the statistical sense, we have spoken with some of the most knowledgeable informants within the chemical industry, many of whom know each other through international meetings of industry associations.

Unlike other case studies in this book, this chapter does not focus on the experience of a single company and its compliance with notification requirements for a single product in Japan, the United States, and Europe. This is due, in part, to the practical reality of how our research developed. We were able through referrals and introductions to speak with many chemical company officials in different major multinational corporations. We determined that, for these regulatory programs, entailing large numbers of new chemical substances each year, evaluating compliance efforts in different companies in different countries would be more appropriate. This greater breadth justified any sacrifice in the depth of detail with respect to specific products, companies, and experiences. Although we report several companies' experiences in notifying regulators about particular chemical products, we also report experienced company officials' broader generalizations about cost, delay, enforcement officials' attitudes, and the effectiveness of regulation in the different regimes. We have considerable confidence about the validity of these generalizations for several reasons. First, the company officials we spoke with had extensive cross-national regulatory experience. Although personally based in either Japan, the United States, or Europe, they had either supervised compliance in all regimes or were responsible for informing company officials of the requirements of more than one regime. Second, the company officials had no obvious incentive to disparage any one country's regulations apart from the impact of the regulation on their company, which was specifically what we were interested in exploring. The officials interviewed, regardless of their own country of origin, generally agreed in their characterizations of the

---

American, Japanese, and European regulatory regimes for new chemicals. In contrast to what some readers might presume to be a corporate incentive to be especially critical of American regulation on grounds of its greater legalism and costliness, the officials we interviewed, while often criticizing some aspects of the American regime, were unanimous in judging it to be generally more flexible, more efficient, and less costly to comply with than the Japanese and the EU regimes.

We conducted in-depth interviews with company officials, government representatives, academics involved in either government or industry, and representatives of industry associations and organizations. In Japan we interviewed company officials from three multinational chemical manufacturing companies (J-1, J-2, and J-3) based in Japan. We also interviewed several research scholars in private institutes dedicated to studying chemical safety; these individuals had retired from major Japanese chemical manufacturing companies but maintain strong influence in the industry and ties to government. We also talked with officials of the Ministry of International Trade and Industry (MITI), the Ministry of Health and Welfare (MHW), the Environment Agency, and a U.S.-based consultant to the Japanese chemical industry. In the United States we have interviewed nine officials in four major multinational chemical manufacturers in the United States (US-1, US-2, US-3, US-4), an Environmental Protection Agency (EPA) official responsible for enforcement of TSCA, and a representative of the chemical industry association. In Europe we conducted interviews with five company officials responsible for product safety and regulatory compliance in five different multinational chemical manufacturers (two in Germany [EU-1, EU-2], two in the Netherlands [EU-3, EU-4], and one in England [EU-5]). Together, these experienced individuals constitute a cluster of “elders” in the international community of chemical risk managers. They are knowledgeable informants concerning the law and practice of chemical risk regulation in the United States, Japan, and Europe.

We conducted some interviews by phone, others in person, and others through written reports prepared by the company. Some interviews used a common questionnaire containing open-ended questions intended to elicit the informants' own descriptions, explanations, and comparisons of the three legal regimes. Other interviews were more casual, with the researcher and the informant meeting or talking on the phone frequently; these “freeranging” discussions often led to valuable insight and information. The pool of interviewees, including the company and government officials, came from a variety of backgrounds and included chemists, lawyers, professors, private researchers and scholars, government officials, plant managers, laboratory researchers, business consultants, and insurance company officials.

In these interviews we sought the insight of company officials on two

---

primary questions. First, in which system is compliance with premarket regulations most costly for their business and why? Second, from the standpoint of operating chemical companies, what are the most significant differences between the chemical control laws of the United States, Japan, and the Western European nations with respect to efficiency and effectiveness?

III. CHEMICAL CONTROL NOTIFICATION LAWS: GENERAL OVERVIEW

The chemical control notification laws of Japan, the United States, and the European Community can be compared with regard to three aspects of the chemical control notification systems: (1) the stringency of the premarket testing and clearance requirements, and the related costs and delays; (2) the cooperativeness of agency decision making regarding compliance and required precautionary measures prior to marketing; and (3) the strictness of regulatory enforcement through inspections and penalties. In each of these categories, we will be considering not only cross-national variation in the substantive requirements of the laws but also the structure and style of enforcement of the law as experienced by regulated enterprises.^[5]

One characteristic dilemma associated with government regulation is the effort to strike the appropriate balance between costs and benefits. National regulatory programs for premarket clearance of chemicals arose in the shadow of existing programs for new pharmaceutical products in the United States and Europe. Such programs, experience showed, were subject to criticism for two kind of errors. Type I errors are errors of leniency, allowing the marketing of products (such as thalidomide in several European countries and Japan) that turn out to have serious harmful side effects. Type II errors entail excessive stringency. They arise, of course, from the effort to avoid excessive leniency. Thus during the 1970s, the U.S. Food and Drug Administration (FDA), many studies indicated, imposed such stringent and costly testing obligations on pharmaceutical manufactures, and took so long to approve applications, that American citizens were unable to obtain valuable new drugs that their counterparts in Europe benefited from months or even years earlier. Moreover, pharmaceutical companies were reluctant to undergo the cost and delays of developing so-called orphan drugs—therapies for conditions and diseases that are relatively rare, and hence would not generate revenues large enough to cover the costs and delays of the regulatory process.^[6] Although industrial chemicals differ from pharmaceuticals in terms of marketing, exposure, and hazard scenarios, premarket clearance regulations on industrial chemicals engender the same dilemma—finding the right balance between excessive leniency (Type I errors) and excessive stringency (Type II).

In designing premarket clearance programs for new chemicals, the

---

European system leans toward the avoidance of Type I errors by imposing mandatory testing requirements on chemical companies for all new chemicals and even for increased production of previously tested chemicals. The U.S. TSCA, in contrast, was crafted in the mid-1970s, at a time in which the FDA “drug lag” had become a serious concern. TSCA seeks to minimize Type II errors by emphasizing company notification of regulators regarding the identity of chemicals rather than universal premarket testing. TSCA puts the burden on the EPA to decide which chemicals must be tested by the applicant company and requires the EPA to make that “notification-only or test” decision quickly, within ninety days. Japan's Kashinhou leans toward avoidance of Type I errors, but errors of a particular kind: the dissemination of nonbiodegradable chemicals that pose a risk of contamination of food sources.^[7] To that end, it mandates a step-by-step system of classification under which the results of the initial testing trigger the requirement or waiver of further testing; it falls between the U.S. and European regimes, since full-scale testing is not required for all new chemicals.

A. Japan: Kashin-hou

The Japanese chemical control law, as originally formulated, reflects that country's painful experiences with the contamination of marine food resources by pollution in the 1950s and 1960s. Although the law as revised in 1986 addresses broader concerns, the focus remains on risk of harm to humans from chemical substances that enter the marine food chain.^[8] Both for chemicals manufactured in Japan and for imported chemicals that are not on an inventory of approved chemicals maintained by MITI and MHW, Kashin-hou requires that the chemical substance be classified according to its level of degradation in water, its accumulation in fish, and its chronic toxicity.^[9]

In principle, the notification and testing proceeds on a step-by-step basis:

First, the company must test the degree of biodegradation of the chemical using a test called the Ready Biodegradability Test.

If biodegradability is high, the substance is considered safe, and the company is free to manufacture.

If the degree of biodegradation is low, the company must proceed to the next level of testing.

The next level of testing is the bioaccumulation level determination, which measures bioconcentration by means of the Flow-Through Fish Test.

If the level of bioaccumulation is high, the company must conduct full chronic toxicity tests, which include carcinogenicity testing, reproduction

---

― 347 ―

tests, teratogenicity tests, as well as toxicokinetics and pharmacological and mutagenicity testing.

If the results of the full chronic toxicity testing are negative, the company can freely produce the chemical.

If the results are positive, the substance will be considered a First-Class Specified Substance, and companies are virtually banned from manufacturing it.

If the level of bioaccumulation is low, the company must conduct Screening Toxicity Tests, which include testing of mutagenicity and the Repeated Dose 28-Day Test.

If the results of the Screening Toxicity Tests are negative, the company can manufacture the chemical without restriction. If the results are positive, the chemical will be classified as a Designated Substance. A Designated Substance will be reclassified as a Second-Class Substance if annual environmental monitoring indicates an increase in the substance in the environment. In such cases, further restrictions on production will be imposed.

Although the law requires the preceding tests on a step-by-step basis, according to our interviews, companies often conduct the Biodegradation, Bioaccumulation, and Screening Toxicity Tests together for low-molecular-weight substances, in order to save time and because generally only 15 to 20 percent of new chemicals have high biodegradability.

After testing is completed, there are three stages to the new chemical notification process in Japan. The first step is an informal interview session, referred to as a “hearing,” three weeks prior to the date of the formal notification before the Chemical Product Committee, composed of chemists appointed by the two agencies.^[10] This hearing is provided independently by MITI and MHW. The next step is the preliminary evaluation, which is another informal session with MITI and MHW a week after the initial hearing.^[11] After responding to MITI and MHW comments, the company submits its data to the Chemical Product Committee.^[12] MITI is responsible for the biodegradation and accumulation concerns, and MHW for the chronic toxicity issues. After the Chemical Product Committee reviews the submitted data, the company receives a notice classifying the notified substance as safe, designated, or first class specified, which entails restrictions or a prohibition on manufacture or import.^[13] For violation of the regulatory requirements, Kashin-hou authorizes imposition of penal servitude of no more than three years, a maximum fine of $10,000, or both, for persons or companies manufacturing or importing First Class Specified Substances.

---

B. United States: Toxic Substances Control Act

Of the many different aspects of TSCA, the category we are most concerned with is the new chemical review process for manufactured and imported chemicals. The TSCA procedure initially emphasizes providing the information government regulators might need to identify and control possible risks, rather than comprehensive testing.^[14] Section 5 of TSCA requires companies to file a premanufacture notice (PMN) with the Office of Pollution Prevention and Toxics of the EPA for any new chemical that is not on the TSCA inventory of existing chemicals.^[15] The company must file the PMN ninety days prior to manufacture or import of a new chemical substance, or before the manufacture or processing of an existing chemical for a “significant new use.”^[16] There are exemptions to the full PMN requirement for test-marketing of chemicals, low-volume release, and polymers, which are not considered chemically active or bioavailable.^[17] The PMN submission must include available data on chemical identity, expected production volume, by-products, intended use, expected environmental release, disposal practices, and human exposure.^[18]

Although the submitter, in contrast to the submissions in Europe and Japan, is not required to conduct new testing, the company must include all existing health and environmental test data in its possession and a description of existing data it knows about. For example, if the company has conducted testing in connection with past or present marketing in Japan or Europe, it must submit that data to the EPA. According to the EPA, less than half of the PMNs submitted include toxicological data.^[19] Using the information on chemical identity submitted by the company, computer analysis, and Structure Activity Relationship (SAR) Data, the EPA compares the chemical to structurally similar chemicals to assess the potential risk. The EPA is concerned primarily with whether the chemical is similar to asbestos, radon, lead, or polychlorinated biphenyl (PCB). If the EPA determines that the chemical may pose an unreasonable risk to health or the environment, it can request additional testing from the company.

Significantly, TSCA states that the EPA must complete its review process on the PMN within ninety days.^[20] For almost 90 percent of the PMNs submitted, the EPA determines that the chemical does not pose an unreasonable risk and places it on the TSCA inventory without restriction.^[21] After filing a “notice of commencement,” the company is free to manufacture the substance without further restriction. Once any company has filed a PMN regarding a chemical substance, it is placed on the TSCA inventory, and no further notification by any company is necessary for that chemical unless the company plans a “significant new use” of the chemical.^[22]

For the other 10 percent of PMNs—those for which the EPA determines that the chemical substance may pose an unreasonable risk to health or

---

the environment—the EPA may request additional toxicological testing, permit manufacture only under conditions specified in a consent order, or prohibit the manufacture of the chemical. If the EPA determines that a chemical presents an unreasonable risk, it has broad discretion under TSCA to require companies to comply with a wide range of possible precautionary measures. These precautions generally are codified into a Significant New Use Rule (SNUR) for that chemical, and other companies that manufacture the chemical will have to adopt the precautions. Unlike the Japanese and European regimes, the EPA is explicitly required by TSCA to consider the economic consequences and to conduct cost-benefit analysis of required precautions.

A company violates TSCA if it manufactures or imports a chemical that is not on the TSCA inventory without filing a PMN, or if it does not comply with requirements of consent orders or SNURs. Fines for TSCA violations, at up to $25,000 per day, are much larger than potential fines in the Japanese or European systems.

C. Europe: Sixth and Seventh Amendments

The EU premarketing directive was adopted in 1979 and took effect in 1981 as Directive 67/548/EEC, the Sixth Amendment to the existing directive on the classification, packaging, and labeling of dangerous substances.^[23] Member states adopted implementing legislation and are charged with implementation of the amendments at the national level.^[24] In 1992 the EU adopted Directive 92/32/EEC, the Seventh Amendment, further harmonizing the procedure for the notification of new substances, adding a requirement for risk assessment, and modifying some of the requirements of the Sixth Amendment.

The EEC directives contain both notification and quantity-triggered mandatory testing requirements prior to marketing or importing of chemicals that are not on the European Inventory of Existing Chemical Substances (EINECS). The EEC directives require only premarketing notification, rather than TSCA's premanufacture notification requirement, so that notification is not required for intermediates (substances produced for use in manufacturing other chemicals but not marketed to third parties).^[25] A company seeking to market a new chemical substance must notify the competent authority in the member state where the substance is produced sixty days prior to marketing.

In contrast to the TSCA inventory, which expands to include each new chemical notified, the EU inventory (EINECS) has been static as of 1981. No new chemicals are added to the inventory; each company marketing or importing a chemical not on the inventory must notify again, even if the chemical has been notified previously by another company. In order to

---

address duplicative test-marketing that occurs because of a static inventory, the EU allows subsequent notifiers to use test data from earlier notifications, with the written permission of the previous notifier, and publishes a list of substances previously notified called European List of Notified Chemical Substances (ELINCS).

For chemical substances that are marketed in quantities of one ton or more a year (Level 0) per producer,^[26] the company must submit to a national administering agency technical information including the identity of the substance, the Chemical Abstract Service (CAS) number, if available, the chemical composition, impurities, the scope of use, production levels for that year, projected production for the future, recommended handling methods and precautions, and proposals for the classification and labeling of the substance. The key distinction from TSCA is that the company must provide, in addition to information on the physicochemical properties of the substance, a base set of toxicological studies, including tests for acute toxicity, mutagenicity, and carcinogenicity, and ecotoxicological studies for acute toxicity for daphnia and fish, and biodegradability.^[27] The company submits three separate dossiers containing physicochemical data such as melting point, flash point, flammability, toxicological data such as acute toxicity and carcinogenicity test results, and ecotoxicological data such as biodegradation and fish and algae daphnia tests. There are exemptions in the EU for test-marketing and research and development, which unlike in the United States do not require an application.^[28]

Under the EU regime, as the quantity of production of the chemical increases, additional toxicological tests are added, and subsequent notifications must be filed. The focus of the administering agency is on the appropriate classification and labeling of the chemical based on its potential toxicity. The enforcing agency of the member state in which the company filed the notification must decide whether a company's testing, notification, packaging, and labeling of a new chemical are in accord with the directives. If the company meets the paperwork requirements of the regulations and follows any labeling requirements, it is free to market the chemical in any member state. According to our interviewees, the agencies enforcing the Sixth Amendment, unlike the EPA, rarely engage in risk assessment and management of risks through controls on the chemical or use restrictions.

The Seventh Amendment specifies that the administering agency of the member state has the ultimate responsibility for conducting risk assessment of new chemical substances according to principles specified by the EC Commission, but a notifier may choose to conduct and submit its own risk assessment with its notification. Whereas the directive introduces the potential for more comprehensive risk assessment, according to our interviewees it remains unclear what impact this will have on the notification

---

process. Prior to the Seventh Amendment's passage, low-volume substances (less than one ton of production per year/per producer) were regulated only at the national level. The Seventh Amendment harmonizes the regulation of those substances and subjects them to a reduced notification procedure. The Seventh Amendment also, for the first time, formally defined a polymer and considerably narrowed the polymer exemption.

Another change brought by the Seventh Amendment responded to certain importers' practice of marketing a low-volume substance through as many as ten or twenty different European corporations to avoid increased testing requirements for larger quantities. Now, a non–European Community producer must nominate one entity, a sole representative, which is responsible for notification of the aggregate import.

IV. COSTS AND DELAYS IN THE THREE NEW CHEMICAL CLEAR ANCE SYSTEMS

All company officials interviewed think compliance with the notification requirements generally is cheapest in the United States and most expensive in the EU. Japan, while occasionally the most expensive, is always the most time-consuming, partly because of the staged character of testing, partly because of delays at the agency decision-making level. The primary factors causing these differences are the level of mandatory testing required and the availability of exemptions.

The experience related by a company official in US-2 vividly illustrates the cost impact of the EU and Japanese mandatory testing requirements. US-2 sought to market a monomer that was not on the TSCA inventory, EINECS, or the MITI inventory. In each instance, the company elected not to claim the identity of the specific chemicals as confidential. The official conducted an estimated breakdown of the costs of collecting the necessary scientific data and filing PMNs under TSCA, Kashin-hou, and the EC directive. For Europe, his estimate was $200,000 for the required base set of testing required. In Japan the estimated cost for biodegradation, bioconcentration, and a series of toxicology tests was $140,000. In the United States the company conducted a standard battery of toxicity tests and was able to submit environmental data on chemically similar substances (SAR Data), at a total cost of $20,500 for testing and $2,500 for the filing fee. Thus, the total cost for compliance with the notification process was $200,000 for Europe, $140,000 for Japan, and $23,000 for the United States.^[29] As this example illustrates, chemical firms' ability to rely on the EPA's use of structural analogies in evaluating the risk of the chemicals, rather than mandatory testing, provides a significant cost savings compared with other systems.

Further savings arise in the United States because (1) TSCA allows the

---

filing of consolidated PMNs for similar chemicals, unlike the Japanese and European systems, which require a separate notification for each chemical substance; and (2) unlike Europe's static inventory, which requires each company marketing a chemical to notify it, TSCA adds previously notified chemicals to the inventory, so that subsequent companies manufacturing the chemical do not have to file PMNs. Consolidated PMNs and “once and for all” notifications contribute to even more stark cost differentials. For example, US-2 wanted to market a family of surfactants consisting of fortyeight chemicals. In Europe, US-2 estimated that the cost of testing was $200,000 each for almost all of the new chemicals (some were on the EINECS inventory), totaling almost $8 million. In Japan the cost, assuming that the chemicals pass the biodegradability tests and no further testing would be required, was $17,000 per chemical,^[30] for a total estimate of $680,000. In the United States the company spent $96,000 in testing for acute environmental harm studies and filed three consolidated PMNs for the forty-eight chemicals, for a total cost of $102,000. Thus the total estimated cost for US-2's compliance with the notification regulations for this family of chemicals was almost $8 million in Europe, $680,000 in Japan, and $102,000 in the United States.

Whereas US-2 reported significant potential cost variation across jurisdictions depending on the testing required, other companies that regularly sell large volumes of chemicals worldwide were less concerned with this type of variation. For example, US-1, a manufacturer of large-volume industrial chemicals, estimated that it spent approximately $500,000 in testing prior to the manufacture or marketing of a typical new product. The standard procedure for US-1 is to conduct the required base set of testing for EU and the bioaccumulation tests for Japan, and then submit the notifications in each region. Because the testing conducted for the other systems is submitted with the TSCA PMN, US-1, in effect, internally harmonizes the regulations for its compliance efforts.

Similarly, EU-5 reported that when it did a worldwide notification, it tended to do the EC directives base set and then add the Japanese protocols. For example the Repeated Dose 28-Day Test of acute toxicity is required for both, but Japan requires additional information, so EU-5 builds on the EU protocol to make it acceptable for Japan. EU-5 sets up one schedule of testing for all regions and designs a whole package of information on the new chemical for use in all of the PMNs. The regulatory manager at EU-5 reported that he tries to convince the businesspeople that “notification is like a building block,” and the company should do everything it can to help them in future marketing efforts. If the molecule is not listed in Japan or Korea, EU-5 tries to conduct tests that will meet these countries' requirements or builds tests into the protocol that will address typical EPA concerns.

---

Although filing PMNs under the U.S. system is almost always less expensive, a company official in charge of product safety and regulatory affairs for EU-1 observed that one can never be completely sure what the cost will be in the United States because of the possibility that the EPA will decide to demand further testing. The flexibility of the system, thus, is accompanied by a degree of uncertainty concerning testing costs. The European system, which requires testing for all chemicals, depending on volume, is very predictable but very costly. The Japanese system, most officials assert, is both unpredictable and costly.

One company official from EU-1 observed that the U.S. regulatory system was preferable in terms of promoting innovation and introduction of new chemicals, largely because of its unlimited exemptions for research and development and for test-marketing of new chemicals.^[31] In the United States, a chemical producer can cooperate with a customer on research and development of a new product without having to file a notification; using these exemptions does require a lot of paperwork and record keeping, but the exemptions, if the EPA grants them, have no time or volume restrictions. In Europe, according to the official from EU-1, if a chemical company cooperates with a customer, it is deemed to have placed the product on the market. The only exemptions for research and development in Europe are for process-based research, which comes with a time restriction, and for scientific research, which comes with a volume restriction.

In addition to the direct costs of the new chemical notification systems, company officials reported various indirect costs such as delay, loss of market opportunity, and diminution of innovation. According to a company official with EU-1, the EU system leads to diminished innovation of new substances, a concentration of innovation in bigger companies, and increased cost and time to market. The official supplied us with figures demonstrating that the number of high volume substances notified by EU-1 has been decreasing over the past five years, especially after the promulgation of the Seventh Amendment, implying a slowdown in innovation. He also reported that his company often notifies first in the United States because it can test the market easier there to know whether its product will be accepted. The company still begins with the EU-required base set of tests for marketing one ton of a new chemical. Then it files in the United States (where the EPA says “just submit whatever you have from doing your base set in the EU”) and moves some of the substance on the market to determine whether there is commercial potential. Based on these efforts, the company will decide whether to go through with the more stringent EU testing required for larger tonnages. In the same vein, another company official described the EU notification system as a “gamble”—if you want to use a new chemical for the market and the notification flops, as it has for his company on occasion in Germany, then you lose a fortune.

---

If you fulfill the notification requirements and the market does not want the product, you have also lost.

An official of US-4, a manufacturer of specialty chemicals, also focused on the costly delays associated with mandatory testing requirements in Japan and the EU. Describing US-4's regulatory experience with registering a new chlorofluorocarbon molecule, he said it reflects the company's general rule of thumb that “getting a chemical to market” will take three to six months in the United States, a year in Europe, and two years in Japan. For this particular chemical, US-4 spent ten months conducting the testing for the EU requirements and forty-five days of regulatory review.^[32] After the three months of regulatory review in the United States, the EPA required additional testing, but its consent order did not require the testing to be done prior to marketing. The company had not received approval from Japan at the time of the interview, but the official expected that it would take eighteen months, less time than usual, because the company had taken a gamble and started the long-term testing early, rather than waiting for the results of the short-term testing.

Japan-based chemical manufacturing companies reported similar concerns with the cost of testing and the time involved in complying with Kashin-hou. Conducting the biodegradation testing alone costs approximately $17,000; on the other hand, the biodegradation, accumulation, and toxicity screening exams together can cost $200,000, and if full chronic toxicity testing is required, the cost can escalate to $4 million.^[33] An official from J-2 discussed an aspect of Kashin-hou it found particularly problematic. During the biodegradation tests, if the new substance degrades into multiple chemicals that have not been tested before, the company must conduct further tests on all these chemicals. Moreover, unlike TSCA and the EC directives, Kashin-hou does not exempt new polymers from notification and requires polymer flow-scheme testing, which costs from $25,000 to $30,000.

A corporate official from J-2 estimated that it takes his company a minimum of one and a half years for testing and collecting data, and at least five more months, including one month for the initial MITI/MHW hearing and three months for evaluation at the Chemical Product Committee and its notification procedure, before it can market the chemical product. Moreover, if the demanding standards for data accuracy applied by the enforcing agencies are not met, then the testing must be repeated until the company “gets it right.” Companies also considered compliance with Japanese regulations more burdensome because there are no explicit guidelines on how the results of the chemical testing will be evaluated, which complicates the company's marketing decision. For example, an official from EU-2 reported that although his firm very often files notifications

---

in Japan, they are accepted only some of the time, and he has no clear understanding of why this is so.

V. AGENCY DECISION MAKING

After a chemical company submits its notification to the relevant agency, its ability to proceed with marketing depends on the competence and caseload of the agency officials, the speed with which they assess the information submitted, and their posture in negotiating conditions and precautions concerning production, shipping, packaging, and marketing. To the company officials we interviewed, the regulatory style of the agencies administering the notification system—their cooperativeness, flexibility, technical expertise, and speed of decision making—had as much significance as the stringency of the legal requirements. All the company officials we talked with had some experience with regulatory agencies in the United States, Japan, and Europe, but they had the most experience with the agency in the location where they operated.

Generally speaking, the company officials saw all the agency regulators as cooperative at the “prefiling” stage, while the company is gathering necessary information and preparing to submit the PMN. The EPA was viewed as somewhat less cooperative with respect to determining appropriate precautions for chemicals that it thought might pose an unreasonable risk. Overall, however, the regulated enterprises viewed the EPA as the most flexible, and as having the most expertise. In both respects, this perception runs counter to the dominant view of U.S. regulatory bodies among multinational corporations.

In the EU, company officials characterized member-state agency decision making as generally rapid, although there is variation among the countries. Although the entire notification process in Japan is the slowest, the evaluation procedure by the Chemical Product Committee generally takes place within ninety days, and J-3 reported a thirty-day evaluation process. In the United States, in 90 percent of the notifications the EPA can be relied on to meet its ninety-day deadline. In the 10 percent of cases where the EPA determines there is some risk, several company officials told us that the negotiation of a consent order regarding precautions usually adds three to six additional months.

A. United States

U.S.-based companies and companies in Japan and Europe reported that the EPA's Office of Pollution Prevention and Toxics is helpful, cooperative, and flexible. For example:

---

• An official from US-4 reported that he could go to the EPA before submitting a PMN and ask: “Do you have any concerns? What do you think?” An EPA official might reply “Well, you're going to have a water discharge here, so you should probably do some fish studies.” The company can ask, “Will we get a consent order (requiring precautions or restricting production)?” Though they will not tell you definitively, because they do not have the final authority, they will say, “If you do all of these tests, it will allow us to make a proper risk assessment.”
• The same official from US-4 said, “We can negotiate with EPA in a ‘scientific way’ regarding what testing should be done.” He also gave examples of industries and industry associations working with the EPA to correct the TSCA inventory, for example, by eliminating 1,300 isocyanates that actually boil down to about eighty distinct chemicals.
• European-based companies described the EPA as responsive and helpful but sometimes “very nagging” and “terribly specific” about the precise identity of a substance. One official from EU-5 commented, “If you genuinely want advice on a notification and what kind of testing you should do, they will be quite helpful with that.” Another from EU-2 said that American regulatory officials are more cooperative: “You can go to them, and they can set up a testing program with the company.” An official from EU-3 said he could “phone EPA officials directly, and they respond most kindly.”
• An official from J-2 commented on the scientific expertise of the EPA regulators, which enables them to conduct not only safety and risk assessment but also cost-benefit analysis.

Reports were more mixed in cases in which the EPA determined that there was some risk in the production or marketing of the chemical. US-4 reported that although the EPA worked with the company in developing appropriate precautions, the negotiation took a good deal of time although there was no litigation. US-3 reported that in one case the consent order negotiation took so long that the company missed its “market window” for the product and discontinued production. Similarly, European officials referred to the almost incomprehensible text of EPA consent orders and reported that the additional testing protocols required by them can be quite time-consuming.

B. Japan

The most significant difficulty with agency decision making reported by Japan-based companies is that they have to deal with both MITI and MHW. The company official from J-2 described the notification process in a case in which MITI found the chemical to be safe and MHW thought it should be classified as a designated substance. According to this official, the division

---

of labor between MITI and MHW complicates the process enormously. Moreover, in Japan, there is no postfiling communication with the evaluation committee, and this makes J-2's decision whether to conduct the costly testing more difficult. At the same time, company officials described Japanese regulatory officials as very serious about punctuality and strict compliance with regulations, and as sometimes rigid:

• An official from US-3 commented, “In Japan, they are there to help, but they will not change the rules regarding testing.” For example, for one chemical marketed, Japanese officials asked the U.S.-based company to perform tests that were physically impossible at room temperature. The protocols are rigid, even if the tests will produce “meaningless results.”
• European officials often commented on the need to employ a Japanese intermediary interface with the agencies. “Contacts have to be arranged; they need everything in the Japanese language.” Another from EU-2 said, “I wouldn't say they are unhelpful. But you have got to have a good person in Japan, who is comfortable with the authorities and respected by them.” He added that EU-2 could conduct a toxicity study that involved a lot of animals for an EU notification, “but if we try to get acceptance of that in Japan we would have to repeat that study to include the Japanese protocols, and use the same number of animals again.”
• An official from J-2 said that what chemical manufacturers in Japan are most apprehensive about is unexpected additional testing that the Chemical Product Committee might request without explanation. This kind of uncertainty makes it especially difficult for manufacturers to plan and schedule production.

C. European Union

The implementing agencies of European nations, our respondents agreed, are the speediest decision makers, once the base set of testing has been accomplished. European officials also are said to have less discretion or flexibility than their U.S., and even their Japanese, counterparts. However, the European company officials we talked with complained about differing interpretations of the law and variation in style of regulation among the member states.

• A director of corporate safety for EU-3, a chemical manufacturer in the Netherlands, reported a differing interpretation about whether a particular chemical was on the EINECS inventory. British regulators told a competitor that the company did not have to notify the substances, whereas the EU-3 official was certain that scientifically the chemical was distinct from the one on the inventory, and EU-3 had been required by

  ---

  ― 358 ―
  Dutch regulators, correctly he believed, to file a notification for the chemical.
• A regulatory manager from EU-5 in England described the difference in dealing with the British and the French authorities. In Britain, after the company submits the dossier, the regulator may have some questions, “but once the ‘day one’ of your regulatory review begins, you can be fairly certain you will get to produce your product sixty days later, even if you have more testing to do.” In France, the company submits the dossier, and only after the sixty-day review period do the regulators begin to ask questions. The manager also described more reasoning and negotiation with British regulators regarding hazard and risk, in contrast to the German and Dutch, who simply say, “Do these tests.”
• A company official from EU-2 in Germany discussed his difficulties dealing with the competition among differing German agencies. He described the overall system of regulation as “having no flexibility to speak of.”

Although all company officials reported some problems and difficulty with regulatory officials, they all said that ultimately they comply rather than contest decisions. The overall impact of agency decision making in this regulatory arena might best be summed up by a German company official who said, “As far as we are concerned, if the authorities do not accept our arguments, then we have to do what they ask.” Even in the United States, where access to courts is not difficult, the delays of litigation make legal challenge infeasible for companies eager to bring the new product to market.

VI. ENFORCEMENT

We asked company officials: What happens if you fail to notify the agency properly about a new chemical, either through oversight or through the belief that it is not really “new”? In Europe and Japan, chemical company officials said, there are fewer inspections of company plants and much less proactive enforcement of the laws. Indeed, the company officials in Europe complained about “commercial injustice,” noting that, whereas their companies, large multinational manufacturers, complied fully with the laws, smaller companies and importers, they believed, could get by without doing so. As a demonstration of the problems with enforcement of the EC directives, company officials in Europe pointed to the July 1996 final report of the European inspection project on the Notification of New Substances (NONS). The focus of the study was coordinated company inspections in fourteen member countries, with a concentration on chemical dyestuffs as “an innovative group of substances with the possibility of having inherently

---

hazardous properties and the potential for high risk of exposure to both workers and the environment.”^[34] The inspection revealed that of the almost four thousand substances that were checked at ninety-six companies, 37 percent were new substances that had not been notified and were thus illegally marketed. The report indicated that of the twenty-nine sample chemical substances analyzed, 31 percent did not conform to the information provided by the company. The project found that the controlling authorities thought that forty-five of the ninety-six companies participating were not working according to the EC directives either because they were marketing non-notified substances or because they lacked sufficient labeling, safety data sheets, registration, or internal control.

In Japan, enforcement is accomplished through cooperation between MITI, MHW, and the chemical manufacturers' association, so that the regulator and representatives of the regulated work together in assessing the appropriateness of the regulation, and there is internal pressure for conformance within the industry association. When a violation is found, company officials told us, the attitude of regulators in Europe and Japan is one of helpfulness and cooperation; the most you get is a “slap on the wrist.” As one regulatory compliance official in Europe reported, “If there exists a good, open relationship between company and authority, then it is hardly imaginable that anything can go wrong.” Indeed, there have been no reported fines in Japan. Although warnings are issued to companies that do not satisfy the requirements of compliance with Kashin-hou, the most likely consequence is that the company must send a formal apology (Shimatsusho).

The United States is a very different story. If a company fails to file a PMN, the EPA can assess a fine of $25,000 per day for every day the nonnotified chemical is marketed.^[35] The imposition of such penalties is not uncommon. Under TSCA alone, for the fiscal year 1996 the EPA assessed almost $10 million in administrative penalties, collected over $15 million in injunctive relief, and negotiated over $22 million in supplemental environmental projects (SEPs), nonmonetary actions that parties agree to carry out as part of settlements.^[36]

Several respondents complained about what they saw as EPA's legalistic and punitive enforcement style. According to an official from US-4, there are three main reasons that violations of the PMN requirements occur despite prenotice consultations with EPA: (1) EPA's strict interpretation of the inventory of existing chemicals; (2) purchase of chemicals from foreign vendors who are unwilling to disclose sufficient information for the purchaser, who might simply produce the chemical itself, to verify full compliance with notification; and (3) simple human error resulting from the administrative burden of tracking and notifying hundreds or thousands of new chemical formulations.

---

One U.S.-based company we interviewed had the dubious honor of receiving the highest initial penalty, $22 million, ever assessed for a violation of TSCA resulting from a strict interpretation of the TSCA inventory. The company had refined its manufacturing process for an existing chemical substance, changing the number of carbon molecules. Believing the refined chemical substance was already covered by the TSCA inventory, it did not file a premanufacture notification. EPA officials, interpreting the TSCA inventory differently, disagreed and imposed the fine, based on the number of days the company had produced the refined product. The company found the fine especially troubling because the chemical substance in question, it strongly asserted, posed no particular risk.

Although the initial fine was automatically reduced to $17 million for self-reporting, and the final negotiated amount was $375,000, the frightening impact of that kind of initial penalty, not just in monetary terms but also as a “public relations disaster,” was noted by all of our interviewees. It seems to be the practice of the EPA to announce huge fines that send a shudder throughout the industry, then negotiate the fines to much smaller amounts. For example, in another recent case, a $3.1 million proposed fine was reduced 99 percent in negotiations to $43,400, primarily because of self-reporting and statute of limitations issues.^[37] The negotiations to obtain a reduction, however, involve time, company expertise, and often expensive legal services, all of which are costs to the company. Unlike the notification stage, companies socked with fines can and do challenge EPA administrative actions before administrative law judges and in court.^[38] Of course, companies are not always successful in such challenges, as a company that operates a small metal recycling and shredding facility employing fewer than twenty people found recently when an administrative law judge upheld a $1.3 million fine for leaving unattended for ten years a pile of electrical equipment contaminated by polychlorinated biphenyl (PCB).^[39] In any case, not only are fines more common in the United States, but litigation is more readily available to force the EPA to justify penalties, or to use as leverage to negotiate down the amount of the fine.

A consultant to chemical companies on TSCA issues told us about discovering that the company from which his client had imported a product had not notified all the chemicals included in the product in order to protect trade secrets. The EPA conducted an inspection of the company, and during the inspection the company realized the discrepancy. That is, violations were found, and the importing company self-reported the problem to the EPA, but still received a large fine.

However, as an official from the EPA we spoke with pointed out, the arguable purpose of statutory fines under TSCA is not to redress harms but to create the incentive for companies to accurately report chemicals so that hazards can be identified and minimized, and to ensure compliance with

---

steps imposed to reduce risk. In the experience of this official, who is charged with enforcement of TSCA, there are few instances of willful violations; the only one she could recall involved a person importing herbal medicines. This was consistent with the EPA official's experience that most of the violations of TSCA occur within small firms, rather than large chemical companies, which have elaborate TSCA compliance programs in place. With regard to EPA enforcement philosophy with respect to TSCA, she stated “it is very much a preventative statute,” and the fact that the neglect of a PMN by one firm may pose no actual harm or risk of potential harm to any person or the environment is irrelevant. In her view, if not punished, “the violation, though only technical, would break down the integrity of the system for preventing the bad chemicals, like PCBs, from getting out. TSCA is about potential for harm or risk, not the harm or risk of this chemical.” From the perspective of this EPA enforcement officer, the criteria for evaluating TSCA enforcement should be whether it is “consistent and efficient,” not necessarily how much it reduces harm or risk.

The tendency to penalize violations, even if they are unintentional and do not lead to harm, is in keeping with multinational corporations' general perceptions of the distinctive legalism of the U.S. regulatory enforcement agencies. Notwithstanding the EPA officials' justifications, it has an alienating effect in the chemical industry. This alienation is exacerbated by the perception of some in the chemical industry that the EPA is more concerned with symbolic politics than with serious risk reduction. It was not only the initially large amounts of the punitive fines under TSCA that disturbed company officials but also what they described as the “political motivations” of the fines. According to an official from US-1, enforcement at the EPA can change depending on the political winds of the administration, making it difficult for companies to predict what the agency may want in the future. This official characterized the “command and control” approach of the EPA as a “joke,” because at the notification stage it required no serious testing yet focused on insignificant minutiae at the enforcement stage. He contrasted this with the certainty and predictability of the European system and the willingness to work with companies that were out of compliance.

VII. EFFECTIVENESS

In each jurisdiction, new chemical notification laws exist alongside laws regulating worker safety, requiring dissemination of material safety data sheets (MSDS), delineating proper waste disposal, regulating air and water emissions, and mandating cleanup of contaminated sites. To the extent that there has been a reduction in exposure to toxic chemicals, it is virtually impossible to sort out the relative contribution of new chemical notification

---

laws, and hence it is impossible to compare the relative effectiveness of TSCA, the EC directives, and Kashin-hou.^[40] Incentives to restrict potentially dangerous chemicals and the way in which they are used also are generated by product liability laws and the prospect of costly lawsuits that they can generate. Moreover, there is disagreement among the environmental and scientific communities over what level of toxicity or exposure presents an unreasonable risk,^[41] as well as disagreement among regulators and regulated about what positive effects to expect from notification laws. For example, with regard to TSCA there is disagreement over the relative importance of the premanufacture notification goals of toxicity information collection, dissemination of toxicity information to the public, and actual prevention of harm, as well as the focus on new chemicals as contrasted with testing existing chemicals.

The effectiveness question is further complicated by the varying rationales that underlie each new chemical notification system. In Japan the laws are targeted primarily at preventing harm to humans from the consumption of contaminated marine food sources. In Europe the EC directives focus on preventing the potential harm to health and the environment from chemicals manufactured and released in large volumes. In the United States the focus seems to be on preventing the marketing of chemicals that are similar in properties to chemicals that are already known to be carcinogenic or harmful to the environment, such as PCBs.

One way to address effectiveness might be to note whether particular chemicals that have been approved and placed on the inventory in one regime (e.g., the United States) have been not approved or restricted in another. We did not find such an inquiry feasible, however. There are over seventy thousand chemicals on the TSCA inventory, and the European inventory is static, that is, approved chemicals are not added to the inventory. We did ask our interviewees about any instances in which one country approved a product without restriction while the same chemical was rejected or restricted under another country's regulatory system. Unfortunately, this question did not offer us much helpful information to assess effectiveness. First, the company officials, especially those directly responsible for health and safety issues, noted that their firm had no interest in manufacturing or marketing unsafe chemicals. This was especially true in the United States, where product liability concerns are paramount. Thus, the officials told us, if there was regulatory concern about a chemical in Europe or Japan, often the company never notified it in the United States or withdrew its notification. On the other hand, chemical company officials rejected the implication that more stringent regulatory requirements in a particular jurisdiction necessarily implied greater effectiveness, which they seemed to define in terms of reducing actual risks. Thus they argued that the mechanical application of EU testing rules that focus particularly

---

on the tonnage of the amount of the chemical produced, or of Japanese rules that sought maximum testing of any chemical that was toxic in any degree to marine organisms, did not necessarily make those systems more “effective,” in terms of actual risk reduction, than the U.S. regulatory regimes.

Nevertheless, if the U.S. system of notification of new chemicals is cheaper and quicker than that of Europe or Japan, and seeks to avoid Type II errors of overregulation, the obvious question is whether it incurs Type I errors of excessive leniency as a result. Although there are no “dead or maimed bodies” to point to, there has certainly been criticism of TSCA and its effectiveness. A 1994 General Accounting Office (GAO) report commissioned by the U.S. Senate Subcommittee on Toxic Substances, Research and Development and the House Subcommittee on Environment, Energy and Natural Resources argued that (1) the legal standards of TSCA—requiring the EPA to demonstrate unreasonable risk before taking action—were so high that the EPA has been able to issue comprehensive regulations to control only nine chemicals in eighteen years;^[42] (2) the EPA review process of new chemicals does not adequately ensure assessment of potential risks and exposure before and after manufacture; and (3) the excessive use of the confidential business information provisions makes the information collected under TSCA less available and useful to state agencies and the public. But the GAO's assessment is not without controversy because it is unclear what criteria EPA should have applied or what social benefits “doing more” would have generated.^[43]

The most significant factor making TSCA cheaper and quicker than the notification systems of either Europe or Japan is the EPA's use of SAR analysis, rather than mandatory testing, to screen chemicals and determine if there is sufficient risk to warrant additional testing. In 1993 the EPA conducted a study with the EU asking what difference European-style testing would make.^[44] The physical properties and hazards identified through SAR for 144 chemicals were compared with properties and hazards identified using test data required under the EC directive. The study found that the accuracy of SAR varied depending on the chemical property tested. SAR was highly effective (93 percent) for assessing biodegradation but only 50 percent and 63 percent effective for predicting boiling point and vapor pressure. EPA officials concluded that SAR was effective in screening new chemicals to determine if there is need of further testing or regulation but less effective in predicting the exact type and level of toxicity, especially for general health effects.^[45] EU officials noted the variation in SAR effectiveness in assessing chemical properties and found SAR to be insufficiently developed for determining eye and skin irritation or sensitization, but more successful at determining acute lethal toxicity.

When we asked chemical company regulatory compliance officials

---

directly about perceived effectiveness of the different systems, their responses were inconclusive. Whereas company officials consistently described the European system as the most expensive, primarily because of the mandatory base set of testing, some believed that no “added safety” came with the high price tag, and others viewed the European system as the safest and most comprehensive. Some company officials viewed the length of time for the required testing of the Japanese system as a costly delay with possibly no corresponding precautionary benefit in terms of safety. Some company officials believed the U.S. system—with no mandatory testing under TSCA but the possibility of testing if there is potential risk—was as effective, if not more so, at preventing the marketing of harmful chemicals.

Two factors contributed to the interviewed officials' estimation of the benefits of the U.S. system: (1) the EPA's commitment to scientifically appropriate risk assessment, and (2) broad EPA discretion to implement precautions. The EPA conducts risk assessment based on the intended use and volume of the chemical. If the EPA finds an unreasonable risk with a particular chemical, it has the discretion to develop mandatory precautions with the company to protect against that risk. Examples of precautions included mandatory wearing of gloves, warning labels, and use of air respirators. Sometimes the EPA also issues what the company official from US-3 referred to as “killer 5(e)“s”; these include requirements such as labels stating that a substance is a “cancer-causing agent,” or precautions, such as mandatory use of an air respirator, either of which might make the customer not want to buy the product. On the other hand, TSCA requires the EPA to consider the economic benefits of the chemical and the costs of the proposed precautions in determining what precautions to impose.

In contrast, testing in the European system is based entirely on tonnage of production per company; the presumption is that increased tonnage increases the risk. The U.S. system considers volume of production, but only as it contributes to the risk of the individual chemical reviewed. In the view of one European official, U.S. regulations are more concerned with the purpose for which the chemical will be used, and they use tonnage as a criterion proportionate to risk. In his estimation, this is a better way of assessing the costs and benefits of additional testing.

The potential fines that the U.S. enforcement system employs made companies more wary of committing violations than did the EU system. In some respects, this may make TSCA more effective by compelling more attention to compliance. Thus, chemical company officials also told us that as a consequence of TSCA's punitive enforcement system, there are increased costs related to more stringent intracorporate review for compliance with TSCA. This seemed especially true for companies exporting chemicals to the United States. For example:

---

• A European company official reported that whenever a chemical is sent to the United States, the company blocks the export and allows only “authorized people in the firm” to release it. Accordingly, “no chemical that is placed on the market goes to America without compliance with TSCA requirements.”
• A regulatory compliance official estimated that assuring compliance with TSCA requires more manpower than the European system, because any product to be sold in the United States that might be composed of new chemicals undergoes a very stringent review procedure in the corporation's regulatory compliance section.

Considering all new chemicals notified under the three regimes, there are clearly more test results collected under the European and Japanese systems than in the United States. But the essential question for assessing effectiveness in preventing harm is not simply whether more data are generated, but what is done with the data that are collected and whether there is sufficient information to assess harmfulness. Whereas more information might always be preferred, there is no indication that the approach of TSCA—screening chemicals before selectively requiring additional testing—has significantly increased the risk to the public from the marketing of toxic chemicals.

VIII. CONCLUSION

Although the U.S. regulatory system in general tends to be designed in a more adversarial and legalistic manner than comparable legal regimes in Japan and Europe, that is not the case with regard to new chemical notification laws, as experienced by numerous officials of multinational chemical companies. They do observe in the United States, however, the characteristic American tendency toward legalistic and punitive responses to violations of the detailed and prescriptive rules, even if the violations are unintentional or not harmful.

The surprising result is the greater flexibility and efficiency of the American regulatory regime. In 1976, when TSCA was enacted, the regulatory regime for premarketing approval and testing of new pharmaceutical products, administered by the FDA, was criticized for unduly delaying provision of valuable new drugs and deterring innovation. One speculative explanation for TSCA's flexibility is that it was a response to this FDA criticism.^[46] The TSCA regime deliberately provided more flexibility by (1) making requirements for extensive testing discretionary, according to the judgment of officials in the implementing agency (EPA); (2) stipulating that the EPA must make that judgment within ninety days; (3) providing exemptions for

---

test-marketing and use of new chemicals for research purposes; and (4) providing EPA with flexible powers to impose custom-tailored postmarketing restrictions on handling, labeling, and use of those chemicals it determines pose special risks. At the same time, the TSCA regime included some features of the pharmaceutical regulation regime that tend to promote regulatory flexibility: rather than delegating implementation and enforcement to state and local government agencies, which often lack expertise and are thought by some to require federal oversight,^[47] implementation is assigned to a centralized cadre of officials in Washington, D.C., who have or acquire significant expertise. The EPA's enforcement system is strict and legalistic partly, perhaps, as an offset to the more pragmatic, less stringent substantive regulatory requirements, and partly as a response to the more open U.S. economy, in which there are many small chemical manufacturers and importers, which are difficult for the U.S. agency to monitor and perhaps cannot be trusted to comply.

In contrast, the European system, which had not experienced the political controversy over product delays in pharmaceutical regulation, became more concerned with promoting harmonization of regulation across member states; hence it sought to restrict member state discretion by imposing mandatory requirements, such as the universal base set of testing. Japan's Kashin-hou, reflecting a political response to a series of environmental and human tragedies stemming from chemical pollution, reflects a stricter and more conservative attitude toward certain risks. Structured to avoid Type I errors (of excessive leniency), it probably makes more Type II errors, delaying dissemination of useful new chemical products. Kashinhou's focus on biodegradation and bioaccumulation, moreover, may make it more tolerant of other, equally troubling risks. In addition, weak enforcement systems in Europe and Japan may well produce more Type I errors in the manufacturing process than does the American system.

The impact of the operational differences between the three notification systems, our interviews indicate, is smaller for large, multinational chemical manufacturers that consistently market a smaller number but larger volume of chemicals worldwide. These companies tend to find it most cost-effective to internally harmonize the requirements for notification in different regimes by adopting the highest standard, generally the European. Such internal harmonization efforts are consistent with the efforts of organizations such as the Organization for Economic Cooperation and Development to harmonize testing protocols and good laboratory practices, and to encourage mutual acceptance and sharing of data.

Variation in the cost of testing and the delay associated with the notification of new chemicals has a much more significant impact on manufacturers of specialty chemicals with smaller markets. Moreover, to the extent that costs and delays in the development of new chemicals dampen innovation

---

in the industry at large, they may in fact have an adverse impact on health and environment. As a company official with EU-1 argued, an effective system must balance the need to protect the health and environment through premarket testing with protection gained from the introduction of improved substances in the market. According to this official, the costs and delays under the European system have significantly discouraged innovation by smaller and medium-sized companies that might have led to newer substances with more desirable properties in terms of health and environmental protection than older, existing chemicals.

Comparison of the costs, delays, style of administration and enforcement, and effectiveness of the new chemical notification schemes of Japan, the United States, and Europe has implications for the rest of the world as well. The developed nations' regulatory schemes provide the primary models for the many developing countries implementing similar notification systems. Indeed, we heard a veritable chorus from corporate officials that their companies were most concerned about the impact of new notification systems in countries like Korea, China, Australia, and New Zealand. Whereas there is general, and obviously growing, acceptance of the need to address and attempt to prevent the harmful effects of potentially toxic chemicals through some type of premanufacture or premarketing screening device, the means employed, as this chapter illustrates, can have significant negative impacts on multinational chemical companies without necessarily increasing effectiveness.

NOTES

We thank our interviewees for their willingness to speak frankly with us and for their generosity in giving us access to their time, expertise, and experience. Mr. Fujie especially wishes to express his appreciation to the government officials, corporate executives, and scholars in Japan who assisted him in his research there and requested anonymity in order to provide their unreserved opinion.

	TOTAL NOTIFICATIONS SUBMITTED
	Domestic: 4,619 Imported: 1,580
	RESULTS OF NOTIFICATION
	Safe: 4,930
	Designated: 344
	Other: 1,169
	(including withdrawn notifications and first class specified substances)

---

12. The Consequences of Adversarial Legalism

Robert A. Kagan

What conclusions can be drawn from the case studies in this volume? Rich in detail, varied by area of law and regulated industry, often containing stories with contradictory outcomes or implications, they show that the world does not always conform to any parsimonious theory. The lessons of the cases, therefore, must be constructed inductively. Generalizations sometimes must be offered tentatively and with significant qualifications and exceptions. This effort to order the findings is organized in sections that address (1) the effectiveness of law in shaping the behavior of large corporations; (2) trends in economically advanced democracies toward convergence in the substantive norms of regulatory law; (3) the continued persistence, salience, and uniqueness of American adversarial legalism; (4) the substantial social and economic costs of adversarial legalism; and (5) adversarial legalism's relatively minimal compensating social benefits.

I. LAW MATTERS

In the regulation of large business firms in economically advanced democracies, law ways can and do change folkways. That message is conveyed by virtually all the case studies in this volume. All the companies studied relate to the relevant laws and legal controls with great seriousness. They maintain substantial staffs of regulatory compliance specialists and lawyers. With rare exceptions, they comply with those laws. Although the regulatory compliance personnel interviewed occasionally offered criticisms of particular regulatory requirements, they did so surprisingly infrequently. To a considerable degree, one can say that corporate officials in these large, multinational companies have accepted—in some cases even “internalized”—the basic norms of the regulatory programs and legal regimes they

---

encounter in economically advanced democracies, and they take the specific rules and institutions that implement those norms as a fixed constraint, part of a business environment that they must try to adapt to as best they can.

This finding has become commonplace. Commenting on their findings in a recent study of fifty firms subject to the British Integrated Pollution Control (IPC) regulatory regime, Alex Mehta and Keith Hawkins observe:

It seems reasonable to suggest that the days of companies consciously polluting the environment for purely financial motives (because it is easier to pay a fine than the costs of compliance) may well be disappearing. Indeed, most managers interviewed agreed with the principle of IPC and with the broad aims of pollution control in general. Many also mentioned a strong sense of moral and legal obligation supporting compliant behavior. No manager thought the IPC regime represented an unreasonable or in some sense illegitimate intrusion by the law.^[1]

Of course, an attitude that views regulation as legitimate does not mean that corporations, or particular subunits, managers, or employees within a corporation, will attend to or adhere faithfully to every one of the myriad regulations that govern contemporary business behavior. But large corporations have entirely self-serving, profit-oriented reasons to adopt a general policy of full compliance. As Mehta and Hawkins found,

Big firms were intensely concerned about the public and political consequences of transgression. Being squarely in the public eye they thought that news of their transgressions which would come to light if they were every prosecuted (regardless of whether they were formally punished) would seriously stain their corporate image. This assault on their reputation could have repercussions for sales and income. …

It is immaterial whether such consequences in fact flow from prosecution. What is more important is that managers believe they do. … The result is that many big firms feel compelled to comply with IPC requirements because they cannot take the risk of their carefully cultivated (and expensive) consumer image being exposed to the uncompromising stare of the media. The bargaining chips held by environmental regulators are the threat of prosecution coupled with exposure to the glare of harmful publicity, which suggests that they enjoy an effective power that exceeds the fines that might be imposed upon a conviction for a violation of IPC requirements.^[2]

With the exception of the Japan-based multinational metal parts manufacturing company studied by Aoki and Cioffi (chapter 2), each of the companies in the other case studies, as well as virtually all of the chemical companies interviewed by Johnson, Fujie, and Aalders (chapter 11), are household names in the United States, Western Europe, or Japan. They all market widely used consumer products or services under their corporate

---

name. Hence, they all appear sensitive to precisely the kind of reputational risk that Mehta and Hawkins write of, and tend to take a “minimax” attitude toward legal or regulatory noncompliance: noncompliance is regarded as risky because it is impossible to predict whether a worst-case scenario—a highly visible environmental problem, or injury to a large number of consumers, or a multi-million-dollar lawsuit—will ensue.

The regulatory risk-aversiveness of large, visible corporations is reflected in many of the case studies. Aoki, Axelrad, and Kagan (chapter 3) report on a prominent Japanese corporation's deep concern that it should never be found to have violated official water pollution standards. Consequently, it established “internal standards” for contaminants in its wastewater that are substantially more stringent than those required by law or regulatory permit; by monitoring compliance with its internal standards, company officials would have an early warning system against violation of the regulatory standard. The multinational pharmaceutical corporation studied by Nielsen (chapter 7) prided itself on having an internal set of rules governing employee discipline and dismissal that would exceed, in terms of due process norms, the legal and regulatory standards of the jurisdictions in which it operated, and on having never lost a lawsuit by a dismissed employee in either the United States or Canada. When the large chemical company studied by Axelrad (chapter 4) became aware of site contamination at one of its manufacturing facilities in Great Britain, it immediately implemented the same set of investigative procedures it uses in the United States, although there are no specific British regulations on the subject. The metal parts manufacturer studied by Aoki and Cioffi (chapter 2) developed a cleaning system in its U.S. plant that dispensed with the use of chlorofluorocarbons (CFCs) well before the deadline called for by the Montreal Protocol and quickly implemented it in its plants in Japan and Asia, although there was no legal requirement there to do so. Thus these case studies provide some evidence for a dynamic toward transnational, “corporation-level” harmonization of regulatory compliance routines in multinational companies, keyed to compliance with the most stringent national standards (sometimes with a margin of error).

Corporate-level “harmonization” of compliance measures, the case studies suggest, is not automatic in every regulatory sector. Although U.S. law required Waste Corp. (chapter 5) to install a “double liner” at its municipal waste facilities in California and Pennsylvania, the company did not volunteer to do so at its facilities in England and the Netherlands, where regulators accepted a different design. Ford Motor Company (chapter 6) moved much more quickly to reduce paint shop odors in the surrounding community (by building a taller exhaust stack) at its plant in Cologne, Germany, than it did in St. Paul, Minnesota. And sometimes “harmonization” is thrust on a company by the extraterritorial reach of national

---

regulatory laws. Kraus (chapter 10) shows that a biologics manufacturing plant in the United States was compelled to adopt tighter tolerances on its processes than had been required by the FDA because it could not market its U.S.-made product in Europe without satisfying the tight tolerance demanded by the European Union (EU) regulatory body. Similarly, as noted by Johnson, Fujie, and Aalders (chapter 11), chemical companies that seek to market new substances in Japan, Europe, and the United States often undertake the expensive toxicity testing required by Japanese and European regulatory agencies, even for products for which the U.S. EPA decides no such testing is necessary.

A disposition toward regulatory and legal compliance by visible multinational corporations does not imply that the meaning of “compliance” is always clear, or that these companies are unwilling to battle agencies or even litigate to obtain favorable interpretations of the law. The U.S. subsidiary of the Japanese metal manufacturing company studied by Aoki and Cioffi (chapter 2) spent thousands of dollars on a legal dispute with state and federal regulators concerning the cleanup of contamination that had been caused by the previous owner of its U.S. plant. A Japanese electronics company (chapter 3) disputed a local agency's method of sampling for water pollution outside the factory. ACME (the chemical company discussed in chapter 9) engaged in extended negotiations and litigation in the United States, Japan, and the EU alike when patent examiners initially did not find its innovation patentable. Ford spent years arguing with New Jersey regulators about the legally required and technically appropriate way of controlling emissions from a paint shop in its assembly plant (Dwyer, Brooks, and Marco; chapter 6). In short, the technical nature of the problem subject to regulation—and the lack of legal specificity about the level of small or remote (but costly to control) risks that can be tolerated—often makes “compliance” a matter of contention. And multinational companies, well armed with experts and money for lawyers and consultants, are formidable actors in the struggle to define legality.

Nevertheless, the dominant impression conveyed by the case studies is that the regulatory officials, not the regulated corporations, had the upper hand in the negotiations about regulatory requirements. In California, Pennsylvania, England, and the Netherlands alike, Waste Corp. was compelled to reformulate its design for a solid waste disposal site to meet regulators' and neighbors' environmental objections, and also to provide costly additional environmental amenities that were only loosely related to the problems the garbage dump created (see chapter 5). Ford Motor Company was forced by regulators and neighbors in both Germany and the United States to build higher exhaust stacks and adopt a more expensive pollution control method (carbon absorption) than it had proposed in its initial permit applications (chapter 6). When regulatory officials at the FDA

---

or the EU agency for regulating biologics demanded more information or tighter tolerances, regulatory affairs personnel in “Company Y” burned the midnight oil to give them what they asked for, whether or not they personally regarded it as reasonable (chapter 10).

The case study method employed does not enable us to assess the other side of the coin. We cannot tell whether or when, in any of the studies, the regulators agreed to a company proposal or permit modification or argument that they “should not have accepted” by some legal or public-regarding standard. But the tone of all the case studies provides no ground for thinking that the corporate officials were able to overwhelm the regulators with argument or expertise, or with political or economic leverage. Above all, it appears that it was the regulatory officials' view of the law that mattered most.

II. SUBSTANTIVE SIMILARITY, SUBSTANTIVE DIVERGENCE

Linked by a global communications system, political elites, advocacy groups, and electorates in all economically advanced democracies lament the same kinds of injustices, environmental harms, and technological hazards. Regulatory officials, scientists, legal scholars, and environmental activists all flit across borders, in person and electronically, sharing knowledge about risks and policy solutions. When the European Union and Japan promulgated legal standards for product liability lawsuits, the standards were remarkably similar to those established by state courts in the United States.^[3] A comparative study by scholars at Resources for the Future stated, “Although U.S. environmental regulations are arguably the most stringent in the world, the differentials between U.S. standards and those of our major industrialized trading partners are not very great.”^[4]

In this respect, the case studies in this book contain no big surprises. Most multinational corporate officials and lawyers interviewed in the course of the Comparative Legal Systems Project claimed that the substantive legal and regulatory standards they encounter in the United States and in other Organization for Economic Cooperation and Development (OECD) countries are basically similar. European representatives of an international petrochemical company told us that safety regulations in Europe and the United States concerning the labeling, packaging, and land transportation of hazardous chemicals were fundamentally the same. Project researchers encountered few criticisms of the United States for enacting safety and environmental regulations or other substantive legal rules that are significantly more stringent than the obligations imposed in other economically advanced democracies.

Most (but not all) of the detailed cases studies, therefore, support the “competitive convergence” hypothesis at the level of basic substantive legal

---

policies and norms. At the same time, they highlight more subtle differences at the level of regulations and administrative decisions that implement those norms. Thus Dwyer, Brooks, and Marco (chapter 6) found that both German and American air pollution standards restricting air pollution from motor vehicle plant paint shops are stringent, but they note that Germany's “statutory ‘state of the art’ requirement” for control technologies “is based on the proportionality principle, under which benefits and costs are balanced,” in contrast with the American lowest achievable emission rate (LAER) standard for control technologies in high-pollution “nonattainment areas.” The American regulations thus were more stringent in the sense of allowing regulatory officials less discretion in adapting general standards to particular situations and to the economic cost of preventive measures. A similar tendency for U.S. regulations to be more inflexible (and hence more stringent) surfaced in many other case studies, such as Axelrad's comparison of British, Dutch, and American regulations governing investigation and remediation of contaminated industrial sites (chapter 4) and Welles and Engel's comparison of the detailed prescriptions for the liners in American municipal waste disposal sites with the more flexible site-specific standards that are employed in the parallel Dutch and British regulatory regimes (chapter 5).

A few case studies turned up more significant differences in substantive law, but differences that did not seem to be particularly significant in practice. The Canadian law against arbitrary discharge of employees, as described by Nielsen (chapter 7), is more comprehensive in its coverage than the patchwork of targeted antidiscrimination statutes and state common law rules in the United States, and Canadian law provides employees with more protection against sudden firing. But when Nielsen compared actual practices in a pharmaceutical company's Canadian and American operations, she found that these substantive differences were less significant in shaping personnel administration and postdismissal legal conflict than certain procedural features of American adversarial legalism, such as the greater risk of a very costly and disruptive civil damage suit in the United States, or the greater likelihood of an investigation by the federal Equal Employment Opportunity Commission (EEOC) or a state antidiscrimination commission. Similarly, according to Aoki and Cioffi (chapter 2), Japanese statutes provide less comprehensive restrictions on industrial waste disposal than does American law, and Aoki, Axelrad, and Kagan (chapter 3) note that the effluent standards under Japan's water pollution control law are, on average, somewhat less stringent than those imposed by federal and state regulations and permits in the United States. Nevertheless, Aoki and his coauthors found that due to informal “administrative guidance” and agreements with local governments, the Japanese factories studied actually

---

controlled industrial wastes and effluents at least as thoroughly as similar factories in the United States that were owned and operated by the same corporation's U.S. subsidiary.

In one case study, substantive legal differences were clearly important in terms of actual impact. Charles Ruhlin (chapter 8) examines a multinational bank's experience in attempting to collect delinquent consumer credit card debt in Germany and the United States. In both countries, consumer debt obligations are governed primarily by contract law, which is basically similar across countries. But in the United States, the bank encounters laws that grant debtors defenses and rights not available to German consumers: (1) federal and state “fair debt collection practice” laws that regulate creditors' dunning techniques, and (2) federal law that makes it relatively easy for debtors to discharge debts (while preserving some assets) by filing for bankruptcy. In consequence, Ruhlin found, the bank is far more cautious about bringing legal action against defaulting debtors in the United States, and it ends up treating a larger proportion of delinquent debt as uncollectible. Yet the bank's greater reluctance to take debtors to court in the United States also is due in part to the procedural characteristics of American adversarial legalism—which make American courts and collection practices far more inefficient, costly, and unpredictable than the less formal debt collection procedures provided by German courts. Ruhlin tells us, however, that Germany has now enacted a law expanding opportunities for consumer bankruptcy. And as this is written, a law curtailing some credit card debtors' bankruptcy rights is moving through the U.S. Congress.

In summary, the studies suggest a good deal of convergence of substantive regulatory norms and standards in economically advanced democracies, combined with continuing divergence in legal or regulatory style. For even when national laws are substantively similar, the chapters indicate, they often are applied and interpreted differently. Japanese and German patent examiners, Deepak Somaya's case study (chapter 9) indicates, are more conservative than their U.S. Patent Office counterparts in agreeing that an application meets the “innovativeness” standard, and the Japanese Patent Office is more insistent that the applicant's claims concerning inventions must be narrowly stated. The chemical company Somaya studied obtained a U.S. patent for a breakthrough process but could not do so in Japan, and it succeeded in Germany only after great difficulty. Similarly, officials in the company studied by Martine Kraus (chapter 10) noted that European regulators adopted a more searching analytic approach than their counterparts at the U.S. Food and Drug Administration (FDA) in evaluating the safety of a process for manufacturing a genetically engineered biological product. EU officials also insisted on tighter tolerances in the manufacturing process. Such differences in what might be called “legal or regulatory”

---

style generally loomed far larger in the experience of the multinational companies studied than did cross-national differences in substantive law.

III. THE SALIENCE AND DISTINCTIVENESS OF ADVERSARIAL LEGALISM IN THE UNITED STATES

According to the chief environmental attorney for a multinational chemical firm, the chief contrast between the legal systems of the United States and other economically advanced democracies is not that the laws or standards are different but that “cultural processes” are different. The salient features of the American process he identified are familiar components of adversarial legalism: more complex and prescriptive bodies of legal rules; a more fragmented system of authority, which means that a company often must satisfy many agencies and respond to scattered claims in different courts; broader provisions for public participation in and influence on administrative decision making; greater uncertainty concerning legal requirements; and longer decision-making delays. For example, this attorney noted that American air pollution permit processes for new or renovated chemical manufacturing facilities, when contrasted with permit systems in other economically advanced democracies, involve more jurisdictional stops, more opportunities for local opponents to appeal to court, and more types of decisions in which approvals must be sought from both federal and state regulators, who often do not agree. This procedural legal maze, in his view, is far more significant than the minor cross-national differences in actual emission standards.^[5] Many of the same dimensions of America adversarial legalism are highlighted by the case studies in this book.

A. Public Participation

Welles and Engel (chapter 5) recount the experiences of a multinational waste management company in obtaining permission to develop or expand solid waste disposal landfills in California, Pennsylvania, England, and the Netherlands. In each of these jurisdictions, permits must be obtained both from a local governmental body that regulates land use and from at least one environmental protection agency. In each jurisdiction, the law provides for some form of participation or input from neighbors or from members of the public who object to the location or design of the landfill facility. But the American legal system, particularly in California, gave local objectors far more legal leverage. If their concerns were rejected in one forum, there were more agencies to which they could complain, and the law gave them greater capacity to challenge adverse administrative decisions in the courts.

Similarly, in both Germany and the United States, the air pollution

---

control system requires companies to apply for permits for process changes that may increase pollution; it also provides for public comment and objections with respect to permit conditions. But as Dwyer, Brooks, and Marco demonstrate (chapter 6), opportunities for public participation under German law are more limited:

In the United States, federal and state law requires every permit decision to be preceded by public notification, an opportunity for public comment, and a meeting where members of the public can voice their concerns and demand answers from regulators (and often company officials). Typically, state implementation plans provide an opportunity for contested administrative hearings (as well as judicial review) to resolve factual disputes underlying the permit decision. Entrepreneurial activists can use these procedures to generate publicity, create and energize a political constituency, and force the agency to address their concerns.

Under German law, … the Land environmental agency has no legal obligation to give public access to the proposed permit and supporting documents … so long as the modification does not increase pollutant emissions. … Even when a German environmental agency must make the draft permit and supporting documents available to the public, … German agencies accept written comments and objections from members of the public, but they are not required to hold a public meeting or initiate a contested administrative hearing.

Thus neighborhood groups and environmental activists in St. Paul, Minnesota, Dwyer and his colleagues report, contested Ford Motor Company's permit application and used public meetings to generate political pressure on the state environmental agency, which created a task force, with citizen group participation, that ultimately resulted in a plant modification to control objectionable odors. In Germany, neighbors filed only written objections concerning odors from a Ford assembly plant's paint shop; although this process resulted in similar plant modifications, Ford's German neighbors did not get to sit on a task force like the one created in St. Paul, which got to hire its own consultant, review the options in detail, and help shape the solution.

B. Legalistic Regulatory Enforcement

For years, observers have criticized many regulatory programs in the United States for employing a legalistic enforcement style, that is, for rigidly enforcing detailed regulatory requirements, declining to make exceptions based on the risks and compliance costs peculiar to the particular case, and for measuring success in terms of numbers of prosecutions and amounts of fines. Legalistic enforcement, it has repeatedly been said, often results in unnecessary regulatory requirements and stimulates resistance on the

---

part of regulated entities.^[6] The Clinton administration's Reinventing Environmental Regulation report of 1995 stated, “The adversarial approach that has often characterized our environmental system precludes opportunities for creative solutions that a more collaborative system might encourage.”^[7] Thus academic analysts, regulatory officials, and political leaders regularly have called for more flexible methods of regulatory enforcement and for redesigning regulation to encourage industry cooperation.^[8] Many efforts to institute such reforms have been reported, especially in the 1990s.^[9] It was therefore somewhat surprising that in the Comparative Legal Systems Project, multinational corporation officials repeatedly referred to—and expressed fear of—legalistic, punitive enforcement of environmental and safety regulations in the United States. A pollution control consultant with wide experience said, “In the United States “if you screw up, you can have stiff fines. You can readily get sued.” In other countries, such as the United Kingdom, “they have fines, but they're peanuts.”

Adversarial enforcement is not universal in the United States. It is not difficult to find cases in which state or federal regulatory agencies declined or simply failed to punish detected regulatory violations.^[10] In some industries, such as meat and poultry processing, companies have managed to ward off legislation calling for large and more automatically imposed administrative fines for violations of safety regulations.^[11] Several corporate officials told us that they enjoyed cooperative, nonlegalistic relations with certain American regulatory agencies. But many corporate regulatory affairs officers told us of incidents in which violations of “prophylactic” regulations (reporting, labeling, design of storage containers, etc.)—violations that entailed no actual harm to workers or to the environment—were met with automatic fines amounting to tens of thousands of dollars. For example, in their study of premarketing regulation of new chemical compounds, Johnson, Fujie, and Aalders (chapter 11) recount the experience of a U.S. chemical firm that had begun marketing a new substance which company chemists believed was essentially identical to others on the long list of substances that had previously been cleared by the U.S. Environmental Protection Agency (EPA) pursuant to the Toxic Substances Control Act (TSCA). Company officials, however, sought confirmation of their action from the EPA. EPA officials, it turned out, disagreed with the company's interpretation of the scope of the prior clearances and found that the company had violated TSCA. Despite the fact that the company had been forthcoming and that it had at least a plausible scientific argument, the EPA assessed a $23 million fine (although it subsequently reduced it to $375,000). Another instance of legalistic, adversarial enforcement, this time by a state environmental agency, is the centerpiece of Aoki and Cioffi's comparison of industrial waste regulation in the United States and Japan (chapter 2).

---

In contrast, in the experience of the corporate officials interviewed, regulators in Western Europe, the United Kingdom, and Japan rarely respond to regulatory violations with formal enforcement and fines so long as the violations are not willful and the regulated company seems genuinely committed to remediation and improvement. The contrast is conveyed by a story told by a regulatory compliance officer in a U.S. subsidiary of a multinational petrochemical company. It arose in a meeting with her counterparts from corporate subsidiaries in the United Kingdom, Canada, and Australia. These nations all have laws requiring chemical companies to maintain detailed inventories of chemicals in process and in shipment. But given the volume and rapidity of shipments in a large chemical company, said the U.S. official, it is “extremely easy to make a mistake.” When a British regulatory compliance official discovered he had made such an error, his attitude, the bemused American official said, was, “Oh dear, I think I'll have to do something about that,” indicating that a call to the relevant agency would take care of such a nonwillful, not immediately harmful violation. In contrast, the American official said, if the U.S. subsidiary's compliance office discovered a similar error, “We'd panic. We'd start calculating the fine we owed,” because reporting can reduce the statutory fine (up to $25,000 a day) by half.

In other countries, regulators were said to be much more inclined to sit down with company officials to work out agreements on technically appropriate remedial measures. Axelrad (chapter 4) describes how Dutch and British regulators informally worked with the regulated company to determine what steps should be taken to investigate the extent of ground and water contamination at an industrial site and how it should be contained and remedied. Both state and federal regulators in the United States, in contrast, inspected, consulted the detailed official regulations, and subsequently mailed the company orders about what it should do—orders that in significant measure were unrelated to the actual tasks of remediation. “In the United States,” company officials told Axelrad, “there's no sitting down with the agency and saying, ‘Let's focus on risk.’”

Regulatory inflexibility in the United States reflects the legal structure and political environments within which American regulators work. Legalistic enforcement is rare in the Japanese system for regulating factory waste disposal, Aoki and Cioffi point out, partly because Japanese waste control regulations are more outcome-oriented—in contrast with American regulations under RCRA, which prescribe in the minutest detail what the company must do to prevent environmental contamination. American air and water pollution permits, our respondents repeatedly observed, are more detailed and specific than those in Japan and Europe. As described by Welles and Engel (chapter 5), federal regulations concerning municipal waste sites in the United States are much more prescriptive, technical, and

---

inflexible than the standards in the Netherlands and Great Britain, which Dutch and British officials can tailor to the particular site in light of evolving technologies and the perceived risks at the particular location. Noting that American regulators acted legalistically even when RCRA rules concerning industrial waste cleanup allowed them discretion, officials at one multinational company told Axelrad (chapter 4) that in his experience American regulators want to issue a permit that is “risk-free to themselves” because they operate in an atmosphere in which they are “pilloried and abandoned if it turns out that they have made a mistake.”

As that observation suggests, behind the greater specificity of American regulations, and hence behind the propensity toward legalistic enforcement, lie cross-national differences in political culture—particularly the greater mistrust of governmental bureaucracy in the United States, which is expressed in the restriction of bureaucratic discretion by detailed rules and ready access to judicial review. In this political environment, “going by the book” provides regulatory officials with protection from at least some kinds of criticism. In countries that are more trusting of government, regulatory accountability rests less on legal mechanisms and more on professional training. Edward Rubin reports that the German Bundesbank's regulations for assuring bank safety and soundness are bound in a pamphlet less than one hundred pages in length. The U.S. Federal Reserve Board's operative statutes and regulations fill several thick binders. Officials hired by the Fed to work on bank regulation, aside from having a law degree or an adequate grade on a civil service examination, receive a few weeks of on-the-job training. The Bundesbank runs a three-year “college” for its regulatory recruits.^[12] When regulatory officials are thoroughly trained professionals, dedicated to a career in the same regulatory program, Rubin notes, authorities can trust them to make programmatically sensible judgments and need not bind their discretion with detailed rules. Repeatedly, officials of multinational corporations whom we interviewed commented on frequency of turnover among the regulatory personnel they deal with in American agencies—which in turn led to variability in American regulators' level of technical knowledge when compared with their counterparts in Europe and Japan. When regulatory personnel are “new on the job,” they are more likely to “go by the book,” applying rules legalistically, because they do not have the specialized knowledge that enables them to confidently negotiate with professionals in regulated enterprises or to defend discretionary decisions to mistrustful superiors or politicians. Legalistic behavior—demanding strict adherence to the rules, demanding more tests and certifications—is a common response to that lack of confidence.^[13]

---

C. The Distinctiveness of American Liability Law

The case studies in this volume and other exploratory interviews by project researchers also underscore the uniquely threatening character of the American liability law regime, particularly tort suits and civil cases that are brought pursuant to regulatory statutes. The distinctiveness of American liability law arises from (1) the readiness with which the liability system can be and is mobilized; (2) the unpredictability and potential magnitude of damage awards (due to both the vagueness of American legal rules for damages and the use of juries); and (3) the costs of the litigation process itself.

The multinational corporation studied by Nielsen (chapter 7) prescribes identical personnel management and employee termination policies for its American and its Canadian branches. Yet in one recent year, Nielsen found, almost 23 percent of “forced separations” of employees in the United States resulted in a lawsuit against the company, compared with 7 percent of forced separations in Canada—even though Canadian substantive law protecting employees against arbitrary dismissal is more comprehensive than is American law. Nielsen's chapter thus supports the general research literature, which suggests that whereas (and perhaps because) American law provides weaker statutory protections against dismissal and less generous postdismissal severance and unemployment benefits than the labor law of Western European states and Canada, American employers are far more likely to be sued for substantial money damages in courts of general jurisdiction on grounds of unjust dismissal or discrimination.^[14]

A senior risk manager for a major chemical manufacturer, 55 percent of whose sales are in the United States, told us that for every one hundred product liability claims against his firm in the United States there are about twenty or twenty-five in Western Europe and none in Japan. Claims filed in the United States (which include actions for products marketed and even manufactured abroad) account for 85 percent of the dollars his company expends on product liability claims, the rest of the world, 15 percent.^[15]

Officials of multinational chemical companies headquartered in other countries told us that they feel threatened by the sheer cost of defending cases filed in American courts as much as by the fear of potential damage awards. The uncertainty of the American product liability system also is widely regarded as troublesome. A Japanese chemical company official reported that after consulting with American legal counsel, his company decided not to market in the United States a household “air freshener” designed to neutralize the odor of tobacco smoke—a product that it sells in large volumes in Japan—because of the possibility of liability suits by individuals harmed by tobacco smoke, whose lawyers might employ a difficult-to-anticipate legal theory.

---

A legal counsel for a non-American motor vehicle manufacturer (echoing the accounts of some tort scholars)^[16] told us that no country in which his firm sells vehicles comes close to the United States in terms of the incidence and cost of product liability litigation. Canada is perhaps the closest, he noted, but it remains a very distant second, for several reasons: (1) Canada requires losing plaintiffs to pay some or all of the defendant's lawyer fees, and in Ontario, the most populous province, contingency fees are prohibited; this, he said, “chills the filing of frivolous cases in Canada”—as opposed to the United States, where he claimed that a significant proportion of cases are filed without much investigation, in hopes of getting a quick “nuisance settlement.” (2) In civil cases, most provinces in Canada (like almost all other countries) do not use juries. In the United States, jury trials in auto product liability cases typically involve complex and conflicting expert witness testimony on such issues as “accident reconstruction, body mechanics, and motor vehicle design engineering and decisionmaking,” and juries, in this corporate lawyer's view, often are swayed by clever lawyering. (3) Canada (like most other countries) has much more specific damage rules and does not impose punitive damages (which even in the United States are actually imposed in only a small percentage of claims but are important, this lawyer claimed, because 10 percent of the liability cases generate 90 percent of the firm's financial exposure). (4) Canadian law provides that if the injured plaintiff was contributorily negligent as a result of alcohol or drug abuse, no recovery is permitted. (5) Canada (like other countries) does not have the highly organized litigation machine that American entrepreneurial lawyers have created—in which plaintiffs' lawyers draw on (and help fund) databases, “litigation kits,” workshops on particular vehicles, and expert witnesses provided by the American Trial Lawyers Association and the Center for Automotive Safety.

IV. EXCEPTIONS TO THE RULE: ADVERSARIAL LEGALISM ABROAD, SUBDUED ADVERSARIAL LEGALISM AT HOME

If the case studies as a whole confirm both the prevalence and the distinctiveness of American adversarial legalism, some of the case studies and exploratory interviews provide three kinds of counterexamples: (1) instances in which a multinational encountered equal levels or more adversarial legalism in other countries, (2) instances in which there were low levels of adversarial legalism in the United States (and abroad), and (3) regional variation within the United States in the amount or intensity of adversarial legalism.

---

A. Adversarial Legalism Abroad?

Somaya's account of ACME chemical company's efforts to patent its innovations in Japan, Europe, and the United States (chapter 9) provides a muted counterexample of the first kind. ACME experienced a great deal of costly, time-consuming, and legally unpredictable legal contestation in all three places. American procedures for granting a patent, in fact, are less adversarial. In contrast to the Japanese and European patent offices, the U.S. Patent Office does not provide for pre-grant “opposition” proceedings by competing companies. In ACME's case, oppositions in Japan and the EU were contentious and costly, and resulted in appeals to court.

On the other hand, Somaya's chapter does not suggest that Americanstyle adversarial legalism is spreading abroad, for the American patent system structures legal contestation of patents quite differently, using a more distinctly American mode of adjudication. Instead of centralizing adversarial opposition to patents in a pre-grant or immediately post-grant administrative procedure, as in Japan and the EU, inventors or companies that wish to challenge a U.S. patent usually do so via post-grant litigation (often as defenses to the patent holders' lawsuits for infringement). Thus the American system pushes adversarial disputes over patents out of a specialized administrative tribunal into a decentralized, somewhat unpredictable court system, in which juries play a significant role. Since an alleged infringer can challenge the patent in court at any time during its term, the legal uncertainty over a patent's validity can linger much longer in the United States. And the courts themselves provide relatively little certainty. Twice, when ACME sued another company for infringement, a U.S. District Court invalidated the company's patent. Twice, the company appealed to the Court of Appeals, which reversed the District Court and sustained the patent. In addition to this characteristic legal uncertainty, Somaya reports that American legal processes for enforcing a patent against infringers are much more adversarial and costly than such actions in other countries. American patent litigation is unique in the extent of its party-driven pretrial discovery, in its reliance on privately hired expert witnesses, and in its threat of imposing treble damages.^[17]

B. Centralization of Authority and Subdued Adversarial Legalism

More interesting, perhaps, are the spheres of law in which adversarial legalism did not appear to be a salient aspect of the multinational corporations' experience in the United States. The common ingredient in these low-adversarialism policy areas is that legal authority is centralized and difficult to challenge in court, either legally or practically; that is, governmental legal authority is exercised more hierarchically and with more finality

---

than is the American norm (and hence more like most parliamentary democracies).

For example, shipping officials in a multinational petrochemical company described as generally nonadversarial and nonlegalistic their relationship with the tank car safety section of the Federal Railroad Administration (FRA), which regulates the design of tank cars and inspects existing cars for safety. In this policy arena, company officials interact repeatedly with a single federal agency, rather than with a multiplicity of federal, state, and local agencies. It is an area of law that has not attracted many lawsuits (against the FRA or the company) by advocacy organizations. In addition, because the FRA has summary powers to suspend the use of equipment it finds dangerous, the company is not inclined to mount legal resistance even to citations or orders it considers unreasonable; the company's primary interest is to keep the trains rolling, even if the FRA requirements are costly to comply with.

Similarly, the regulatory affairs official in the biotech company studied by Kraus (chapter 10) described the biological products office of the U.S. FDA as somewhat legalistic (in the sense that it is a stickler for strict compliance with highly detailed rules) but not adversarial. Disputes over company applications for an FDA facilities license or new product license rarely end up in court. Here, too, regulatory power is concentrated in a single regulatory office in Washington, and that office that has crucial powers: it can delay the marketing of products in which the company has invested huge amounts of money, and it can publicly question the safety practices of a health care products company, which has an enormously important economic stake in a reputation for absolute safety. In these circumstances, the regulated entity has strong incentives not to appeal even those agency decisions that it considers unreasonable. It is for similar reasons that banks tend not to litigate against the Federal Reserve Board.^[18] Such agencies have strong hierarchical powers; that is, they rarely are subject to challenge and reversal in court, or by other agencies. And to refer to the typology of decision-making methods set forth in chapter 1, hierarchical authority structures push policy implementation or dispute resolution into the cell labeled “expert or political judgment” or “bureaucratic legalism,” rather than adversarial legalism.

Consider, too, the report by Johnson, Fujie, and Aalders that numerous multinational chemical company officials view the U.S. TSCA regime, as administered by the EPA, as less legalistic and more economically efficient in screening new chemical substances than comparable regulatory systems in Japan and the European Union. This surprising finding stems partly from the distinctive political climate in which TSCA was enacted, which resulted in a statute that focuses analysis only on high-risk new chemicals,

---

so as to avoid regulatory delays that would slow the movement of wellunderstood, beneficial new chemicals into the marketplace. But it is also noteworthy that TSCA concentrated decision-making authority for new chemicals in one agency in Washington, D.C. Chemical company executives regard the EPA officials who administer TSCA as having a reasonably high level of scientific expertise. Chemical companies rarely appeal to court the EPA's decisions pursuant to TSCA (for that would only prolong delays in getting products to the market); nor do pro-regulation advocacy groups. Hence, the TSCA process is dominated not by lawyers but by scientific experts.

The TSCA exception must be tempered, on the other hand, by Johnson Fujie, and Aalders's report that in the experience of multinational chemical companies, the EPA is far more likely than its Japanese and European counterparts to take formal legal action, including the imposition of substantial fines, if a company is found to have violated any of the highly technical filing and testing regulations. This reflects the general tendency of American agencies to respond to detected violations with legal sanctions, in order to emphasize their authority and protect themselves from political criticism.

C. State-by-State Variation in the United States

In their account of the permitting process for municipal waste disposal facilities in California, Pennsylvania, Great Britain, and the Netherlands, Welles and Engel (chapter 5) highlight the prevalence of state-level variation in adversarial legalism within the United States. Whereas Waste Corp. encountered extremely costly litigation in California, it did not encounter lawsuits in Pennsylvania. The state of Pennsylvania has constructed a much more hierarchical decision-making process—and, in that respect, one that more closely resembles the British and Dutch systems. In California the project proponent had to obtain separate permits, certifications of compliance, or agreements from sixteen local, regional, and state agencies, each imposing different legal and analytical requirements. In Pennsylvania a single state-level environmental agency issued the facility's solid waste permit and coordinated the issuance of permits (a total of six) required from other state agencies. In Pennsylvania, appeals by local opponents and environmental groups were channeled into a state-level administrative hearing process. But in California, local opponents (with help from a rival waste disposal company) hauled Waste Corp. into court more than once, challenging both the environmental impact report and permit conditions that had been approved by the local agency.^[19] Thus the Pennsylvania case, too, suggests that adversarial legalism is less common when regulatory authority is more centralized or hierarchically organized.

---

Dwyer, Brooks, and Marco (chapter 6) describe two kinds of legal conflict that Ford Motor Company encountered in seeking to expand or change production patterns in two of its automobile assembly plant paint shops, one in Minnesota and one in New Jersey. Referring to the air pollution permitting process at its Minnesota plant, Ford officials describe the Minnesota Pollution Control Agency (MPCA) as cooperative, constructive, and nonlegalistic. Also, when Ford sought a formal “innovative technology waiver” from the U.S. EPA “new source performance standards,” based on its plan to develop new low-volatile-organic-compound (low-VOC) paints, the federal agency announced its intention to grant the exception. Legal conflict arose as a result of opposition from local community groups and environmental advocates who had long been unhappy about odors emanating from Ford's plant. These local opponents challenged and then appealed the EPA waiver, politically pressured MPCA to ask Ford to undertake an odor abatement study, and requested a contested hearing before an independent administrative law judge with respect to the MPCAissued permit—all of which resulted in the creation of a citizens task force that ultimately worked out a cooperative solution. In New Jersey, it also took several years for Ford to obtain a permit. There the conflict arose not because of adversarial opposition from neighbors and environmentalists but because of resistance by officials in the New Jersey Department of Environmental Protection (NJDEP), whom Ford officials characterized as legalistic, distant, and noncooperative—although NJDEP officials blame Ford for failing to provide adequate information to evaluate the permit and predict the consequences for air quality. Conceivably, the difference between the MPCA and the NJDEP regulatory behavior reflects the fact that New Jersey, unlike Minnesota, is in a “nonattainment area” under the Clean Air Act, which puts regulatory officials under some pressure to bring about measurable reductions in pollutants even in expanding factories.

V. THE COSTS OF ADVERSARIAL LEGALISM

Notwithstanding the exceptions mentioned, the dominant message of the case studies in this book, along with project researchers' exploratory interviews with a wider number of business executives, lawyers, and consultants, is that adversarial legalism is predominantly an American phenomenon and that it is rather pervasive across business sectors and legal policy areas. Moreover, multinational enterprises generally regarded the “friction costs” associated with American adversarial legalism as both substantial and troublesome. The precise magnitude of those “extra” legal-procedural costs is difficult to specify. The financial statements and accounting schemes of the corporations studied were not designed to identify those costs, which tend to be buried in a variety of other categories. Nevertheless,

---

qualitative information about the friction costs, especially when viewed in the light shed by the explicit cross-national comparisons, provides some useful insight into the consequences of adversarial legalism. The following subsections organize this qualitative information in terms of various kinds of costs.

A. Legal Services

All modern societies need lawyers to help communicate the requirements of complex and changing legal rules, to structure contracts, and to defend the accused. But the United States seems to spend much more on them than comparable countries. In 1991 aggregate expenditures on legal services in the United States amounted to more than $100 billion, perhaps 2 percent of the gross national product.^[20] Ordinary observation leaves little doubt that when American business executives are engaged in negotiating sales franchises, seeking approval for real estate projects, issuing stock, and launching new products, they are surrounded by larger phalanxes of expensive attorneys than their counterparts in Europe or Japan—where the legal risks corporations face are less problematic.

In this project's interviews of officials in multinational enterprises, respondents often mentioned that their company spends far more money on legal services in its American operations than in its parallel operations in other countries. American subsidiaries consult lawyers more often and longer on a wider range of matters, ranging from selecting pollution control equipment to managing problem personnel and conducting sales transactions. They do so, project researchers repeatedly were told, because (1) American law is generally more complex, changeable, and difficult to master, and (2) the legal sanctions for being wrong are generally much higher.^[21]

The U.S. subsidiary of PREMCO, the Japan-based manufacturing company studied by Aoki and Cioffi (chapter 2), has spent more money on lawyering—both in-house and outside counsel—than PREMCO has in all its much larger worldwide operations. Nielsen (chapter 7) found that when the pharmaceutical company she studied decided to terminate individual employees in the United States, it consulted its attorneys earlier, more frequently, and in more depth than when terminating employees in Canada. Ruhlin (chapter 8) observes that because German law concerning debt collection is much simpler and more uniform than American law, the German credit card division of Credit Co., a multinational bank, contrasts sharply with Credit Co.'s American operations: the German division does not have to maintain separate in-house counsel's offices to deal with litigation management, consumer bankruptcy, and debtors' counterclaims.

---

And unlike its U.S. counterpart, the German division does not feel obliged to provide ongoing, intensive legal training for collection agents.

Similarly, in one of the project's exploratory interviews, a corporate general counsel in the U.S. office of a multinational chemical firm said that in the company's U.S. subsidiary—in contrast to its European corporate parent—a wide variety of documents are routinely reviewed by the legal department. One stimulus to this procedure was a lawsuit for breach of contract in which a drawing made by a company sales representative was held, much to the company's surprise, to have been evidence of a contract. The suit cost the company roughly $500,000 in damages and $500,000 in disputing and lawyering costs. Now, the lawyer said, “Half of our salesmen are scared to death to do anything without consulting a lawyer.” The U.S. subsidiary employs six specially trained nonlawyers who spend at least half their time fielding legal questions that salespeople routinely ask about sales contracts. To further avoid legal problems, the corporate attorney added, “When European people [from the parent firm] come over here, we have to forbid them from talking or writing letters to anyone.”

Most shockingly, Welles and Engel found that Waste Corp. (chapter 5) spent a staggering $15 million on legal services in the course of its efforts to obtain approval for a municipal solid waste landfill in California; for over ten years, the company had approximately seven lawyers on retainer, busy addressing, inter alia, two major administrative appeals and three extended lawsuits. In Pennsylvania, the same company retained seven lawyers (but only part-time) for the five years it took to get a permit there, a process that entailed two administrative appeals but no lawsuits and “only” $1.45 million in lawyering costs. In England, by contrast, the company retained two lawyers, part-time, for an eight-year process that also included at least one administrative appeal; its legal costs there were about $137,000. And in the Netherlands, despite having undergone two administrative appeals, the company did not have to retain lawyers at all (since lawyers are not required in administrative appeals) and spent “less than “50,000” on legal services.

B. Accountability Costs

Viewed in cross-national perspective, American regulatory regimes generally impose more extensive and specific requirements concerning reporting, record keeping, testing, employee education, certifications, and so on.^[22] In addition to the costs of complying with substantive regulatory standards, therefore, regulated firms in the United States usually must spend more than in their overseas operations in proving that they are complying. Consider, for example, Dwyer, Brooks, and Marco's comparison between

---

the air pollution permits that German and U.S. regulators issued to Ford Motor Company, when Ford sought to expand production or use new paints on its vehicles (chapter 6). In both countries the regulatory authorities required Ford to install similar VOC-reducing technologies, such as carbon adsorption units. But the American permits, the authors observe, impose much more detailed record-keeping and reporting requirements, designed to ensure that Ford does not exceed the planned level of emissions. Ford's New Jersey permit, these authors tell us, tell us, set hourly, daily, and annual VOC emission limits for each separate step of the painting process; included limits on the amount of volatile organic solvent (VOS) for each of the several coatings; specified the type of fuel and fuel usage rates for afterburners; and specified “destruction efficiency” and operating temperature for the afterburners. The permit also required Ford to install a continuous VOC monitor and recorder after each carbon adsorber; and to seek agency approval for type and calibration of the monitors and recorders; to conduct three types of tests on the efficacy of the controls; to test the solvent content of the paint monthly, and get prior agency approval for its testing methods; to maintain all records—including daily production rates, colors, paint and solvent usage, VOC content of the paint, natural gas consumption (for the thermal oxidizers), and monitoring data—onsite for at least five years, to be made available to regulatory officials on demand; and to submit a quarterly report to the agency covering daily use of coatings and solvents for each stage of the painting processes and daily emission levels.

The German permit for Ford's Cologne plant was also detailed, but basically it set an overall annual limit for VOC emissions, rather than specific step-by-step limits. It required Ford to engage in continuous monitoring of certain parameters, to conduct certain tests, and to keep records of solvent use. But with respect to the number and specificity of the testing, monitoring, and reporting procedures, the German permit was far less demanding—one might say it “displayed much less mistrust”—than the New Jersey permit. For example, after initial tests to be sure the system was operating, the German permit required Ford to undertake performance testing only every three years.

Another telling comparison of regulatory accountability regimes emerged from the visit of two project researchers (Kagan and Marco) to a U.S. plastic products plant in Texas. The plant is closely modeled on its corporate owner's Japanese factory for making the same products. Japanese engineers working at the Texas plant stated that the relevant air pollution control standards in the two nations are roughly the same, as are the basic pollution control systems in the two installations. Thus the Texas and Japanese factories use similar methods to control fugitive emissions from the complex system of pipes and ducts. But in the United States, firms are

---

required to keep more detailed records of emissions and to file much lengthier reports to the state agency that enforces federal regulatory standards. In the Texas factory, for example, there are at least four thousand points in the manufacturing process at which emissions can leak out. Each one must be tagged, monitored, and recorded with time and date, and total emissions must be estimated on a quarterly basis. The Texas plant hires outside consultants to perform this time-consuming task. In Japan, we were told by corporate officials, government audits are less frequent, and fines for detected noncompliance are issued only in instances of continued noncompliance. Once a company has acquired a good reputation for compliance, as this firm's Japanese plant has, it need not invest in the extra layer of monitoring required in the Texas plant. The Japanese engineers did not think fugitive emissions in their Japanese plant were any worse than those in the Texas plant.

One consequence of the more demanding regulatory accountability regimes in the United States, several studies in this book indicate, is that American branches of multinational corporations maintain substantially larger intracorporate regulatory compliance staffs than do similar business operations abroad. Thus Nielsen (chapter 7) shows that in addition to its “human resources department,” the American division of a pharmaceutical company created a special Affirmative Action Office—an additional layer of bureaucracy that must be consulted in all probation and dismissal cases involving female employees, racial minorities, and employees over forty years of age. The company does not have such an office in its branch in Canada, where litigation and regulatory agency intervention are less frequent and less threatening. Moreover, the character of corporate compliance staffs often differs. In Aoki and Cioffi's study of waste regulation (chapter 2) and in Aoki, Axelrad, and Kagan's study of water pollution control (chapter 3), a multinational manufacturer's U.S. subsidiary confronted a more legally complex and potentially punitive regulatory regime than its sister operations in Japan. In the U.S. subsidiaries, regulatory affairs departments concentrated much more on acquiring legal expertise, whereas regulatory compliance specialists in the Japanese factories concentrated much more fully on technical or engineering issues. Similarly, the more threatening and costly American litigation system often induces companies in the United States to spend more on what might be called “defensive legal medicine.”^[23]

1. WHEN ACCOUNTABILITY DISPLACES RESPONSIBILITY

The cross-national differences in accountability costs are symbolized most strikingly by the huge volume of supporting evidence companies must supply to regulators in the United States to demonstrate that the firm has met legal standards. Axelrad (chapter 4) recounts the experience of B

---

Corporation when it notified regulatory authorities in the United States, England, and the Netherlands that it had discovered solvents which had leaked from deteriorating underground tanks and pipes. In each jurisdiction, B Corporation embarked on discussions with regulators concerning further soil and groundwater testing and remediation. The American regulators, however, demanded far more comprehensive analysis, more voluminous documentation, and more costly reports. B Corporation's corporate regulatory compliance group officials said that documents submitted to American regulators for such contaminated sites typically fill a fourdrawer filing cabinet, compared with one foot of depth in a single file drawer for the documents submitted to the other countries. And behind each additional ten pages of documentation, they emphasized, lie scores of hours which company officers must devote to research, testing, measurement, analysis, and preparation and checking of draft reports. All in all, B Corporation officials estimated that “extra” studies, submissions, and negotiations with U.S. regulators added $8 to $10 million to the costs of designing the cleanup plan for the two sites in the United States (out of total costs per site of an estimated $22 million), whereas the “extra” regulatory accountability costs for comparable site investigations and cleanup planning in the United Kingdom and the Netherlands were negligible.^[24] As of the time of Axelrad's interviews, remediation efforts in England and the Netherlands were well under way. But in the American sites, action remained on hold while the firm waited to learn if officials considered the company's analysis sufficient. In this case, therefore, the additional demands of the U.S. regulatory regime confirmed the maxim that when pushed too far, accountability (proving one has done the right thing) can displace responsibility (doing the right thing).^[25]

2. TWO EXCEPTIONS TO THE PATTERN

Two case studies highlight policy spheres in which regulatory accountability requirements in the United States appear to be less demanding than those imposed by agencies in other countries. Johnson, Fujie, and Aalders (chapter 11) point out that the American regime for premarketing clearance of new chemicals (under TSCA) does not require extensive toxicity testing for all new products, but only for 25 percent or so—those for which the EPA, based on the company's submissions, concludes that the potential risks justify the costs and delays of testing. European Union regulations, in contrast, require extensive testing for all new chemical products produced in volume, regardless of resemblance to previously tested chemicals and regardless of levels of risk. In Japan, too, an innovative company must establish biodegradability characteristics for all new chemicals, and toxicity levels for most; hence obtaining approval for marketing new products in Japan is much more costly than in the United States. Kraus (chapter 10)

---

shows that a multinational manufacturer of genetically engineered biologics found that EU demands for information concerning the company's manufacturing methods were more analytically searching than the U.S. FDA's regulatory requirements. The EU regulators compelled the company to undertake additional testing and quality assurance analyses for a production process already approved by the FDA.

In these two regulatory arenas, American regulatory authority is concentrated in a single federal bureau in Washington, D.C. (rather than, as is often the case, scattered in an overlapping welter of federal, state, and local agencies), and those regulators have considerable technical expertise. It is noteworthy, too, that in these two spheres of regulated activity—genetically engineered health products and production of new chemical substances— the United States seems to have a competitive advantage in launching useful new products.

C. Opportunity Costs

One characteristic regulatory tool, prominent in a number of the regulatory programs described in this book, is prior governmental review and approval of new construction projects, new health care products, and changes in industrial production, all with the purpose of ensuring compliance with safety and environmental standards. Such permitting systems are consciously designed to slow the headlong rush of development and technological change, to force consideration of their potential adverse side effects. This prior regulatory scrutiny, however, may also impose opportunity costs on society, as well as on the businesses that suffer regulatory delays. The community must wait longer for improved goods and services and for the jobs that accompany the new ventures. Kraus (chapter 10) tells us that for every month that an FDA manufacturing process review kept a new genetically engineered blood product off the market, the manufacturer had to forgo serving a market that may be worth more than $13 million in monthly sales, and individuals who were eager for the new genetically engineered (and probably safer) version of the product had to wait longer to obtain it.

The regulatory review systems of other nations sometimes impose even greater delays—and higher opportunity costs—than American procedures. A U.S. General Accounting Office study found that it took six to seven years, on average, to win approval for a patent in Japan, compared with about nineteen months in the United States.^[26] When “Company Y” (chapter 10) sought regulatory approval of its facility for manufacturing a genetically engineered biological product, it took more than two years to obtain a license from the European Union's Committee for Proprietary Medicinal Products plus the relevant agencies in certain member states, almost as long

---

as the FDA process in the United States. Johnson, Fujie, and Aalders (chapter 11) found that premarketing clearance procedures for new chemicals typically took longer in the EU and Japan than in the United States. TSCA provides a uniquely rapid, no-testing “track”—unless EPA officials decide (which they must do within a ninety-day period) that the chemical requires more in-depth analysis. Also, the United States provides a faster clearance system for new chemicals produced in small volumes for experimental or “market-testing” purposes, which apparently makes it advantageous to market new chemicals first in the United States.

Nevertheless, this project's research as a whole suggests that American legal and regulatory regimes more often impose longer delays and larger opportunity costs than comparable regimes abroad. In contrast to the FDA and TSCA regimes, decision-making authority in the United States generally is more institutionally fragmented than in parallel regulatory processes abroad, requiring more applications and administrative reviews, and hence more possibility of costly delay. Moreover, American opponents of new projects generally have more opportunities to challenge regulatory approvals in court, which can result in substantial opportunity costs. The greater prospect of judicial review, moreover, often seems to make American regulatory officials more cautious and legalistic in reviewing proposals than their counterparts abroad. According to an environmental consultant with a great deal of cross-national experience, because the regulatory permitting process for an industrial project in the United States entails a great deal more legal formality, organizational complexity, and documentation than in Western Europe, “It takes less time overseas. The cost for initial studies is less. We are not required to accumulate as much information. In Europe [the time from application to permit averages] one third less.”^[27]

Consider, for example, Dwyer, Brooks, and Marco's comparison of the American and German permit systems for ensuring that changes in motor vehicle factory production and painting processes do not result in increased air pollution or obnoxious odors. Regulatory officials in both countries required Ford Motor Company to install similar pollution control technologies and to build higher exhaust stacks to diffuse odors. But for two factories in western Germany, the time from permit application to approval was five months and seventeen months, respectively; for Ford's plants in Minnesota and New Jersey alike, it was over four years. The opportunity costs for Ford's American operations were tempered because the state agencies allowed Ford to continue production during the permitting process, albeit at a reduced volume. But the way the American permit process operated, that chapter indicates, suggests that its inherent potential for imposing costly delays is much greater than the German process. In fact, American automakers have complained that the greater prescriptiveness of American air pollution rules concerning paint shop emissions often

---

leads to more regulatory inflexibility and delay than they encounter abroad, sometimes throwing tightly coordinated major model changes off schedule. In another study, Richard Brooks describes how a vehicle manufacturer, frustrated by regulatory delays and resistance in seeking to modify its air pollution permit, ended up contracting out the painting of small truck bodies to a distant firm—and shipping them hundreds of miles by rail back to the assembly plant.^[28]

Axelrad (chapter 4) shows that in seeking regulatory approval of methods for investigating contamination and remediation of industrial sites, it took B Corporation scarcely more than a year in Great Britain. In the Netherlands, the regulatory process took slightly over five years. But in the United States (at two different sites) it took fifteen and eleven years, respectively. As noted earlier, this involved substantial opportunity costs: the actual remediation efforts moved forward much more rapidly in the United Kingdom and the Netherlands, while that crucial step remained delayed by regulatory red tape in the United States.

Similarly, as noted earlier, Welles and Engel (chapter 5) report that it took Waste Corp. eleven years to complete the permitting process for a municipal waste disposal site in California, which delayed the opening of the landfill for five years. The process for comparable facilities in the United Kingdom took eight years, despite similarly intense local objections. In Pennsylvania and the Netherlands the permit process took about five years, including the processing of local objections and administrative appeals. A company with lesser financial resources and staying power than Waste Corp. might have given up, especially in California—which is precisely what project opponents (and competitors) often hope to achieve by initiating litigation. But although the opponents of Waste Corp.'s California project did not succeed in stopping the project, by constructing consecutive legal obstacles they did compel a sharp reduction in the size of the planned landfill.

D. Divisiveness and Delegitimation

One of the more intangible costs of adversarial legalism is its corrosive effect on personal and institutional relationships. When a regulatory inspector and a regulated enterprise become locked in a legalistic, adversarial posture, the exchange of information and cooperation that are essential to effective regulation often are reduced.^[29] When regulatory rule making is only a prelude to litigation, a National Academy of Sciences analysis of EPA policy making observed, contending interests are more likely to resist compromise.^[30] In an excellent comparative study, John Braithwaite found that American regulations concerning quality of care in nursing homes are enforced more legalistically (complete with monetary fines and the threat

---

of court orders) than the Australian regulatory regime. Nevertheless, Braithwaite observed, quality of care in American facilities is on average worse. While there may be multiple reasons for that disparity, Braithwaite argues that legalistic enforcement has played a significant role, encouraging the development of large corporate-run homes (that can better afford to comply with costly regulations concerning facilities) and to a ritualistic, penalty-avoiding attitude toward compliance, rather than to a more caring attitude.^[31] As Steven Kelman puts it, “Adversary institutions fail to create any relationship among the parties to a policy disagreement. Participants in adversary institutions remain separate individuals, physically proximate only in order to argue before a third party. They do not talk to each other.”^[32]

Some of these effects were evident in the case studies in this book and were mentioned in the exploratory interviews that led to them. Environmental compliance officials in B Corporation, a multinational chemical manufacturer, contrasted the quality of the company's communications with British and American regulators. In British manufacturing plants, they told Axelrad, regulatory restrictions concerning control of industrial water pollution, air pollution, and other wastes are combined in a single “unified permit.” British regulators, in consequence, can trade off special efforts by the firm or more costly restrictions in one area for greater leeway in another area. B Corporation officials plausibly argued that this give-and-take led to more efficient overall remediation plans than in the fragmented American regulatory system, in which the company typically must deal with a multiplicity of state and federal regulatory offices, each operating under its own specific and highly detailed statutory instructions.^[33]

Nielsen (chapter 7) describes a pharmaceutical company that has institutionalized a probation process to deal with allegedly problematic employees and, if dismissal is decided upon, procedures for negotiating a termination settlement. Nevertheless, as noted earlier, Nielsen found that 23 percent of employees dismissed by the company's American operations in a typical year subsequently brought a lawsuit against the company— compared with 7 percent in Canada. The disparity arises, it seems likely, from features of the American legal system that provide stronger incentives to sue rather than to negotiate—much higher money damages for unjust or discriminatory dismissals, contingency fee arrangements for compensating lawyers, and the law's failure to compel losing plaintiffs to pay winning defendants' counsel fees.

California's complex permitting scheme for municipal waste disposal, as described by Welles and Engel (chapter 5), seems to discourage initial compromise by project opponents because they gain a great deal of bargaining power by initiating litigation, which can impose costly delays on

---

the community and the waste disposal company. To settle the third of the costly lawsuits designed to block its project, Waste Corp. agreed to pay the objecting city government $3.5 million for public works unrelated to environmental risks associated with the landfill—such as the remediation of environmental problems elsewhere in the city that had been caused by entirely different enterprises. In contrast, Welles and Engel's chapter suggests, legal objections to a proposed landfill in England and the Netherlands are directed to a single administrative body or administrative court (rather than into courts of general jurisdiction), and higher-level governmental bodies can override a local government's objections. Hence, local opponents have more incentive to engage in dialogue and to reach a compromise. In California, the leverage provided by adversarial legalism gave objectors incentives to hold firm.

Aoki and Cioffi (chapter 2) and Aoki, Axelrad, and Kagan (chapter 3) show how American adversarial legalism can erode regulated entities' respect for the law and incentives to internalize regulatory norms. The American subsidiary of PREMCO, a large metal products manufacturer, was battered by heavy fines for violations of RCRA waste product collection and storage regulations. One consequence was that the subsidiary made significant changes in environmental management, policies, and practices. Nevertheless, both environmental and line managers continued to view a visit from regulatory officials with trepidation. They continued to resent the detailed, complex character of the RCRA rules. In PREMCO's parallel Japanese factory, no citation or fine for waste disposal rules has ever been imposed. Japanese law gives the company more discretion in solving the collection and disposal problem. There, regulators' visits are viewed as neither annoying nor threatening. Compared with their counterparts in PREMCO-USA, shop-floor workers in PREMCO-JAPAN seem to have more fully internalized waste control norms—which are perceived as a matter of company policy, not governmental regulation. Similarly, among the subsidiaries of Q Corp, a large manufacturer of electronic components, only Q USA has dragged its feet in developing an environmental management and auditing plan that would pass muster with an international standards body. The American officials have all they can handle, they say, simply trying to keep abreast of and in compliance with the more detailed, complex, and potentially punitive American regulatory regime.

Thus, by making legal and regulatory processes costly, complex, frustrating, and threatening, adversarial legalism tends to alienate citizens and business executives from the law and from the state, eroding respect for bureaucrats, courts, and active government. A national business and professional elite that trades jokes which demean regulatory bureaucrats and lawyers, I suspect, also is more likely to support reforms that diminish

---

regulatory obligations and rights to litigate—without providing alternative mechanisms to ensure accountability, sensible regulatory protections, and justice itself.

VI. COMPENSATING SOCIAL BENEFITS?

The costs that flow from American adversarial legalism —“defensive medicine,” larger expenditures on lawyers and consultants and paperwork, opportunity costs, the defensiveness it often stimulates—might be justifiable if adversarial legalism, by virtue of the legal threats it generates, also stimulated higher levels of legal compliance and correspondingly greater social benefits. Lyle Scruggs's study of comparative environmental performance in rich democracies (measured as rates of reduction of major sources of pollution) suggests the opposite. Using OECD data, Scruggs found that notwithstanding (or perhaps because of) its more prescriptive, fiercer environmental laws and enforcement style, the United States made less progress in the 1970–90 period than Germany, the Netherlands, Japan, Sweden, and the United Kingdom, where regulatory policy is forged and implemented in a more consensual manner.^[34] And although it cannot be claimed that the case studies in this book are fully representative, the representatives of multinational corporations studied generally could not point to significant public-regarding precautions taken in their American operations that were not also taken in their installations abroad. More specifically:

• PREMCO, a metals manufacturing company (Aoki and Cioffi; chapter 2), encountered much more legalistic enforcement of industrial waste disposal regulations in its U.S. plant (replete with large financial penalties and heavy legal expenses) than in comparable factories in Japan. In consequence, the American operations beefed up their environmental management system. Yet officials with experience in both countries assert that waste collection and disposal practices in the Japanese plant are identical, and are implemented by the company with much more care.
• Aoki, Axelrad, and Kagan (chapter 3) found that in “Q Corp's” Japanese electronic parts factory, control of industrial effluent has been more consistently effective than in the corporation's similar U.S. factory, notwithstanding the fact that statutory standards in the United States are more stringent and prescriptive and that “Q USA” (unlike “Q Japan”) has been subjected to costly regulatory penalties.
• Nielsen (chapter 7) indicates that the American legal system encourages dismissed employees to sue their former bosses more often than their counterparts in Canada. But the American branch of the company Nielsen

  ---

  ― 401 ―
  studied does not respond by being more solicitous of workers' claims than the Canadian branch; if anything, the American branch is more aggressive in firing employees who are initially tagged as unsatisfactory, for it uses the company-prescribed probation process less often than the Canadian branch.
• The multinational company discussed by Kraus (chapter 10) makes genetically engineered biological products; its officials are quite cognizant of the enormous liability risks that would arise in the United States from the sale of contaminated product, and they design their quality control system accordingly. Yet when the company sought manufacturing licenses, the European Union's regulatory body insisted on more probing analyses of quality control than did the U.S. FDA. The EU's standards led the company to establish certain controls that had not been stimulated by the more legalistic American legal and regulatory system.
• Axelrad (chapter 4) finds that U.S. regulations concerning the study and remediation of contaminated industrial sites are significantly more detailed and prescriptive than Dutch and British regulations. A multinational chemical company that dealt with all three jurisdictions described both federal and state regulators in the United States as much more demanding, in terms of required investigatory data. But the “extra” research and reviews required in the United States, company officials convincingly assert, did not result in any additional protections for the public because much of that work was demanded as a matter of regulatory routine and produced information that was irrelevant to the risks posed by the contamination at the particular sites. Indeed, as noted earlier, American regulatory legalism delayed the actual cleanup efforts enormously, while in England and the Netherlands the company moved forward much more quickly to contain any remaining risks to human health.

In three other case studies, it is more uncertain whether American adversarial legalism led to “extra” public protections that might be thought to outweigh its higher costs:

• Credit Co., a multinational bank with credit card operations in the United States and Germany, encounters more adversarial legalism in seeking to collect unpaid debts in the United States (Ruhlin, chapter 8). Because of differences in the law and in the efficiency of the court system, American debtors are much more likely than German debtors to raise legal defenses based on the complex regulations governing direct collection efforts, and much more likely to use the bankruptcy law to block collection efforts. Credit Co., accordingly, “writes off” a larger proportion of delinquent consumer debt in the United States; this might be considered a social cost, since the bank's losses are presumably passed

  ---

  ― 402 ―
  on to other customers in one way or another. But the American system arguably provides certain benefits as well—greater legal protection from overly aggressive or manipulative collection efforts; the “fresh start” that bankruptcy provides debtors who encountered sudden, crushing financial reversals (as a result of serious illness, for example, or divorce or job loss). At the same time, one could argue that Germany's more generous unemployment and public health care systems do a more consistent job of providing for victims of economic reversals of fortune, and that it is precisely the greater difficulty and expense of collecting debts through the courts that lead credit card companies in the United States to resort to extended and aggressive dunning methods.
• Dwyer, Brooks, and Marco (chapter 6) show that in seeking permits from air pollution control authorities, Ford Motor Company encountered more adversarial legal conflict in Minnesota and legalistic resistance in New Jersey than it experienced in its applications for two sites in Germany. It took far longer to obtain the permits in the United States than in Germany. Yet Ford was obliged to install similar pollution control and odor reduction technologies in its assembly plant paint shops in Germany and the United States. However, Dwyer, Brooks, and Marco indicate that the controls on the U.S. plants are tighter, that the permits more demanding, and that Ford may have done more to incinerate odors (as opposed to merely diffusing them) in its Minnesota plant than in Cologne. Although one could argue that those “extra” benefits provided by the U.S. regulations do not outweigh the far greater “friction costs” in the United States, others might disagree. Moreover, it appears that the adversarial proceedings in Minnesota seem to have been the primary precipitant of the odor control plan eventually adopted.
• Welles and Engel (chapter 5) found that in seeking permission to construct solid waste disposal facilities in the United States, the United Kingdom, and the Netherlands, Waste Corp. encountered in all jurisdictions a complicated set of regulatory controls and procedures that provided an opening for public and local concerns or opposition. In each jurisdiction, the regulatory process was effective in that it compelled Waste Corp. to modify its original plans to build in more extensive protections against pollution, odors, traffic, monitoring, and the like. In California and Pennsylvania, however, the permitting process required the corporation to expend vastly more money on lawyers and studies than in England or Holland; and in California, lawsuits and other adversarial challenges resulted in extra multiyear delays in obtaining a final permit. Did adversarial legalism produce protections or benefits for the public in the United States, over and above those provided by the British and Dutch regulatory systems?

Welles and Engel note that Waste Corp. was required to install a

---

― 403 ―

thicker, double lining at its California and Pennsylvania sites than was required in the United Kingdom and the Netherlands, and they are uncertain whether the Netherlands site includes a comparable leachate collection system. On the other hand, the designs of the liners and the leachate systems in the United States are products of the detailed, nationwide legal standards in the federal RCRA law; they were not the products of the local regulatory process described in the case studies.

However, in order to curtail the delays of adversarial legalism in California, Waste Corp. agreed to extensive “exactions.” It “donated” lands (well removed from the waste site) for parks and recreation. It paid for revegetation research, offside tree planting, and cleanup of hazardous waste facilities (totally unrelated to Waste Corp.'s wastes) in other parts of the county. It also agreed to pay for governmental inspection vehicles, county staff members, and “road access fees.” Together, Waste Corp. estimated, these exactions cost $10 million—compared with the roughly $3 million in exactions in connection with obtaining a permit in Pennsylvania, just under $2 million in the United Kingdom, and some noncostly modifications of the waste facility itself in the Netherlands. The much more costly and extensive amenities and “side payments” Waste Corp. provided in California might be considered “extra” social benefits that were generated by adversarial legalism. Perhaps they could be considered, as Welles and Engel suggest, compensation to the community for negative externalities of the waste disposal operation that could not be directly mitigated—compensation that communities in Pennsylvania, England, and the Netherlands did not receive because their regulatory system did not provide local objectors with the weapons of adversarial legalism. From another perspective, however, the costly California exactions might also be considered the product of legal extortion—social benefits that in fairness ought to be publicly debated in a legislative forum and paid for by the taxpaying public.

In sum, in a number of case studies, American adversarial legalism generated extra costs and no extra social benefits; in a few case studies, the balance of costs and benefits is more difficult to assess. Of course, the balance may be different with respect to other spheres of regulation, other industries, and other kinds of regulated companies. In some areas of American environmental regulation, some of our interviewees pointed out, technology forcing and tough enforcement have compelled American firms to take the lead in developing and implementing control technologies that firms overseas employ later, at lower cost and with less need for governmental coercion. As mentioned earlier, Richard Wokutch found that despite OSHA's legalistic regulatory style, large automobile manufacturers in

---

Japan, where regulators are far less legalistic, have better occupational safety and health practices than comparable American plants, but working conditions in small Japanese subcontractors were comparatively worse than those in small American parts suppliers.^[35] At the same time, the costs of American adversarial legalism undoubtedly bear more heavily on smaller firms, which can less easily afford to hire lawyers and consultants and to bear the delays that the larger multinationals described to us.^[36]

Some might argue that even if adversarial legalism does not yield better precautionary measures than less legalistic controls abroad, things would be worse without it because the United States is different. American business firms, some have argued, are less deferential to governmental authority than are Japanese and European firms.^[37] Perhaps Waste Corp.'s extended adversarial legal struggles in California and Ford Motor Company's lengthy permit processes in Minnesota and New Jersey reflect a tendency of American corporate officials to resist regulatory demands. The more rambunctious American business style, the argument goes, is one reason that American tort and regulatory law are—and ought to be—more adversarial, legalistic, and deterrence-oriented. Others would disagree, arguing that there is no empirical basis for concluding that American business firms, overall, are more resistant to regulation than British, Dutch, or German enterprises. The case studies in this book cannot resolve that issue. But they provide no hint that American corporate subsidiaries acted rambunctiously in other countries, resisting reasonable regulatory demands. Indeed, in most other countries, U.S. multinationals have a reputation for being especially good regulatory citizens. That weakens the argument that adversarial legalism is needed in the United States to tame a cowboy business culture, but it does not resolve it, since multinationals may not be representative of the American business culture as a whole.

More fundamentally, even if one assumes that adversarial legalism adds something to the deterrence equation, the overall size and distribution of those social benefits are highly uncertain. Nor is it at all clear that the presumed deterrent regulatory effects of adversarial legalism are large enough to offset the substantial social costs—defensiveness, litigation and lawyering expense, accountability costs, opportunity costs, and alienation— highlighted by the studies in this book. Perhaps most important, as suggested in the discussion of “divisiveness and delegitimation” in the preceding subsection, adversarial legalism may discourage the kind of business-government cooperation that is essential to the full achievement of regulatory goals. In analyzing national rates of pollution reduction, Scruggs found that “neocorporatist” governmental systems (such as Germany, the Netherlands, Sweden, and Japan), where strong industrial associations participate in consensus-seeking methods of policy making and implementation, exhibited stronger environmental performance than more “pluralist”

---

countries such as the United States, where policy making and implementation are more conflictual^[38]—and hence conducive to a prescriptive, legalistic, adversarial, and costly style of regulation that generates further mutual mistrust.

Finally, it might be argued that given American political culture, adversarial legalism is necessary to maintain the legitimacy of legal and regulatory processes. In a nation that mistrusts political authority and corporate power, adversarial legalism is valued, perhaps, not because it is efficient or effective but because it supports civic values such as governmental accountability and transparency, checks and balances, responsiveness to minority and local interests, and equal protection of the law. Several case studies in this book underscore the role of adversarial legalism in giving citizens a larger and more powerful voice in the regulatory permitting process than is enjoyed by their counterparts abroad. Adversarial legalism thus democratizes regulation to a certain extent. In some ways, this may add to its legitimacy. Moreover, some multinational corporate officials we spoke with suggested that despite its additional costs and inefficiencies, adversarial legalism helps ensure the integrity and evenhandedness of American regulatory and legal processes. Compared with the closed-door corporatist decision making of many other nations, American adversarial legalism provides both domestic and foreign companies greater assurance that they are competing on a level regulatory playing field and that all business firms enjoy legal recourse against official arbitrariness or favoritism. The American court system, some corporate executives note, is very expensive, but its powerful tools of legal discovery engender confidence that contracts can be enforced, trade secrets protected, and extortive demands repulsed. After all, viewed in comparison with other economically advanced democracies, the United States has been remarkably successful in fostering vibrant financial and venture capital markets, promoting innovation and entrepreneurial activity, and facilitating industrial restructuring; this, too, suggests that in some respects the American legal system has favorable effects on economic growth and competitiveness, at least in cross-national comparative terms.

That said, it does not follow that all aspects of American adversarial legalism are necessary or desirable. The studies in this volume suggest that in many particular spheres of legal activity, adversarial legalism is economically wasteful and sometimes counterproductive, imposing costs that other nations do not, without producing compensating social benefits. Even in a successful economy, wasteful legally imposed expenditures and legally created delays and uncertainties are worth worrying about.

---

VII. CONCLUSION

Governance in the United States is pervaded by adversarial legalism. Compared with other economically advanced democracies, American processes for policy implementation and dispute resolution generally involve more formal legal contestation, of a particularly adversarial and costly kind. American regulatory programs involve more detailed and complex bodies of rules, harsher penalties, a more legalistic enforcement style, more fragmented institutional structures, and more frequent recourse to courts to challenge administrative decisions. Liability law and civil litigation are more frequently employed in the United States than in Western Europe, Japan, and British Commonwealth nations, and those processes are more punitive, more expensive, and more unpredictable. In consequence, besides the direct costs of complying with federal and state laws and regulations, productive enterprises operating in the United States bear the additional costs of coping with adversarial legalism. These costs, we have seen, often are considerable.

Of course, no legal or regulatory system is free of exasperating “transaction costs.” When Shakespeare's Hamlet soliloquized about reasons for shuffling off this mortal coil, he cited “the law's delay” in the same breath as “the insolence of office” and “the pangs of dispriz'd love.” Nevertheless, in contemporary Denmark (or England), the law's delay is far less burdensome than in the United States, and its processes are significantly less costly. Comparative studies in numerous policy areas, as well as the studies of multinational corporations reported in this book, indicate that businesses in the United States usually face much higher dispute-resolution and regulatory “process costs” than do parallel business operations in other economically advanced democracies. These costs include:

1. Larger expenditures on lawyers, specialized regulatory affairs personnel, consultants, documentation of compliance, and other formal accountability mechanisms.
2. Longer delays in gaining regulatory approval for new projects and products.
3. More frequent legal conflict, higher monetary penalties, more diversion of managerial time, and more frequent public embarrassment— all of which stem from the more legalistic, unforgiving, and punitive American style of regulatory enforcement.
4. Much larger expenditures on liability insurance, litigation-related legal services, litigation avoidance, and compromise of legally justified claims and defenses—all of which stem from the more adversarial, expensive, and unpredictable American methods for adjudicating

   ---

   ― 407 ―
   civil cases, paying lawyers, and covering accident victims' lost wages and medical needs.

Indeed, according to the studies in this book and additional exploratory interviews that did not result in case studies, the costs, uncertainties, and delays associated with the American legal and regulatory style generally are more salient and significant to multinational enterprises than are crossnational differences in substantive law and regulatory standards.

At the same time, despite adversarial legalism's demanding processes and despite the deterrent threat of its harsher legal sanctions, multinational enterprises, the studies in this book tentatively suggest, often achieve levels of environmental protection and safety in their European and Japanese operations that are similar to those in their more legally threatened American facilities. This accords with numerous cross-national studies of environmental and safety regulation.

Finally, adversarial legalism, because it more often subjects productive enterprises to wasteful expenditures and unpredictable risks, has certain alienating effects. Multinational enterprises, comparing their experience in the United States and in other countries, often come to regard American legal and regulatory processes as arbitrary, threatening, and counterproductive.

These generalizations have been painted with a rather broad brush, however. American governments project an immense, constantly shifting array of rules into hundreds of different industries and thousands of individual enterprises. The policy areas covered by existing studies and by the case studies in this book represent only a sample of all the existing areas of law and commerce. It will not be difficult for knowledgeable readers to point to instances in which substantive differences in national regulatory or legal rules—such as American restrictions on banking activity, or German restrictions on employee layoffs—are much more important than differences in legal and regulatory style, or in which the fierceness of American law or regulation produces higher levels of protection than the legal regimes of comparable countries. The legalistic American approach to water pollution control, according to a study by Kathryn Harrison, has been more effective than Canada's cooperative style in producing compliance with certain regulatory standards on the part of pulp and paper mills.^[39] Japan's nonlegalistic mode of legal control, it appears, is far less effective than American adversarial legalism when the regulatory goal is to change established social norms and practices, such as promoting equal opportunity for women in employment or protecting workers and other nonsmokers from exposure to “secondhand” cigarette smoke.^[40] Moreover, some multinational companies we have studied pointed to instances in which legal processes in

---

Japan (e.g., in obtaining patents, in gaining approval for new chemical products) or in Europe (e.g., in obtaining licenses for genetically engineered products) were as slow, complicated, and unpredictable as in the United States, or even a little more so.

Nevertheless, the overwhelming weight of the existing literature and of our current research supports our generalizations concerning the greater pervasiveness, salience, and economic significance of American adversarial legalism. Thus far, the exceptions we have encountered do not come close to obviating the rule.

By and large, the case studies in this volume were not able to provide meaningful quantitative estimates of the “extra” costs specifically attributable to American adversarial legalism, either at the level of the whole economy or at the level of particular firms or industrial sectors. Since our research focused on large multinational enterprises that do have operations in the United States, we encountered no direct evidence that the burdens and risks associated with adversarial legalism scared away foreign businesses or drove out American firms. The United States benefits in this regard from the enormous size of the market it offers both American and foreign businesses, which appear to treat adversarial legalism as a price to be paid to gain access to this market. It does seem clear, however, that the costs of American adversarial legalism, as distinct from the costs of complying with substantive regulatory standards, rise to an economically significant level and often are very significant to individual enterprises. These costs detract from the economic efficiency of the American economy, saddling it with costs of a kind that are not borne by other national economies.

That does not exactly prove that the costs of American adversarial legalism outweigh its social benefits. The United States is a particularly diverse, undeferential society. Multinational enterprises operating in America undoubtedly pay more for security guards, employee drug testing, and the like than they do in parallel facilities in Japan or Germany; such extra security costs, in comparative terms, operate as a drag on the American economy. Still, given the social conditions in the United States, it would be foolish not to invest in the extra protection. Similarly, American regulators work in a more diverse, less deferential, and difficult-to-regulate business environment than their Japanese, Canadian, and European counterparts. Sprawling and decentralized, the United States does not have the strong trade associations, respected national bureaucracies, and comprehensive social welfare systems that characterize Western European states. Thus, to maintain similar levels of protection for its citizens, it might be argued, the United States has little choice but to employ the shock troops of adversarial legalism—legal rights and penalties, lawyers, and litigation. Adversarial legalism may be a costly, inefficient method, but without it, conceivably, substantive legal and regulatory policies would be evaded more often.

---

Moreover, the same legal arrangements that lead to high litigation costs, unpredictability, and defensiveness—easy access to courts, entrepreneurial lawyering, extensive pretrial discovery, strong legal remedies, a politically responsive (rather than a narrowly professional) judiciary, institutional fragmentation (checks and balances)—also are used to attack and deter governmental unresponsiveness, bureaucratic arbitrariness, abuse of economic power, discrimination, and negligence. Adversarial legalism makes the American political and legal system more responsive to new ideas and justice claims often generating regulatory ideas and norms that other countries later emulate. Thus, despite its failings and inefficiencies, some would argue, adversarial legalism makes a positive contribution to social welfare.

Adversarial legalism's admirable functions, however, do not justify ignoring its costly excesses. Some defenders of American adversarial legalism, afraid that conservative reformers will throw the baby out with the bathwater, tend to dismiss attacks on the system's costs and inefficiencies as unrepresentative exaggerations. They act as if any diminution of adversarial legalism in any sphere of law or any regulatory program would have adverse effects. That surely is not the case. Although the techniques of adversarial legalism apply across a broad range of American policy arenas, they were institutionalized and are now invoked on a program-by-program, case-by-case basis. They can be reassessed and moderated in the same way.

NOTES

---

---