regulation-innovation dynamics for drugs can provide very valuable lessons for the on going shaping and re shaping of public policy across a wide range of other sectors in the economy.

The conclusions on the regulatory experience in the drug industry are relevant for other industries. The first conclusion is that none of the assessments of the over all effects of the drug lag made thus far have identified fully and quantified accurately the total societal benefits and costs resulting from the lag, but they have laid the beginnings of a good foundation for doing so. Parenthetically, in the absence of a definitive impact assessment, it is not surprising that the 1962 Amendments have not been amended, despite strong criticism of their effects on innovation.

The second conclusion is that there is a causal relationship between drug regulation and innovation: drug regulation of the sort imposed by the 1962 Amendments has increased the cost of new drug development; cost increases, in turn, have altered the relative abilities of firms to pursue drug research and development ("R&D"), and thus have affected R&D concentration in the drug industry. Increases in development times additionally have shortened effective commercial patent lives of new drugs. Higher monetary and time costs, by affecting structural conditions in drug markets, ultimately have directly and indirectly impacted on both the magnitude and pattern of new drugs developed by firms.

This study is structured along lines suggested by the above description of its orientation and conclusions. This statement of introduction serves as section I. Section II reviews the well traveled ground of the economic and medical literature on the drug lag, touching lightly on findings in regard to the existence of the lag and its relation to the 1962 Amendments, and somewhat more heavily on assessments of the full societal impacts of the Amendments and/or the lag.

Sections III, IV, and V relate respectively to the ways described above in which the 1962 Amendments have affected R&D activity in the drug industry. Section III sets out the effects of the Amendments on the development costs and periods for new drugs. Section IV assesses the impact of higher R&D costs on the essential structural elements of drug markets — economies of scale, concentration of R&D efforts, and product patent life — that affect the capabilities and incentives for innovation. Section V examines firm R&D strategies that reflect the cost and structural influences described in sections III and IV. Finally, section VI reviews the findings of the study, suggesting methods for improving the assess

ment of the full impacts of the 1962 Amendments and articulating the "lessons from the drug lag" that may be most useful for other areas of regulation.

II. THE EVIDENCE ON THE Drug Lag

This section assesses the economic and medical evidence on the three major hypotheses in the drug lag debate: (1) that the U.S. is experiencing a gap between the present record of drug innovation and that of the past, or between our record and that of other countries; (2) that this gap, or lag, results largely from the stricter testing and approval standards imposed by the 1962 Amendments; and (3) that this lag, on balance, imposes costs on society that significantly outweigh its benefits.

The studies on the drug lag issue range from simple data presentation that offer only limited implications regarding the existence of a drug lag to sophisticated medical and economic analyses that offer more substantial conclusions about all three hypotheses. This section briefly reviews the more important of these studies and their conclusions on the magnitude, causes, and impact of a drug lag in the United States.'

The Drug Lag and Its Causes

The simplest type of evidence on the existence and magnitude of a drug lag compares the rate of introduction of all or some subset of new drugs before and after some point in time, usually 1962. These studies generally agree that both the overall rate of new drug introduction and the rate of introduction of new chemical entities ("NCES") have declined substantially during the 1960s and at a slower albeit still substantial rate during the 1970s.

Yet this approach has shed little light on the three hypotheses. These conclusions are challenged by other comparisons of the rates of introduction of "significant" or "important" therapeutic discoveries before and after 1962. Moreover, the subjective nature of selecting which discoveries to include in the comparative studies limits their use as a precise measure of innovative achievement in drugs. The degree to which comparisons based on such selective categories conflict with those based on more inclusive measures of drug R&D output qualifies any firm conclusions about the significance of the observed changes in the rate of innovation. A

2 An earlier, more comprehensive analysis of these studies appears in Schifrin & Tayan, The Drug Lag: An Interpretive Review of the Literature, INT'L J. HEALTH SERVICES, (Winter 1977).

much greater limitation of these comparative studies is that as the relevant data are more closely observed, the downturns in drug innovation clearly began before the Amendments were passed and long before they were implemented. Thus such studies tell us, at most, that recent rates of drug innovation, by some measures, are lower than previous rates; they do not tell us whether the observed rates are below the normal or expected rates and if so, what factors, including the 1962 Amendments, are the cause.

A major improvement on this comparative approach was made by Peltzman' in an imaginative, if not fully successful, effort to measure the drug lag and to assess its impacts. Peltzman first defines the lag in an intertemporal fashion, as the difference between the actual flow of NCEs each year after 1962 (through 1969) and the flow predicted for each year from regression analysis of the determinants of the pre-1962 annual rate of introduction of NCES. Peltzman finds the actual post-1962 rates were approximately half of the predicted post-1962 benchmark rates.

Peltzman's work has been criticized on several counts. The most important is that he has overstated the lag by failing to give proper weight to the downward trends in drug development that began to appear in the late 1950s. If other factors were contributing to this decline before 1962, then identifying it as a post-1962 phenomenon wholly attributable to the 1962 Amendments clearly is in error. Yet Peltzman's study should not be dismissed solely on this basis, for it measurably raised the level of discourse on the drug lag. Furthermore, its qualitative conclusions about the drug lag and the role of the 1962 Amendments generally have been substantiated.

Subsequent analyses of the drug lag in the United States have not used intertemporal comparisons of drug innovation rates. They have avoided the errors of failing to account for diminishing research opportunities, exogenous increases in R&D costs, and other influences on drug innovation not related to the new regulations by using international comparisons of drug innovation. Since major changes in drug research opportunities, methodology, and productivity would affect innovation rates in many countries, the effects of a significant regulatory change made only in the United States would be measured more accurately by the differences in the

3 Peltzman's work has been published in several forms. His most complete presentation is S. PELTZMAN, THE BENEFITS AND COSTS OF NEW Drug RegulatiON, R. Landau, ed. 1973); see also S. Pri1zman, RegulatION OF PHARMACEUTICAL Innovation, THE 1962 AMENDMENTS (1974); An Evaluation of Consumer Protection Legislation: The 1962 Drug Amendments, 81 J. POL. Econ. 1049-91 (1973).

innovation rates in the United States and other countries. Most of these studies have used Great Britain as the basis of comparison for the United States, but some also have included other European countries, Canada, and Japan.

Wardell has provided the most thorough and persuasive medical assessments of the drug lag using international comparisons. Comparing the United States and British records for three time periods since 1962 (1962-1971, 1971-1974, and 1972-1976), he found that drug availability is more constrained in the United States in three respects: there are more drugs available in Britain that are not available in the United States than vice versa; drugs that are available in both countries are more often introduced in Britain before being introduced in the United States; and drugs that are available in both countries are more likely to be approved for a wider range of indications in Britain than in the United States.

These phenomena were first observed in the original study of the drug lag during the decade 1962-1971. In his second study, covering 1972-1974, Wardell found some aspects of the earlier lag to have narrowed, but for drugs introduced in the three years after the first study the same sorts of lag in the U.S. appeared as had in 1962-1971. Finally, in a recent study of drugs introduced during the 1972-1976 period in Britain and the United States, Wardell again found a narrowing of the original lag, but nevertheless a continued lag for the newly-introduced drugs in the United States. Wardell found that the overall lag in the United States relative to Great Britain has diminished in recent years. Yet there still is a lag, more significant in some therapeutic areas than others, in the availability, time of introduction, and range of application of drugs in the United States.

Grabowski has contributed greatly to identifying and estimating the drug lag in the United States by combining the best features of Wardell's and Peltzman's methodologies.' Like Wardell,

4 For a listing of Wardell's significant drug lag publications through 1975, see Schifrin and Tayan, supra note 1. Among his many later studies, most of which are available from the Center for the Study of Drug Development, University of Rochester Medical Center, Rochester, New York, are Wardell, 24 CLINICAL PHARMACOLOGY AND THERAPEUTICS 499-524 (1978); Wardell, Development of New Drugs Originated and Acquired by U.S. Owned Pharmaceutical Firms 1963-76 (unpublished manuscript).

5 H. Grabowski, Drug Regulation and Innovation (1976); see also H. Grabowski, Regulation and the International Diffusion of Pharmaceuticals, Conference on the International Supply of Medicines, American Enterprise Institute, Washington, D.C. (September 15, 1978); Grabowski, Vernon & Thomas, THE Effects of Regulatory Policy ON THE ÎN CENTIVES TO Innovate: An International Comparative Analysis, reprinted in Impact

Grabowski use; international comparisons of NCE introduction rates; and like Peltzman, he employs econometric methods to estimate the extent of the lag. His key findings are that both the U.S. and Great Britain (and by extension, probably all countries) have experienced a declining productivity in drug R&D, meaning lewer R&D outputs (NCE:s) per R&D dollar input. However, Grabowski also finds the productivity decline in the United States during the 1960s was approximately twice that of Britain. He attributes half of the U.S. decline and all of the British decline in drug R&D productivity to various factors, most notably a worldwide “depletion of research opportunities," that affect all nations. The remaining half of the productivity decline in the United States (the United States "drug lag") is attributed to regulatory policy.

Grabowski thus avoids the pitfalls encountered by Peltzman by using international rather than intertemporal comparisons. His definition of the regulatory-induced lag is much different than Peltzman's, and his estimation of it is decidedly smaller. Yet his qualitative findings that drug innovation in the U.S. has been influenced negatively by the 1962 Amendments are in agreement with Peltzman.

Peltzman, Wardell, and Grabowski, particularly the latter two, offer persuasive evidence that since the early 1960s there have existed differences in the availability of drugs in the United States relative to Great Britain that may properly be labelled a drug lag. Data provided by others, particularly de Haen" and Lasagna,' reinforce and extend this conclusion. Their calculations of new drug introductions in the United States, Great Britain, Germany, and France show that, to some extent, each country experiences a drug lag relative to at least a few others, and that the magnitude of the lag varies from one class of drugs to another. Their major point is that by most measures of new drug innovation the United States clearly lags behind most other Western countries in both the rate and timing of such introductions. Thus the first of the drug lag questions whether such a lag in fact exists must be

-

of Public Policy on Drug Innovation and Pricing, (Mitchell and Links, eds. 1976); Grabowski, Vernon & Thomas, Estimating the Effects of Regulation on Innovation: An International Comparative Analysis of the Pharmaceutical Industry, 21 J. Law & Econ. 133-63 (1978).

6 de Haen, The Drug Lag Does It Exist in Europe, 9 Drug IntelligenCE AND CLINICAL PHARMACY 144-50 (1975).

7 Lasagna, Research, Regulation, and Development of New Pharmaceuticals Past, Present, and Future, Part II, 263 AMERICAN Journal of Medical Science 66-78 (1972).