Most countries regulate manufacturer prices for pharmaceuticals, either directly (France, Italy) or indirectly through controls on reimbursement (Germany, Japan) or profits (the UK). It is widely believed that drug prices are lower in countries with strict price regulation than in countries with less restrictive regulation (the UK) or no regulation (the US). For example, the BEUC (1989b) concluded that prices in the UK and the US were, respectively, 20 and 54 percent above the EEC average, whereas those in France and Italy were, respectively, 30 and 28 percent below the EEC average. The US General Accounting Office (GAO, 1992, 1994a) concluded that prices in the US in 1992 were 32 percent higher than prices in Canada and 60 percent higher than prices in the UK. A UK Department of Health study (DOH, 1997) concluded that US prices 88 percent higher than the UK in 1992. A recent US study (US H.R. Minority Staff, 1998) reported that drug prices in the US were 70 percent higher than in Canada and 102 percent higher than in Mexico. In the US, these studies have contributed to proposals (so far not enacted) for drug price controls – for example, President Clinton’s Health Security Act (1993) and the Prescription Drug Fairness Act of 1999 (H.R.644). In addition to these comparisons of average price levels, a growing number of countries, including Italy, Spain, the Netherlands, Canada and Japan, use international comparisons in their regulation of prices for individual drugs.

The first purpose of this paper is to report indexes of manufacturer-level drug prices for six major markets – the UK, Canada, France, Germany, Italy, and Japan – relative to the US, using comprehensive data and more appropriate methods than those used in previous studies. Our data are from Intercontinental Medical Systems (IMS), a market research firm that collects data on drug sales worldwide. IMS data for all outpatient drug sales in 1992 are used to construct price indexes based on all molecules that match across the countries under comparison, including branded products (whether sold by the originator company or a licensee, hereafter ‘originator’ and ‘licensed’ products) and generic products, and all formulations, strengths and packs. Standard price indexes, weighted by either US consumption patterns or the comparison country’s consumption, are computed. The indexes with US quantity weights show average prices for each country, relative to the US, as follows: the UK –17 percent, Canada +2 percent; Germany +25 percent; France –32 percent; Italy –13 percent; and Japan –12 percent. Thus this analysis, using a comprehensive market basket including generics and appropriate weighting, shows that crossnational price differences are less than suggested by previous studies. The indexes with UK consumption weights show the UK either lowest based on price per gram or third lowest, after France and Italy, based on price per pill. The bias in previous studies for the US is shown to result from selection of very small, unrepresentative samples of leading branded products, exclusion of all generics, and reporting of unweighted averages which give undue weight to the highest priced products. The Appendix to this paper details the differences between our price comparisons and those made by the GAO and BEUC.

The cross-national diversity in range of products available and in patterns of drug consumption implies that there is no single ‘correct’ measure of price differences. Methodological judgements are unavoidable – in particular, choice of sample, weights and measure of price – and the most appropriate measure depends on the perspective of the analysis. However, while some choices depend on perspective, the analysis here clearly shows that a robust estimate should be based on a representative sample, including generics, and standard weighted indexes, not simple averages, as used in BEUC (1989), GAO (1992) and the OECD Pharmaceutical Purchasing Power Parities (OECD, 1993).

The second purpose of this paper is to examine the extent of competition under alternative regulatory regimes. Regulation is often rationalised by the assumption that price competition is weak because: insurance makes patients insensitive to prices; physicians who are primary decision-makers may not know product prices and/or may be imperfect agents for patients; patents intentionally limit competition from generically equivalent substitutes; and therapeutic substitutes are imperfect. Retail pharmacy prices and other aspects of retail pharmacy are also regulated in countries that regulate manufacturer prices. Previous studies have found evidence of price competition in the UK, the US and Germany (Reekie, 1996; Towse and Leighton, 1999). However, both theory and casual empirical evidence suggest that strict price regulation may undermine competition. Generic market shares of off-patent products are significantly lower in countries with strict price or reimbursement regulation, such as France, Italy or Japan, than in the US, the UK, Canada and Germany, which have less strict price regulation. Whether regulation reinforces or undermines competition is an important empirical question, as different countries evaluate possible changes in their regulatory regimes. This paper estimates the effects of generic competition, therapeutic competition and other factors in the seven countries with their different regulatory regimes.

The main findings are that generic competition significantly reduces prices in countries with free pricing (the US) and moderately constrained pricing (the UK, Germany and Canada), whereas generic competition is ineffective and may be counterproductive in countries with strict price or reimbursement regulation (France, Italy and Japan). One plausible explanation is that in regulatory regimes that drive down the originator price over the life-cycle, generic equivalents are often licensed co-marketers or minor ‘new’ versions of old molecules introduced by manufacturers as a strategy to obtain a higher regulated price. By contrast, in countries with free pricing and price sensitive purchasers, generic entrants must compete on price to gain market share. For therapeutic substitutes, the results here confirm previous findings (Reekie, 1996; Towse and Leighton, 1999) that successive molecules to enter a therapeutic category do so at lower prices than those of established entrants.

This analysis has important implications for the methodology of drug price comparisons and for policy. First, robust price comparisons require representative samples and standard indexes. Limiting the sample to leading branded products and use of unweighted averages can yield very misleading results that tend to systematically overestimate prices in unregulated or less regulated markets, compared to strictly regulated markets.

Second, regulation clearly undermines competition in the off-patent, multisource sector, which is contrary to sound economics and to the stated aim of policy in most countries. Innovative products are appropriately protected from generic competition for the life of the patent, in order to yield a return on R&D investments. However once the patent has expired, price competition between generic substitutes can yield significant savings to consumers and payers. Off-patent drugs account for 88 percent of reimbursable packs sold on average for member states of the EU (European Commission, 1998), and this off-patent share is expected to grow as patents expire on many of the current leading drugs. Increasing competition in the off-patent sector to free up ‘headroom’ in public budgets to pay for innovative, patent-protected products was suggested by the Bangemann Round Table discussions on the European single market for pharmaceuticals (European Commission, 1998; Danzon, 1998). Designing regulatory systems to promote competition in the off-patent sector is an important issue for all governments concerned with obtaining maximum value from health spending. The evidence here suggests that useful pro-competitive strategies include: (1) permitting pharmacists to substitute between generically-equivalent products (generic substitution), unless the physician indicates otherwise; and (2) promoting competition in retail pharmacy, by deregulating dispensing fees.

For the on-patent sector, policy conclusions are more tentative because the empirical evidence is less robust and because optimal competition policy must weigh the consumer benefits of lower prices against the need to provide an opportunity for originator firms to recoup their R&D investments, which is the intent of patent protection.

Whether pharmaceutical prices should differ between countries and, if so, the appropriate magnitude of such differences, has been addressed elsewhere (Danzon, 1997b, c) and is not discussed in detail here. Economic theory indicates that uniform prices would not be optimal and that price differences (Ramsey pricing) for patented products are the most efficient practical strategy to pay for the common costs of R&D that serve all consumers. The potential for price differences within the EU arises because health care is a national policy prerogative and different countries have pursued very different regulatory strategies. The growth of parallel trade, whereby wholesalers import products from lower-priced to higher-priced countries has narrowed sustainable differences, particularly on high-volume branded products. Regulation based on cross-national price comparisons has similar effects. Because of both of these factors, a comparison based solely on recently-launched branded products might show smaller differences between EU countries than the differences reported here based on 1992 prices which may still reflect, via the older products, the regulatory regimes and exchange rates that prevailed in the 1980s . Although the price data used here may not accurately reflect current price differentials on newly-launched products, these data do show the methodological issues raised by cross-national price comparisons and the biases in previous comparisons, as well as the broad effects of regulatory systems on competition.

The report is structured as follows. Section 2 describes the data. Section 3 reports price and quantity indexes, and compares the results here with those in previous cross-national price comparisons. Section 4 outlines a simple model of drug prices and the expected effects of regulation. Section 5 describes the empirical model and methods. Section 6 reports product-level regression analysis of product prices and tests for significant differences between countries. Section 7 reports similar analysis for prices at the molecule level, where the molecule price is a weighted average over all products in the molecule. Section 8 concludes.