ArticlePDF Available

Effects of Regulation on Drug Launch and Pricing in Interdependent Markets

November 2012
Advances in Health Economics and Health Services Research 23:35-71

DOI:10.1108/S0731-2199(2012)0000023005

Source
PubMed

Authors:

University of Pennsylvania

This study examines the effect of price regulation and competition on launch timing and pricing of new drugs. Our data cover launch experience in 15 countries from 1992 to 2003 for drugs in 12 major therapeutic classes. We estimate a two-equation model of launch hazard and launch price of new drugs. We find that launch timing and prices of new drugs are related to a country's average prices of established products in a class. Thus to the extent that price regulation reduces price levels, such regulation directly contributes to launch delay in the regulating country. Regulation by external referencing, whereby high-price countries reference low-price countries, also has indirect or spillover effects, contributing to launch delay and higher launch prices in low-price referenced countries. Referencing policies adopted in high-price countries indirectly impose welfare loss on low-price countries. These findings have implications for US proposals to constrain pharmaceutical prices through external referencing and drug importation.

. Launch and Molecule Count and Mean and Median Launch Delay by Country and Subclass.

…

. Marginal Effects for Launch Model.

…

. (Continued )

…

. Marginal Effects of Prior Foreign Launch on Launch Hazard in Low-Price EU Countries for Superior Subclasses.

…

. Determinants of Launch Prices, OLS, and Normal Random Effects Regressions.

…

Figures - uploaded by Patricia Danzon

Content may be subject to copyright.

Content uploaded by Patricia Danzon

Content may be subject to copyright.

EFFECTS OF REGULATION ON

DRUG LAUNCH AND PRICING IN

INTERDEPENDENT MARKETS

Patricia M. Danzon and Andrew J. Epstein

ABSTRACT

Purpose – This study examines the effect of price regulation and

competition on launch timing and pricing of new drugs.

Methods – Our data cover launch experience in 15 countries from 1992 to

2003 for drugs in 12 major therapeutic classes. We estimate a two-

equation model of launch hazard and launch price of new drugs.

Findings – We ﬁnd that launch timing and prices of new drugs are related

to a country’s average prices of established products in a class. Thus to

the extent that price regulation reduces price levels, such regulation

directly contributes to launch delay in the regulating country. Regulation

by external referencing, whereby high-price countries reference low-price

countries, also has indirect or spillover effects, contributing to launch

delay and higher launch prices in low-price referenced countries.

Implications – Referencing policies adopted in high-price countries

indirectly impose welfare loss on low-price countries. These ﬁndings have

The Economics of Medical Technology

Advances in Health Economics and Health Services Research, Volume 23, 35–71

All rights of reproduction in any form reserved

ISSN: 0731-2199/doi:10.1108/S0731-2199(2012)0000023005

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

implications for US proposals to constrain pharmaceutical prices through

external referencing and drug importation.

Keywords: Pharmaceutical economics; drug pricing

INTRODUCTION

New drugs are potentially global products and can contribute importantly

to health outcomes for consumers and expenditures for payers. Prompt

launch is also critical for drug manufacturers, given the ﬁxed patent life

over which to recoup the high costs of R&D.

In fact, in a study of the 1990s

launch experience of 85 drugs in 25 industrialized countries, roughly half

of the potential country-compound launches occurred, and many of the

eventual launches involved months or years of delay (Danzon, Wang, &

Wang, 2005). Launch lags in the United States in the 1960s and 1970s

were attributed to increased regulatory requirements for proof of safety and

efﬁcacy (e.g., Grabowski, Vernon, & Thomas, 1978;Peltzman, 1973). How-

ever, in the 1990s the United States and the European Union harmonized

and accelerated new drug review, such that remaining differences in regis-

tration requirements cannot fully explain observed launch lags, especially

between EU countries.

By contrast, price regulation has become more complex and potentially

contributes both direct (domestic) effects on launch lags in the regulating

country and indirect (spillover) effects on launch in other countries. Price

regulation may delay launch directly through three mechanisms. First,

regulation that reduces the manufacturer’s expected price and net present

value (NPV) of revenues reduces incentives for launch, especially in small

countries and for drugs with low expected sales volume, assuming some

ﬁxed costs of launch. Second, negotiation strategies may lead to strategic

delay by ﬁrms or regulators to inﬂuence the ultimate price. Third, regula-

tory processes may entail pure bureaucratic delay. The welfare con-

sequences of these direct/domestic effects of a country’s regulation on its

citizens are ambiguous a priori, because any foregone health beneﬁt due to

fewer/later launches may be offset by savings from lower drug prices

and better pre-launch information about a drug’s relative safety and

effectiveness.

More problematic from a social welfare perspective are the indirect/

spillover effects that arise when one country regulates its drug price with

PATRICIA M. DANZON AND ANDREW J. EPSTEIN36

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

reference to the price of the same drug in other countries (‘‘external

referencing’’). For example, many EU countries cap their prices at the mean,

median, or minimum price in a group of other EU countries. By under-

mining market segmentation and price discrimination, external referencing

by high-price countries creates spillover incentives for a ﬁrm to seek higher

prices or not launch in lower-price referenced countries. The welfare

consequences in the referenced low-price countries are clearly negative, since

they suffer reduced access to new drugs and possibly higher prices due to

external referencing by other countries, with no offsetting beneﬁts. Parallel

importation by high-price EU countries from low-price EU countries is a

second spillover mechanism that may also contribute to non-launch and

higher prices, and hence negative welfare consequences for the exporting

countries.

Understanding these indirect effects of one country’s regulation

on launch in other countries is particularly important as the scope of refer-

encing is increasing in the EU and in other regions, and the United States

considers both referencing foreign drug prices and/or legalizing parallel

trade (‘‘drug importation’’).

Previous studies of launch experience for new drugs (Danzon, Wang, &

Wang, 2005;Kyle, 2006, 2007;Lanjouw, 2005) have generally concluded

that price regulation contributes to launch delay. As detailed below, how-

ever, these prior estimates are inconclusive because: their proxy measures of

regulation were at best rough and sometimes inaccurate, leading to possible

measurement error and some contradictory results; they lacked product-

speciﬁc data on prices of competitor products, a critical benchmark for

regulation and hence expected launch prices; they estimated launch equa-

tions that presupposed regulatory effects on launch prices, but were unable to

corroborate with evidence on launch prices; and none clearly distinguished

the indirect spillover effects of regulation from benign determinants of

launch sequencing.

This chapter provides robust estimates of the effects of regulation on both

launch delay and launch prices. We distinguish clearly between the direct/

domestic effects of a country’s own regulatory system and the indirect/

spillover effects on other countries, and clarify the negative welfare con-

sequences of the spillover effects. We use the average price (lagged one

quarter) of established products in the same subclass as the most accurate

available measure of the effect of regulation on expected prices and hence

the direct effect of regulation on launch.

We distinguish the inﬂuence of

products in the same subclass vs. older subclass, to provide evidence on

whether the availability of older, cheaper products constrains launch prices

for new drugs in innovative subclasses and hence the extent of static

Effects of Regulation on Drug Launch and Pricing 37

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

vs. dynamic competition. We also provide evidence on (lack of) ﬁrst-mover

advantage and on the inﬂuence of local corporations on launch timing and

prices in some countries. Additionally, to identify the impact of external

referencing spillovers separately from rational launch sequencing, we

compare the effects on launch delay and pricing in low-priced EU countries

of prior launch in high-priced EU countries with prior launch in high-priced

non-EU countries.

Our study uses quarterly sales data, by drug, from IMS Health from 1992

to 2003 for 15 major countries and 12 major therapeutic classes, all of

which experienced entry of a new subclass during our study period (e.g., in

the antiulcerant class, the proton pump inhibitors [PPIs] displaced the

H2-antagonists). We estimate a two-equation model of launch hazard and

launch price for molecules in the superior subclasses.

We ﬁnd that launch hazards and prices of new drugs are positively

associated with prices of established products in the same new subclass,

but ﬁnd no evidence of effects of old subclasses on new subclasses. Thus,

to the extent that regulation reduces drug prices, it directly contributes

to launch delay/non-launch, but availability of cheaper drugs in older

subclasses does not constrain drug prices for new subclasses. Regulatory

referencing by high-price EU countries contributes to launch delay/

non-launch in low-price EU countries. The evidence here shows that

indirect regulatory strategies that rely on referencing to foreign prices or

importing from foreign countries can impose signiﬁcant welfare loss on these

foreign countries due to launch lags, non-launch, and higher prices. The

model implies that such effects will be greater, the larger the referencing

country relative to the referenced country and the greater the potential price

difference. Thus, this evidence is particularly relevant as the United States

considers external referencing and/or legalizing drug importation/parallel

trade, given the huge size of the US market (almost 40% of global

pharmaceutical sales) relative to potential referenced countries.

It is also

relevant to the growing concern over high prices of drugs in developing

countries, which are plausibly attributable in part to the risk of referencing

or importation by higher-priced countries, as well as other factors.

In this chapter, we ﬁrst review the relevant literature and describe the

main types of price regulation and their expected direct and indirect effects

on drug launch. We then outline our theoretical model, the empirical

speciﬁcation, data, and variable deﬁnitions. The ﬁnal sections report the

empirical evidence on direct and indirect effects of regulation on drug

launch timing and launch prices, and draw policy conclusions.

PATRICIA M. DANZON AND ANDREW J. EPSTEIN38

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

LITERATURE

Several studies have examined effects of price regulation on lags in new drug

launch. Danzon, Wang, and Wang (2005) studied the launch experience of

85 new drugs in the 25 leading markets in the 1990s, focusing on drugs that

had met registration requirements of one of the two strictest agencies (the

US Food and Drug Administration [FDA] and the UK Medicines Agency)

and hence could potentially meet registration requirements in other

countries. This analysis is used as a proxy for the direct regulatory effect

on a new drug’s expected price the average price of all competitor products

in the class, measured at the time of the new drug’s global launch. This

class-wide average price at global launch may be an imprecise measure of

expected price for a new drug in a new subclass, because the average

includes older subclasses and generics and because several years may have

elapsed between ﬁrst global launch and launch in a particular country. Our

present study estimates separate effects of (a) the average price of other

on-patent drugs in the new subclass (‘‘superior’’ brands) vs. (b) the average

price of older brands in other, related subclasses (‘‘inferior’’ brands), and

quarter prior to launch). We ﬁnd that launch of new brands is affected only

by prices of brands in the same subclass. That prices of new drugs are not

constrained by prices of older, cheaper brand, or generic substitutes has

important implications for regulatory policy.

Kyle (2007) used a more heterogeneous sample of compounds and

countries, a longer time period (1980–1999), and measured regulation with a

vector of regulatory structure indicators and a price rank indicator. She

concluded that price controls reduced launch probability in countries that

impose them. However, the regulatory indicators are from 2000/2002; hence

they postdate the 1980–1999 analysis period and do not reﬂect signiﬁcant

changes in several countries within the analysis period. The indicators

incorrectly classify some countries and are missing for others. More funda-

mentally, within each measured regulatory type, there are major unmea-

sured differences across countries and over time that affect prices. Kyle’s

only measure of expected price is an indicator for a country’s price rank in

2002, which is invariant across classes and years. This was negatively related

to launch probability prior to 1995, contrary to predictions, and insigni-

ﬁcant after 1995. Kyle’s (2006) analysis of launch in the G7 countries uses a

single binary indicator for drug price controls. This is negatively associated

with launch probability in some speciﬁcations, but is insigniﬁcant in others.

Effects of Regulation on Drug Launch and Pricing 39

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

Thus, this evidence for the direct effects of regulation on launch proba-

bilities from these analyses using regulatory indicators is inconclusive.

With respect to spillover effects of regulation, these prior papers use

indicator variables for prior launch in a high-price vs. low-price country.

However, none tests for differential effects between countries that are

closely linked vs. those that are not, which is necessary to distinguish

regulatory spillovers from other unobserved factors that may lead to

sequenced launches, such as coordinated regulatory ﬁlings. Finally, these

papers show that launch by a local corporation increases launch proba-

bilities on average, but there is no analysis of differential effects across

countries and no compelling evidence on whether this reﬂects a true

experience advantage or simply political bias towards local ﬁrms.

Lanjouw (2005) examined launch in a large, heterogeneous sample of new

drugs in 68 countries between 1982 and 2002, using covariates measured at

ﬁrst global launch. She also lacked data on competitor prices and measured

regulatory stringency with binary indicator variables that are invariant

across classes and over the 20-year study period. Given the imprecise

measure of regulation, the conclusions on direct effects of price regulation

are inconclusive. Lanjouw did not address spillover effects.

None of the previous papers examine effects of regulation on launch

prices. Kanavos and Costa-Font (2005) argue that parallel trade has led

regulators in the lower-price EU countries, such as Italy, France, and

Portugal, to accept higher prices, but they do not directly test this. Similarly,

Grossman and Lai (2006) present a theoretical model in which regulators in

parallel export markets accept higher regulated prices under regimes of

international patent exhaustion (i.e., parallel trade is permitted), but no

empirical evidence is presented. Our chapter is the ﬁrst to provide evidence

of effects of regulatory spillovers on launch prices.

PHARMACEUTICAL PRICE REGULATION AND

ITS MEASUREMENT

New drugs facetwo possible regulatory hurdles: registration and price approval.

Registration

In all countries, drug registration requires proof of safety, efﬁcacy, and

manufacturing quality as a condition of market access. In the 1990s, the

US FDA and counterpart agencies in Europe harmonized some data

PATRICIA M. DANZON AND ANDREW J. EPSTEIN40

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

requirements and adopted measures to accelerate review while retaining

autonomy in decision-making. Since 1995, the European Medicines Agency

(EMA) has offered both centralized and mutual recognition procedures that

can lead to simultaneous registration of new drugs in all EU countries, as an

alternative to the traditional country-by-country review through national

drug-approval agencies.

Thus, cross-national differences in drug registra-

tion regulation cannot explain large systematic differences in launch timing

among EU countries or between the United States and European Union.

Japan is an exception in retaining special requirements, including clinical

trials on Japanese citizens.

Price/Reimbursement Regulation

Once a new drug clears registration hurdles, most countries with national

health insurance systems require price approval as a condition of reimburse-

ment.

Countries use one or both of two criteria to set launch prices:

(a) ‘‘internal referencing’’ to prices of one or more competitor products in

the same therapeutic class, with potential for mark-ups for superior efﬁcacy

or other factors,

or (b) ‘‘external referencing’’ to the minimum, median, or

mean of prices of the same drug in speciﬁed comparator countries. Most

price regulatory regimes disallow post-launch price increases, and price cuts

are sometimes mandated; hence, the launch price is critical to a drug’s life

cycle price proﬁle. Internal referencing may entail bureaucratic delay and

possibly strategic delay if ﬁrms (regulators) hold out to achieve a higher

(lower) price. However, regulators should have incentives to weigh any costs

associated with launch delay against the beneﬁts in lower prices, with no

signiﬁcant spillover to other countries.

By contrast, because external referencing regimes benchmark their price

to the price of the same drug in other countries, they create incentives

for ﬁrms to delay launch in referenced lower-price countries until prices

have been established in other potentially higher-price countries and/or

regulators in low-price countries accept higher prices. Thus, higher-price

countries that reference lower-price countries can impose launch delay and

possibly contribute to higher prices in the referenced countries.

Measurement

Identifying the contribution of these regulatory regimes to drug launch

experience in speciﬁc countries is complex because most countries use

Effects of Regulation on Drug Launch and Pricing 41

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

multiple forms of regulation, including both internal and external referen-

cing, and the details of each regulation type differ across countries. The

effects of internal referencing depend on the choice of comparator products

and criteria for and size of mark-ups, if any. External referencing effects

depend on countries referenced, whether minimum, median, or mean price is

used, the other countries that the referenced countries reference, exchange

rates, and whether external referencing is only one among several factors

considered. The complexity of these regulatory details and changes within

countries over time make it impossible to accurately categorize countries

by regulatory indicators or derive a ﬁrm’s optimal launch sequence and

minimum reservation price for each drug–country pair. However, a clear

prediction is that referencing of low-price countries by high-price countries

will exacerbate launch delay and possibly raise prices in the low-price

countries, especially if they are small relative to the referencing countries. Of

the countries in our data during our study period, France, Japan, Canada,

the Netherlands, Italy, Mexico, and Brazil nominally used both internal and

external referencing for some drugs in at least part of our time period.

Greece and Portugal also used external referencing for most of the period.

In Germany and the United Kingdom, a new drug could be launched and

reimbursed without regulatory approval of price, although governments

monitor average international price differences and other control mechan-

isms applied.

In the United States, private sector plans negotiate

prices, and some public sector plans (e.g., Medicaid, the Veterans Admini-

stration) mandate either a discount off the average or the lowest price to

private plans.

The inability of regulatory indicator variables to accurately measure these

complex regulatory systems is evidenced by results from international price

comparisons. For example, although France, Japan, and Canada all use

both internal and external referencing, weighted price indexes for 1999 and

2005 show Canadian and French prices roughly 30% lower than the

United States, whereas Japan’s prices were higher than the United States

(Danzon & Furukawa, 2003, 2008).

Rather than using regulatory type indicators, we use average prices of

competitor products as the most accurate measure of the net direct effect of

each country’s regulatory system on expected prices for new drugs. These

lagged prices can be treated as exogenous, because prices are constrained by

regulation in most countries and incentives for competitors to cut price as an

entry deterrence strategy are weak or nonexistent (see footnote 5). To test

for spillover effects, we categorize countries as high-price EU countries

(UK, Germany, Netherlands, Sweden), low-price EU countries (France,

PATRICIA M. DANZON AND ANDREW J. EPSTEIN42

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

Spain, Portugal, Italy, Greece), and high-price non-EU countries (US,

Canada, Japan, Switzerland), based on average prices in our dataset and

evidence from earlier price comparisons (Danzon & Furukawa, 2003).

Parallel Trade

Parallel trade between EU states is legal under the Treaty of Rome. In our

data, parallel imports are reported only in the high-price EU countries,

but the exporting country is not reported. Parallel exporting countries are

mainly the low-price EU countries (France, Spain, Italy, Greece, and

Portugal) (Burstall, 1998). Parallel export risk may raise a ﬁrm’s reserva-

tion price for launch in low-price countries within the European Union.

Measurement of parallel trade risk is described below.

THEORETICAL MODEL AND EMPIRICAL

SPECIFICATION

If markets were separable and prices were unregulated, proﬁt-maximizing

ﬁrms would set prices independently for each country and would launch

promptly after registration in all markets where the expected net present

value of revenues exceeds country-speciﬁc ﬁxed launch costs. With price

regulation and potential spillovers, a necessary condition for launch of drug

sin country jis that expected net present value of revenues exceeds country-

speciﬁc ﬁxed costs plus any revenue loss in other countries due to spillovers:

t¼1(½PsjtðPbjt ðR;NbjtÞ;Pgjt ðNgjt Þ;Yj;PsktÞCsjt Qsjt ðQjt;Nbjt ;NgjtÞ

X

kaj

XjktðPsjt ;Rkt;Ikt Þ)ert dt4FjtðHÞð1Þ

The expected price of product s,P

sjt

, is assumed to depend on: P

bjt

, the

average price of competitor brand products, which depends on regulatory

regime Rand on number of brand competitor products in the subclass

(or launch order if ﬁrst-mover advantage is signiﬁcant) N

bjt

; the average

price of competitor generics P

gjt

, which depends on number of generic

competitors N

gjt

; per capita income, Y, which may affect demand and

regulatory stringency; and P

skt

, the price of drug sin countries k,y,KZ0

Effects of Regulation on Drug Launch and Pricing 43

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

that are referenced by j. Expected sales volume Q

sjt

depends on aggregate

sales in the class Q

, and on the number of brand and generic competitors

bjt

and N

gjt

Expected net revenue from launch is decreasing in

¼q(P

)/qP

, the spillover effect of P

on revenues in country k,

which either references to jor derives parallel imports Ifrom j. F

¼F

þF

is total ﬁxed cost of drug launch, including registration cost F

and price

approval cost F

, which may be lower if the launching corporation is home-

based H;Tis the duration of the economic life of the drug indexed by t;and

ris the discount factor.

Eq. (1) yields the ﬁrm’s reservation price for launch in country j.It

is increasing in X

if launch in jmay reduce revenue loss in country k

due to spillovers from country jð@PAsk

sj =@Xjk40Þ.IfFis invariant across

countries and drugs, reservation price is decreasing in expected market

size ð@PAsk

sj =@Qsj o0Þ. More generally, ﬁrms are less likely to launch drugs in

countries with low expected sales, due to low volume or because price

regulation reduces expected prices.

The regulator’s reservation or maximum offer price depends on the

regulatory regime. Under internal referencing, the regulator’s offer depends

directly on prices of substitute products ð@POffer

sj =@Pbj 0Þand evidence of

incremental beneﬁt. If there is diminishing incremental beneﬁt from addi-

tional drugs in a subclass, offer prices are likely to be non-increasing in

launch order (N

bjt1

). Under external referencing, although the regulatory

formula is based on prices in comparator countries, prices of existing

products provide an empirical proxy for achievable price levels under

country j’s referencing formula. Regulatory offers may be related to per

capita income ð@POffer

sj =@Yj0Þ, given the political pressures in most coun-

tries to constrain health spending within budgets that depend on per capita

income. Potential budget impact may lead regulators to offer lower prices

for drugs with relatively large expected volume, Q

, other things equal

ð@POffer

sj =@Qsj 0Þ. If local corporations are favored, @POffer

sj =@Hj0.

If POffer

sj PAsk

sj 0 and a launch price can be agreed within this range,

launch occurs. If not, negotiations may continue and launch may ultimately

occur if either POffer

sj increases, PAsk

sj falls, or some mechanism is negotiated

to bridge the difference, such as a price–volume offset.

In our data, we observe only the launch date and launch price condi-

tional on launch, not the date of registration approval or the ‘‘ask’’ and

‘‘offer’’ prices. We therefore estimate reduced form equations for the launch

hazard and launch price as a function of the determinants of the ﬁrm’s ask

PATRICIA M. DANZON AND ANDREW J. EPSTEIN44

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

price and the regulator’s offer price. The reduced form launch hazard

equation is:

hsjt ¼hP

bjt1;Pgjt1;Qjt1;Xjt1;Nbjt1;Ngjt1;Ijt1;Yj;PsKt1;H



(2)

Because the bargaining range P

Offer

P

Ask

W0 also deﬁnes the range for

the price, conditional on launch, the reduced form launch price equation

includes the same variables as the launch hazard equation:

Psj ¼fP

bj;Nbjt1;Pgj;Qj;Xj;Ij;Yj;PsKt1;H



(3)

Identiﬁcation of structural parameters from the reduced form estimates is

not possible because most variables enter both the ‘‘ask’’ and the ‘‘offer’’

equations and with the same expected sign. In theory, expected price and

market size at global launch t

inﬂuence the decision to seek registration

approval and hence could identify the launch hazard equation, whereas

launch price depends on (lagged) contemporaneous competitor prices. In

practice, both pre-launch values and change over time of the expected price

and quantity variables were insigniﬁcant in the launch equation, after

controlling for contemporaneous values.

Launch Estimation

We estimate the launch hazard using a maximum likelihood discrete-time

implementation of a proportional hazards model based on complemen-

tary log–log regression.

In the clog–log analysis, each drug is eligible for

launch in all countries starting from its quarter of ﬁrst launch in any country

in our sample (‘‘global launch’’), and remained eligible until it launched.

Thus each drug sin each country jcontributes t

observations, the number

of quarters from product s’s global launch through either ﬁrst launch in

country jor the end of our study period.

To account for the intramolecule

clustering across countries, we used robust standard errors or molecule

random effects.

The hazard of launch is h

sjt

, the probability that drug slaunches in

country jin period tconditional on not having previously launched. Using a

clog–log speciﬁcation implies that

hsjt ¼1exp expðlðtÞþbGsjtÞ

(4)

where G

sjt

is a vector of explanatory variables as outlined above. To facilitate

interpretation, we present marginal effects calculated at the regressors’

Effects of Regulation on Drug Launch and Pricing 45

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

mean values. The standard errors of the marginal effects were calculated using

the delta method.

One limitation of the clog–log speciﬁcation is failure to account for

possible time-invariant unobserved characteristics common to a molecule

across countries that inﬂuence launch. Unobserved heterogeneity may lead

to coefﬁcient attenuation and biased estimates of duration dependence

(Heckman & Singer, 1984;Lancaster, 1990). We address this issue by adding

a normally distributed term for the drug-level heterogeneity n

hsjt ¼1exp expðlðtÞþbGsjt þnsÞ



(5)

Split-Population Estimates

The clog–log speciﬁcation embeds the assumption that the probability of

launch goes to unity as time goes to inﬁnity. Some of the molecules in our

sample might not meet drug-approval requirements or would have limited

market potential in some countries, however. We therefore also estimate a

discrete-time split-population model with time-varying covariates (Jenkins,

1995;Schmidt & Witte, 1989), which allows for some empirically estimated

sample-wide proportion of drugs never to launch.

Launch Price

We use ordinary least squares with molecule-clustered standard errors to

model the log of launch price of drug sin country j, conditional on launch-

ing. To account for unobserved molecule characteristics we also report

results from a generalized least squares (GLS) random effects estimator.

DATA AND VARIABLE DEFINITIONS

Data

We use data from IMS Health’s Midas database on drugs in 15 countries

for 12 therapeutic classes, all of which experienced the launch of a new

subclass shortly before or during our study period, 1992–2003.

The data

for each molecule include active ingredient, originator corporation(s) and

marketing companies, pack description, launch date, therapeutic class, and

PATRICIA M. DANZON AND ANDREW J. EPSTEIN46

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

quarterly data on outpatient sales at manufacturer prices (revenue in local

currency) and unit volume (IMS standard units)

from 1Q 1992 through 4Q

2003. After the data were screened for internal consistency, revenue was

adjusted for inﬂation using country-quarter-speciﬁc Producer Price Indexes

available from the International Monetary Fund, with 2003 as the base year,

and converted to US dollars using the average 2003 country-speciﬁc

exchange rate. Price per dose for each drug was calculated on a quarterly

basis as the ratio of total revenues to standard units sold.

Of the 375 molecules in the dataset, 116 are classiﬁed as in new subclasses

(‘‘superior’’) and 259 are classiﬁed as in older subclasses (‘‘inferior’’).

Among the superior molecules, 221 of their potential 1,740 drug–country

launches had occurred prior to our study period; we observe 885 of the

1,519 potential superior drug–country launches during our 12-year study

period.

Variable Deﬁnitions

Regulation/Expected Price

We use the (log-lagged) unweighted average price of competitor brand

(originator and licensed, including parallel imports) products in the same

therapeutic class as a comprehensive measure of the direct effect of price

regulation on expected price for a new drug. The price of close competitor

products is the explicit regulatory benchmark in internal referencing

regimes; it should also be a rough proxy for the net effect of external

referencing formulae. In free-pricing countries, the average price of esta-

blished products provides an estimate of the expected price of a new drug,

assuming competition constrains similar products to have similar prices. We

distinguish between average prices for superior and inferior subclasses, to

test for differential effects within vs. between subclasses, as proxies for

static vs. dynamic price competition in free-pricing countries; in regulated

markets, analogous effects operate via the regulator’s choice of comparator

products.

Launch Order and First-Mover Advantage

Number of brand competitors in the subclass was included to test for ﬁrst-

mover advantage and/or diminishing incremental beneﬁt. It was insignif-

icant and replaced in the launch equation with an indicator variable for

quarters in which there are no molecules in the country-subclass. Indicator

variables for launch order are retained in the launch price equation.

Effects of Regulation on Drug Launch and Pricing 47

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

Expected Sales Volume

The (log-lagged) total number of doses sold in the same therapeutic class as

the new drug is included as a measure of expected volume.

The expected

effect on launch hazard and price is uncertain a priori, depending on whether

the ﬁrm’s opportunity cost of delay dominates the regulator’s concern

over budget impact. It was insigniﬁcant and was dropped from the price

equation to conserve degrees of freedom.

External Referencing Spillovers

To test for indirect effects in low-price countries of regulatory referencing by

high-price countries, we included three count variables that measure the

number of countries in which a molecule has already launched, categorized

by low-price EU countries (France, Italy, Spain, Portugal, and Greece),

high-price EU countries (Germany, the Netherlands, Sweden, and the

United Kingdom) and high-priced non-EU countries (the United States,

Japan, and Canada). Our categorization of low- and high-price EU countries

is supported by actual average prices (see below). These variables are also

interacted with indicators for whether the potential launch is in a low-price

vs. high-price EU country. Although the external referencing hypothesis

predicts that launch delay and price in low-price EU countries will be

positively related to prior launch in high-price EU countries, this could also be

explained by rational launch sequencing, even without external referencing.

An additional prediction of the external referencing hypothesis alone, which

we test, is that the association between launch in a low-priced EU country

and prior launch in a high-priced EU country should be greater than the

association between launch in low-price EU countries and prior launch in

high-price non-EU countries.

In the price equation, we include similar interactions, except that we use

the Minimum Own Price in high-price EU, low-price EU, and high-price

non-EU countries, deﬁned as the lowest price received for the molecule in

any country where launch has already occurred, for each country group,

rather than simple count variables for number of prior launches. Estimates

using Maximum Own Price were similar to those reported here using

Minimum Own Price. Both variables could not be included together due

to collinearity.

We also include a dummy variable Any PI Share in Subclass, to test

whether risk of competition from parallel import reduces the propensity to

launch or reduces launch prices.

Because the IMS data do not identify the

country from which PIs originate, we cannot test directly whether volume

PATRICIA M. DANZON AND ANDREW J. EPSTEIN48

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

of parallel exports reduces launch hazards in exporting countries. Rather,

the propensity to parallel export is subsumed in the country-ﬁxed effects.

Launch Timing

A quadratic in years since global launch is included to control for the decline

in incentives for launch with time since global launch, because patents run

regardless of launch and compounds may undergo obsolescence due to entry

of newer compounds. An indicator for molecules launched before 1990

controls for their relatively old age. An indicator for molecules launched

since 1996 tests for effects of the EMA regime. It is expected to be positive if

the cost-reducing effects of the EMA coordinated registration process

outweighed the increased risk of spillovers. Molecules launched during

1990–1995 are the referent category.

Country of Domicile

Previous studies have found that on average new drugs launch more quickly

in the home country of the originator ﬁrm (Danzon, Wang, & Wang, 2005;

Kyle, 2006, 2007); however, differences across countries, by type of local

ﬁrm, and effects on launch prices have not been examined. Elsewhere

(Danzon & Epstein, 2009), we report results for three categories of local

corporations: Local Originators, Solo Licensees, and Local Co-marketers.

These effects are included in the regression estimates but not discussed in

detail here.

Country and Year Effects

We include country-ﬁxed effects to capture other country-speciﬁc factors

that may affect launch delay and launch prices (controlling for expected

price, volume, and per capita income), in particular, pure bureaucratic delay

and parallel export risk. Germany is excluded as the referent country.

The Dollar–Euro/ECU exchange rate and the PPI are included in the

price equation to control for exchange rate and indexing trends that could

bias our dollar-denominated estimates of competitor prices and launch

prices. Year effects were included in some speciﬁcations, but were generally

insigniﬁcant and are not reported here.

Product Characteristics

The price equation includes product characteristics that affect price per

dose, including pack size, pill strength (grams per unit), and indicator

variables for specialized formulations (oral delayed and non-oral solids),

with oral solids (basic tablets and capsules) as the referent formulation.

Effects of Regulation on Drug Launch and Pricing 49

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

DESCRIPTIVES

There are substantial differences in the number and timing of molecule

launches by country and subclass (Table 1). For the superior subclasses,

Germany and the United States (two free-pricing countries) have the most

molecules ever launched (88 and 86). Sweden, the United States, the Nether-

lands, Germany, and the United Kingdom (all higher price, less-strictly

regulated) also have the shortest median launch delay (17.4–18.7 months).

Japan, Portugal, and France (all strict price-regulated countries) have the

fewest superior molecules (53, 62, and 69 respectively) and the longest mean

launch lags (41, 31, and 37).

Table 2 reports unweighted mean prices by country-subclass-license type,

to provide a rough measure of benchmark competitor prices used by

regulators and ﬁrms in forming price expectations. Note that these means

reﬂect differences in molecules and formulations, in addition to price differ-

ences for identical products, and hence are not valid indexes of cross-

national price differences for a standardized basket of drugs.

These unweighted mean prices show that, for originator/licensee superior

products in the EU, the strictly price-regulated regimes (France, Spain,

Portugal, Greece, and Italy) have relatively low prices. Thus, we classify

them as ‘‘low-price EU markets.’’ The countries with no or less extensive

price/reimbursement regulation (Germany, the United Kingdom, the

Netherlands, and Sweden) have higher mean prices, and we classify them as

‘‘high-price EU markets.’’ The United States has the highest prices, followed

closely by Canada; we also classify Switzerland and Japan as ‘‘non-EU high-

price markets’’ based on other price index comparisons (Danzon &

Furukawa, 2003), although Japan’s unweighted mean prices are quite low

in Table 2. Originator prices are lower in inferior than superior subclasses,

and country rankings are similar but with smaller differentials. Other brand

prices are generally higher than for Unbranded Generics, as expected.

prices generally fall between generic and originator prices, as expected;

however, these differentials are not based on a standardized product mix and

do not provide an accurate measure of originator/PI price differentials.

DETERMINANTS OF LAUNCH

Table 3 reports marginal effects from our basic launch speciﬁcation,

using clog–log estimates with either robust, clustered standard errors, or

molecule-level normal random effects to control for unobserved molecule

PATRICIA M. DANZON AND ANDREW J. EPSTEIN50

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

Table 1. Launch and Molecule Count and Mean and Median Launch Delay by Country and Subclass.

Country Superior Subclasses Inferior Subclasses

Molecules Launches Mean Launch

Delay

Median Launch

Delay

Molecules Launches Mean Launch

Delay

Median Launch

Delay

High-price EU countries

Germany 88 72 18.5 9.5 131 18 30.4 17.5

UK 80 58 18.7 6.5 116 24 69.2 41.5

Netherlands 73 56 18.1 10 112 21 43.7 15

Sweden 77 62 17.4 7 82 19 49.9 18

Low-price EU countries

France 69 53 30.9 29 105 19 87.6 59

Greece 72 55 30.1 22 107 37 116.8 54

Italy 76 61 24.8 21 127 26 74.7 48.5

Portugal 62 48 37.0 33.5 113 26 85.8 67

Spain 76 62 28.1 21 112 23 43.9 31

High-price non-EU countries

Canada 73 62 25.6 16 98 22 91.9 66.5

Japan 53 42 41.0 40 158 43 63.3 28

Switzerland 78 63 23.9 18 111 21 55.5 47

USA 86 72 17.9 8 97 23 86.4 62

Low-price non-EU countries

Brazil 71 60 31.2 20.5 92 40 107.1 90.5

Mexico 72 59 28.6 17 105 28 84.0 55

Note: Launch delays measured in months.

Sample includes all molecules present and new launches occurring in our data during 1992–2003.

Launch delays are calculated only for country launches that occurred during 1992–2003 (regardless of when the global launch occurred).

Effects of Regulation on Drug Launch and Pricing 51

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

Table 2. Mean Price per Dose by Country, Subclass and License Type in 2003.

Country Superior Subclasses Inferior Subclasses

Originator/

Licensee

Unbranded

Generic

Parallel

Import

Other

Brand

Originator/

Licensee

Unbranded

Generic

Parallel

Import

Other

Brand

High-price EU countries

Germany 30.31 0.49 9.23 0.94 2.56 0.32 0.33 0.75

UK 13.26 0.64 1.36 0.61 6.18 0.93 0.46 1.27

Netherlands 35.53 0.05 2.22 7.75 10.79 0.45 0.61 0.72

Sweden 24.41 0.34 1.79 0.45 13.08 0.26 0.42 0.98

Low-price EU countries

France 6.52 0.41 N/A 0.89 0.28 0.39 N/A 0.75

Greece 12.84 1.11 N/A 0.56 6.26 0.45 N/A 0.80

Italy 2.24 0.41 N/A 0.77 0.30 0.31 N/A 0.61

Portugal 1.83 0.44 N/A 0.78 0.33 0.20 N/A 0.34

Spain 6.13 0.33 N/A 1.06 0.28 0.63 N/A 0.69

High-price non-EU countries

Canada 49.30 0.57 N/A 0.60 1.03 0.93 N/A 0.57

Japan 12.27 0.44 N/A 0.73 2.62 1.26 N/A 0.59

Switzerland 34.64 0.81 N/A 0.76 5.81 0.37 N/A 1.43

USA 52.70 0.90 N/A 14.22 11.05 1.43 N/A 12.72

Low-price non-EU countries

Brazil 2.34 0.39 N/A 0.95 4.21 0.21 N/A 0.35

Mexico 18.34 0.69 N/A 1.65 5.84 0.26 N/A 0.53

Note: All prices are ex manufacturer prices per standard unit in 2003 US Dollars.

Sample includes all molecules present in our dataset in 2003.

PATRICIA M. DANZON AND ANDREW J. EPSTEIN52

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

Table 3. Marginal Effects for Launch Model.

Variables Clog–Log With

Clustered SEs

Clog–Log With

Normal REs

Log average price of superior brands

(Lag 1Q)

0.0053



0.0086



[0.0027] [0.0032]

Log average price of inferior brands

(Lag 1Q)

0.0031 0.0053



[0.0029] [0.0031]

Log total volume of all drugs in class

(Lag 1Q)

0.0034 0.0012

[0.0027] [0.0035]

Num generic manufs per molc in superior

subclass (Lag 1Q)

0.0002 0.0005

[0.0002] [0.0004]

Num generic manufs per molc in inferior

subclass (Lag 1Q)

0.0001 0.0001

[0.0001] [0.0001]

No molecules in superior subclass D.V. 0.0135 0.0089

[0.0105] [0.0129]

No molecules in inferior subclass D.V. 0.0210



0.0264

[0.0074] [0.0181]

Time since global launch (years) 0.0291



0.0271



[0.0046] [0.0049]

Time since global launch squared (years) 0.0011



0.0009



[0.0002] [0.0002]

First global launch before 1990 D.V. 0.0002 0.0131

[0.0090] [0.0149]

First global launch in [1996-end] D.V. 0.0023 0.0058

[0.0066] [0.0108]

Num already launched (UK, Germany) 0.0290



0.0301



[0.0071] [0.0072]

Num already launched (Sweden,

Netherlands)

0.0245



0.0239



[0.0052] [0.0059]

Num already launched (Italy, France) 0.0126



0.0185



[0.0055] [0.0058]

Num already launched (Spain, Portugal,

Greece)

0.0031 0.0024

[0.0030] [0.0040]

Num already launched (Canada, Japan,

Switzerland, USA)

0.0089



0.0079



[0.0030] [0.0035]

Any PI share in subclass D.V. 0.0008 0.0041

[0.0072] [0.0092]

Launch by local originator corporation D.V. 0.1286



0.1822



[0.0458] [0.0420]

Launch by solo licensee corporation D.V. 0.0328



0.0424



[0.0130] [0.0157]

Launch by local co-marketer

corporation D.V.

0.0340



0.0239

[0.0145] [0.0160]

USD to (ECU or Euro) exchange rate 0.0044 0.0115

[0.0209] [0.0255]

Effects of Regulation on Drug Launch and Pricing 53

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

Table 3. (Continued )

Variables Clog–Log With

Clustered SEs

Clog–Log With

Normal REs

UK D.V. 0.0101 0.0090

[0.0085] [0.0111]

Netherlands D.V. 0.0276



0.0343



[0.0082] [0.0106]

Sweden D.V. 0.0249



0.0269



[0.0099] [0.0116]

France D.V. 0.0340



0.0452



[0.0078] [0.0110]

Greece D.V. 0.0341



0.0418



[0.0086] [0.0116]

Italy D.V. 0.0296



0.0361



[0.0084] [0.0109]

Portugal D.V. 0.0405



0.0536



[0.0084] [0.0115]

Spain D.V. 0.0261



0.0292



[0.0079] [0.0112]

Canada D.V. 0.0310



0.0389



[0.0084] [0.0108]

Japan D.V. 0.0436



0.0613



[0.0083] [0.0114]

Switzerland D.V. 0.0313



0.0378



[0.0089] [0.0119]

USA D.V. 0.0283



0.0387



[0.0085] [0.0104]

Brazil D.V. 0.0326



0.0395



[0.0084] [0.0111]

Mexico D.V. 0.0345



0.0426



[0.0086] [0.0115]

Constant

Therapeutic class ﬁxed effects included? Yes Yes

Num observations 23,400 23,400

Number of molecule-level clusters 111 111

Model log-likelihood

Mean of dependent variable 0.0378 0.0378

Note: All prices are ex manufacturer prices per standard unit in 2003 US dollars.

Standard errors in brackets.



Signiﬁcant at 10%;



signiﬁcant at 5%;



signiﬁcant at 1%.

D.V., dummy variable; Num, number; manuf, manufacturer; molc, molecule.

PATRICIA M. DANZON AND ANDREW J. EPSTEIN54

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

characteristics. The clog–log and random effects estimates are generally

consistent. Country and class-ﬁxed effects are included.

Direct Effect of Price Regulation

Launch hazards are signiﬁcantly related to mean prices of competitor

brands in the same subclass: for superior products a 10% increase in

competitor prices is associated with a 0.053 percentage point increase in the

launch hazard, which implies an elasticity of 0.0014 at the 3.8% mean

launch hazard for superior molecules per quarter. Thus to the extent that

regulation reduces prices, it reduces incentives to launch. These estimates

may underestimate the magnitude of regulatory effects, if our measure of

mean price, based on all originator and licensed products in the subclass,

includes some that are not used as comparators by regulators. Cross-class

price effects are insigniﬁcant, indicating that regulatory benchmarking

and/or competitive constraints operate primarily within rather than between

subclasses, and that dynamic competition is driven by product character-

istics other than price. Similarly, the effects of number of generic compe-

titors are negative but statistically insigniﬁcant, providing further evidence

that availability of older, cheaper generic substitutes is not a signiﬁcant

deterrent to the launch of new brand products, even in older subclasses

where generics are more numerous, possibly because generic substitution is

mostly within rather than between molecules.

Unit volume for the therapeutic class – whether measured contempor-

aneously, at global launch, as a growth rate, or by subclass – was not

signiﬁcant. This may reﬂect offsetting incentives for regulators to constrain

prices more stringently for compounds with large potential sales, whereas

ﬁrms prefer to launch these drugs more promptly.

Launch Timing and Sequence

Launch hazards appear ﬁrst to decrease then increase with time since global

launch, reaching a minimum at 13.5 years from global launch.

This

average quadratic speciﬁcation reﬂects the diverse launch patterns in

Table 1, which shows median launch lags for superior drugs of less than

one year in the high-price EU countries and the United States, followed

by launch within the second year for all other countries except France,

Portugal, and Japan, where launch typically occurs later.

Effects of Regulation on Drug Launch and Pricing 55

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

The coefﬁcient on the post-1996 global launch indicator is negative but

insigniﬁcant for superior drugs. Taken at face value this suggests that on

average the EMA process has not reduced launch lags, possibly because

price approval is the rate-limiting regulatory hurdle and any cost-reducing

effect of accelerated approval is offset by increased risk of spillovers.

Indirect Regulatory Effects: Cross-National Spillovers

The evidence strongly supports the hypothesis that launch in low-priced EU

countries is adversely affected by the risk of spillover to higher-price EU

countries through external referencing and possibly PI risk. For superior

drugs, the coefﬁcients on number of countries in which launch has already

occurred are positive and signiﬁcant, with the exception of prior launch

in the three lowest-price EU countries, Spain, Portugal, and Greece.

The marginal effect is largest for prior launch in the United Kingdom or

Germany, at 0.030, and is 0.025 for prior launch in Sweden or the

Netherlands. By contrast the marginal effect of prior launch in Italy or

France is 0.013, and for high price non-EU countries the marginal effect

is only 0.009.

This pattern conﬁrms that ﬁrms delay launch in low-price

EU countries until launch has occurred in higher-price EU countries.

Moreover, because Spain, Portugal, and Greece reference the lowest prices

in a group of relatively low-price countries, including France and Italy, a

ﬁrm’s optimal launch strategy plausibly leads to launching last in these three

countries, after higher prices have been established in the countries that

they reference.

To explore further the spillover effects, we estimated a speciﬁcation that

includes counts of prior launches in high-price EU, low-price EU, and other

high-price countries (Canada, Japan, Switzerland, and the United States),

together with interactions between these launch counts and indicators for

whether the current observation is a low- or a high-price EU country.

Marginal effects of these interactions are reported in Table 4. The marginal

effect of a prior launch (from zero to one) in a high-price EU country on

launch in a low-price EU country is 0.0018, whereas the effect of a prior

launch in another low-price EU country is only 0.0005, and the difference is

statistically signiﬁcant. Similarly, the marginal effect on launch in a low-

price EU country is greater from a prior launch in a high-price EU country

than from a launch in a high-price non-EU country, as expected, because

referencing and parallel trade within the EU is only to EU countries. This

evidence thus supports the hypothesis that the observed pattern reﬂects

PATRICIA M. DANZON AND ANDREW J. EPSTEIN56

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

spillovers, rather than other unobserved factor that may be correlated with

launch timing across countries.

Parallel import presence in the class is not associated with launch hazard

in the importing country, after controlling for country-ﬁxed effects. This

may simply reﬂect the high correlation between the PI indicator and the

country indicators for the four importing countries – Germany, Sweden, the

Netherlands, and the United Kingdom – and the high PIs presence across

classes in those countries. It is also more likely that parallel trade risk

leads primarily to non-launch or launch delay in the parallel export coun-

tries, not in the importing countries. We cannot measure this effect in the

exporting countries because our data do not report PI country of origin;

hence, it is subsumed in country effects.

Country-Fixed Effects

For superior drugs, compared to Germany, the referent country, other

country effects are all negative. Marginal effects are smallest for the United

Kingdom (0.011) and other relatively free-pricing countries; marginal

effects are largest for Japan (0.044), probably reﬂecting its unique clinical

Table 4. Marginal Effects of Prior Foreign Launch on Launch Hazard

in Low-Price EU Countries for Superior Subclasses.

Marginal Effects of Prior Foreign Launch in Low-Price EU Countries

Clog–log with robust clustered SEs

Net effect of a single prior launch in a

high-price EU country

0.0018



[0.0004]

Net effect of a single prior launch in a

low-price EU country

0.0005



[0.0002]

Difference 0.0013



[0.0003]

Net effect of a single prior launch in a

high-price EU country

0.0018



[0.0004]

Net effect of a single prior launch in a

high-price non-EU country

0.0006



[0.0002]

Difference 0.0012



[0.0004]

Standard errors in brackets.



Signiﬁcant at 10%;



signiﬁcant at 5%;



signiﬁcant at 1%.

Effects of Regulation on Drug Launch and Pricing 57

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

trial requirements; and the major EU parallel export countries (Spain,

France, Greece, Portugal, and Italy) are all signiﬁcantly negative, as are

several other countries.

Split-Population Estimates

The split-population estimates are generally consistent with the clog–log

estimates (see Danzon & Epstein, 2009). However, the split-population

estimates imply that the probability of never launching is highly signiﬁcant

for 27.7% of inferior molecule–country pairs, compared with only 4.9%

of the superior molecules. This provides further evidence that certain mole-

cules, especially late entrants in inferior classes, are not marketable in certain

countries.

Whether this reﬂects inability to meet regulatory or market

requirements cannot be determined with our data.

DETERMINANTS OF LAUNCH PRICE

Table 5 reports the determinants of (log) launch price using ordinary least

squares (OLS) with robust, clustered standard errors, and GLS random

effects to control for unobserved molecule heterogeneity. Our discussion

focuses on estimates that include gross domestic product (GDP) per capita;

excluding GDP changes primarily the country-ﬁxed effects, as reported

below. Year-ﬁxed effects were also included, to control for any bias in our

inﬂation and exchange rate adjusters, but these coefﬁcients were generally

insigniﬁcant and are not reported. Therapeutic class effects were omitted

because they are highly collinear with competitor prices and order of entry

within class.

Competitor Prices

For both superior and inferior products, launch prices are signiﬁcantly

positively related to prices of competitor products in the same subclass

(elasticity of 0.12 for superiors and 0.17 for inferiors). The cross-subclass

elasticity of inferior drug prices on launch price of new superior drugs is also

positive but smaller (0.08). This conﬁrms the earlier evidence, that launch

hazards are inﬂuenced mainly by prices of established products within

subclass, implying that dynamic competition between subclasses is based on

PATRICIA M. DANZON AND ANDREW J. EPSTEIN58

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

Table 5. Determinants of Launch Prices, OLS, and Normal Random

Effects Regressions.

Variables OLS w/Robust Clustered SEs Normal Random Effects

Superior brands’ price

missing D.V.

0.0615 0.0636 0.0644 0.0685

[0.1879] [0.1886] [0.1134] [0.1138]

Log avg price of superior

brands (Lag 1Q)

0.1574



0.1586



0.1244



0.1283



[0.0394] [0.0395] [0.0201] [0.0201]

Inferior brands’ price missing

D.V.

0.1093 0.114 0.004 0.0157

[0.1523] [0.1541] [0.1118] [0.1121]

Log avg price of inferior

brands (Lag 1Q)

0.0393 0.0387 0.0844



0.0830



[0.0398] [0.0397] [0.0206] [0.0206]

Generics’ price missing D.V. 0.0674 0.0689 0.0095 0.0115

[0.1168] [0.1165] [0.0758] [0.0761]

Log avg price of generics in

class (Lag 1Q)

0.0292 0.0294 0.0169 0.0162

[0.0295] [0.0296] [0.0194] [0.0195]

Time since global launch

(years)

0.0427 0.0413 0.0391 0.0367

[0.0265] [0.0265] [0.0260] [0.0261]

Time since global launch

squared (years)

0.0028 0.0027 0.0005 0.0003

[0.0018] [0.0018] [0.0018] [0.0018]

First brand launch in

ctry-subclass D.V.

0.1998 0.2034 0.2132 0.2164



[0.1656] [0.1656] [0.1307] [0.1312]

Second brand launch in

ctry-subclass D.V.

0.3496



0.3486



0.2819



0.2770



[0.0824] [0.0824] [0.0749] [0.0751]

Third or fourth brand launch

in ctry-subclass D.V.

0.2412



0.2399



0.1859



0.1816



[0.0611] [0.0613] [0.0555] [0.0557]

High-price EU min own

price missing D.V.

0.2170



0.2138



0.0649 0.0631

[0.0821] [0.0821] [0.0575] [0.0578]

Log min own price in hi-price

EU (Lag 1Q)

0.2179



0.2174



0.1000



0.1012



[0.0623] [0.0622] [0.0261] [0.0262]

Low-price EU min own price

missing D.V.

0.0185 0.0161 0.0251 0.03

[0.0524] [0.0527] [0.0483] [0.0485]

Log min own price in

low-price EU (Lag 1Q)

0.0243 0.0234 0.0221 0.0188

[0.0394] [0.0390] [0.0279] [0.0280]

High-price non-EU min own

price missing D.V.

0.1054 0.1076 0.1019



0.1049



[0.0649] [0.0651] [0.0554] [0.0556]

Log min own price in hi-price

non-EU (Lag 1Q)

0.2682



0.2677



0.1433



0.1425



[0.0522] [0.0519] [0.0251] [0.0252]

Any PI share in subclass

D.V.

0.022 0.0232 0.0207 0.0236

[0.0795] [0.0786] [0.0650] [0.0653]

Log GDP per capita 0.862 1.8350



[0.9232] [0.7426]

Launch by local originator

corporation D.V.

0.0311 0.0332 0.0693 0.0739

[0.1009] [0.1011] [0.0694] [0.0696]

Launch by solo licensee

corporation D.V.

0.0742 0.0708 0.0543 0.0452

[0.0770] [0.0768] [0.0623] [0.0625]

Effects of Regulation on Drug Launch and Pricing 59

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

Table 5. (Continued )

Variables OLS w/Robust Clustered SEs Normal Random Effects

Launch by local co-marketer

corporation D.V.

0.0031 0.0038 0.0265 0.0288

[0.0971] [0.0980] [0.0782] [0.0785]

USD to (ECU or Euro)

exchange rate

0.1475 0.1373 0.0265 0.0042

[0.6339] [0.6234] [0.4524] [0.4542]

Country-speciﬁc quarterly

Producer Price Index

0.0089



0.0065 0.0088



0.0038

[0.0044] [0.0039] [0.0044] [0.0039]

Avg pack size (up to 100) 0.0118



0.0118



0.0094



0.0093



[0.0017] [0.0017] [0.0010] [0.0010]

Pack sizeW100 D.V. 1.1996



1.2014



0.9891



0.9930



[0.1880] [0.1888] [0.1114] [0.1118]

Avg pill strength (g) 0.4791



0.4928



0.0577 0.0766

[0.2452] [0.2457] [0.2737] [0.2738]

Form: oral solid delayed

D.V.

0.1014 0.1092 0.1059 0.0926

[0.1994] [0.1988] [0.1823] [0.1829]

Form: injectable D.V. 2.0793



2.0865



1.7654



1.7827



[0.3009] [0.3025] [0.0885] [0.0887]

Form: other 0.0523 0.0551 0.0793 0.0734

[0.1683] [0.1663] [0.1702] [0.1708]

UK D.V. 0.3218



0.2751



0.2829



0.1861



[0.0957] [0.0827] [0.0875] [0.0784]

Netherlands D.V. 0.0707 0.0734 0.0669 0.0711

[0.1034] [0.1022] [0.0798] [0.0802]

Sweden D.V. 0.1746 0.0541 0.3307



0.0743

[0.1684] [0.0859] [0.1335] [0.0844]

France D.V. 0.2055



0.2382



0.1272 0.1956



[0.1117] [0.1098] [0.1003] [0.0969]

Greece D.V. 0.2805 0.3932



1.1945



0.2393



[0.7187] [0.1198] [0.5893] [0.1050]

Italy D.V. 0.19 0.3485



0.0835 0.2521



[0.1922] [0.1031] [0.1664] [0.0970]

Portugal D.V. 0.373 0.3098



1.2153



0.2341



[0.7279] [0.1107] [0.5961] [0.1077]

Spain D.V. 0.1941 0.2498



0.7733



0.1710



[0.4840] [0.0992] [0.3931] [0.0928]

Canada D.V. 0.0439 0.0363 0.0338 0.0197

[0.1078] [0.1054] [0.0923] [0.0925]

Japan D.V. 0.1482 0.5852



0.3752 0.5573



[0.5017] [0.2021] [0.3967] [0.1221]

Switzerland D.V. 0.1417 0.2091



0.6109



0.1389

[0.4039] [0.0911] [0.3162] [0.0888]

United States D.V. 0.1838 0.5272



0.35 0.3833



[0.4014] [0.1347] [0.3125] [0.0987]

Brazil D.V. 1.2852 -0.3136



3.1815



0.2182



[1.6940] [0.1111] [1.3787] [0.0940]

PATRICIA M. DANZON AND ANDREW J. EPSTEIN60

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

non-price product attributes. Generic prices have no signiﬁcant effect on

launch prices of new superior brands, suggesting weak price competition

between new brands and old generics.

Launch Timing and Sequence

Controlling for inﬂation and exchange rates, launch prices for superior

products decline with time elapsed since global launch, at an estimated rate

(from the random effects model) of –3.8% (p¼0.05) per year.

This

suggests that manufacturers cannot obtain a higher price simply by holding

out in launch price negotiations.

For superior products, there is no evidence of ﬁrst-mover advantage in

prices, although second and third entrants do receive higher prices relative

to later entrants in a class.

Cross-National Spillovers

For superior products, launch prices increase with the lowest price previ-

ously received in other high-price EU countries, whereas effects of launch in

Table 5. (Continued )

Variables OLS w/Robust Clustered SEs Normal Random Effects

Mexico D.V. 0.9276 0.2477



2.3375



0.1610



[1.2631] [0.1215] [1.0153] [0.0960]

Constant 7.0696 1.2244 16.8192



0.8265

[8.7738] [0.8798] [7.1791] [0.7247]

Log GDP per capita

included?

Yes No Yes No

Year ﬁxed effects included? Yes Yes Yes Yes

Observations 950 950 950 950

Number of molecule-level

clusters

109 109 109 109

0.89 0.89 0.87 0.87

Mean of dependent variable 0.74 0.74 0.74 0.74

Standard errors in brackets.



Signiﬁcant at 10%;



signiﬁcant at 5%;



signiﬁcant at 1%.

D.V., dummy variable; Num, number; avg, average; ctry, country.

Effects of Regulation on Drug Launch and Pricing 61

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

low-price EU countries is insigniﬁcant. The minimum own price received in

non-EU countries is signiﬁcantly positive for superior molecules, but

insigniﬁcant for inferior molecules, which launch less frequently in high-price

non-EU countries such as the United States and Canada.

We also estimated equations with interactions to test the hypothesis

that spillovers to low-price EU countries are largest from high-price EU

countries. Marginal effects of the interactions are reported in Table 6,

which parallels Table 4 for the launch model in structure and results. We

ﬁnd that the effect on launch price in a low-price EU country is larger for a

10% increase in a drug’s minimum own price in high-price EU countries

than from a 10% increase in minimum own price in other low-price EU

countries. Speciﬁcally, the difference in the marginal effects in the minimum

own price elasticity based on the OLS model is 0.39, and based on the RE

model is 0.26. Similarly, the effect on launch price in a low-price EU country

is substantially larger for a 10% increase in a drug’s minimum own price in

high-price EU countries than it is for a 10% increase in minimum own price

in high-price non-EU countries. The differences in the own price elasticity

are 0.37 from the OLS model and 0.24 for the RE model, consistent with

Table 6. Spillover Effects of Own Price on Launch Price in Low-Price

EU Countries for Superior Subclasses.

Marginal Effects of Log Min Own Price on Launch Price in Low-Price EU Countries

OLS w/Robust

Clustered SEs

Normal

Random Effects

Net effect of log min own price in high-price

EU countries (Lag 1Q)

0.4236



0.2508



[0.0861] [0.0845]

Net effect of log min own price in low-price

EU countries (Lag 1Q)

0.0366 0.0096

[0.0591] [0.0432]

Difference 0.3870



0.2604



[0.1103] [0.1055]

Net effect of log min own price in high-price

EU countries (Lag 1Q)

0.4236



0.2508



[0.0861] [0.0845]

Net effect of log min own price in high-price

non-EU countries (Lag 1Q)

0.0551 0.0096

[0.0837] [0.0778]

Difference 0.3684



0.2411

[0.1641] [0.1580]

Note: All prices are ex manufacturer prices per standard unit in 2003 US dollars.

Standard errors in brackets.



Signiﬁcant at 10%;



signiﬁcant at 5%;



signiﬁcant at 1%.

PATRICIA M. DANZON AND ANDREW J. EPSTEIN62

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

the hypothesis that launching ﬁrst in high-price EU markets can inﬂuence

prices in low-price EU markets. This evidence from launch prices thus

validates that launch delay in low-price markets may ultimately yield higher

prices in these markets through spillovers from higher-price markets, in

addition to avoiding contamination of those high-price markets.

The indicator for PI presence has no signiﬁcant effect on launch prices

of superior drugs. This is consistent with theory and prior evidence, that PI

supplies are limited and middlemen capture most of any price differential.

Controls

The US Dollar per Euro/ECU exchange rate, which declined from a high

of 1.38 in 1992 to a low of 0.83 in 2000, is insigniﬁcant for superior pro-

ducts but large and signiﬁcant for inferior products. This suggests that our

exchange rate adjustments accurately tracked manufacturer’s pricing

adjustments for the more broadly diffused superior products, which are

at higher risk of referencing and parallel trade. By contrast, our exchange

rate adjustment apparently over-adjusted for launch pricing for inferior

products, which were less likely to launch in the United States and therefore

plausibly, were priced independently of the USD/Euro exchange rate.

For superior drugs, the country-speciﬁc Producer Price Index (our price

deﬂator) is signiﬁcantly negative in the OLS speciﬁcation, suggesting that

launch prices of drugs on average have not kept pace with economy-wide

inﬂation, but this conclusion is tentative because the RE estimate is

insigniﬁcant.

Country-Fixed Effects

The country-ﬁxed effects in Table 5 are dramatically different depending on

whether GDP per capita is included. Without GDP controls, prices for

superior molecules in the United States and Japan are signiﬁcantly higher

than in Germany; prices in Switzerland and Canada are higher but insigniﬁ-

cant, whereas prices in Brazil, Mexico, and all other EU countries are

lower, except the Netherlands and Sweden that are similar to Germany.

However, after controlling for GDP per capita, Brazil, Mexico, Spain,

Portugal, and Greece have prices signiﬁcantly higher, and Switzerland,

the United Kingdom, and Sweden have prices signiﬁcantly lower than

Germany, based on the RE estimates.

Effects of Regulation on Drug Launch and Pricing 63

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

Although these are not pure hedonic country effects, taken at face value

they imply that the prevailing spread in drug prices across EU countries is

compressed relative to the counterfactual of differentials based on per

capita income. This may explain at least in part why Spain, Portugal, and

Greece, the lowest-price EU countries in our sample, adopt more stringent

price controls than the higher-price northern countries. However, these

low-price EU countries appear to be constrained in their ability to keep

price differentials in line with income differentials, in part due to external

referencing by and parallel importing to higher-income countries. The

evidence here conﬁrms that the resulting interconnectedness across coun-

tries contributes to higher prices – in addition to delay or non-launch – of

new drugs in these low-price EU countries, as ﬁrms seek to avoid spillovers

to prices in higher-price EU countries.

CONCLUSIONS

Our ﬁndings demonstrate price regulation has both direct (country-speciﬁc)

and indirect (spillover) effects on launch timing and launch prices of new

drugs. Launch timing and prices are signiﬁcantly related to prices of

established products in the same subclass, our measure of price regulation.

Thus, to the extent price regulation reduces price levels, it contributes

directly to longer average launch delays observed in countries that adopt

such regulation. Our estimates suggest that the magnitude of these direct

effects is quite small, although downward bias due to measurement error is

possible. Welfare conclusions on the direct effects of price regulation are

ambiguous, assuming that regulators weigh the beneﬁts of lower prices

against any welfare loss from reduced access to new drugs for their citizens.

We also ﬁnd signiﬁcant evidence of indirect or spillover effects of the

external referencing form of price regulation. Speciﬁcally, regulatory refer-

encing by high-price countries to lower-price countries within the EU creates

incentives for manufacturers to delay launch in low-price countries until

higher prices have been established in other countries that might reference

these low-price countries. Consistent with such strategies, we show that

launch in higher-price EU countries is associated with increased launch

hazard in lower-price EU countries, and launch prices in low-price

EU countries are directly related to prior launch prices in high-price EU

markets. This evidence, that spillover effects are greater from high-price

EU countries to low-price EU countries, than from high-price non-EU

countries, in both launch and price models, supports the hypothesis that

they are attributable to referencing and possibly parallel trade, not to other

PATRICIA M. DANZON AND ANDREW J. EPSTEIN64

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

unmeasured factors that may lead to closely sequenced launches across

countries. To the extent that referencing or parallel importation by higher-

price countries leads to longer launch delay or higher prices in lower-

price countries than would otherwise occur, a welfare loss is imposed on

low-priced countries by the higher-price countries that adopt these regula-

tory strategies.

Although the lower-income EU countries regulate their drug prices,

Spain, Portugal, and Greece nevertheless had relatively high drug prices

given their income, whereas high-price EU countries had lower drug

prices relative to their per capita GDP. Both the theory and evidence

presented here suggest that external referencing has contributed to this

convergence of pharmaceutical prices relative to GDP among EU countries.

Parallel trade effects are expected to be smaller, and empirical estimation

is problematic because our data do not identify exported volumes or source

countries. With that caveat, the evidence on launch timing suggests that

country-ﬁxed effects are negative for countries that are signiﬁcant parallel

exporters.

We ﬁnd no evidence of ﬁrst-mover pricing advantage, and no evidence

that manufacturers can raise prices simply by delaying launch. Although

prices of new drugs are constrained by prices of established drugs, due to

either regulation or competition, such effects occur primarily within

subclasses, with weaker constraints from older, cheaper drugs.

This evidence, that policies of external referencing (and, with weaker

evidence, parallel importing) impose an external cost on the referenced or

exporting countries, is based on the European Union, where such policies

already exist. However, it has important implications for the US debate

over drug price controls through external referencing and drug importation.

The theory above indicates that the launch lag externality will be greater

if referenced countries are small and low price, compared to a much larger,

higher-price referencing or importing country. Because the United States

has both higher brand prices and much higher total volume than most

potential reference or exporting countries, the impact on these countries

if the United States were to adopt referencing or drug importation

would potentially be much larger than the European Union effects docu-

mented here.

NOTES

1. Pharmaceutical R&D takes on average 8–12 years and costs roughly $802m

(in 2001 US dollars) per new compound approved in the United States (DiMasi,

Hansen, & Grabowski, 2003).

Effects of Regulation on Drug Launch and Pricing 65

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

2. Negative external effects may accrue to other countries if the regulated prices

result in suboptimal contribution to joint costs of R&D, but these effects are small

for countries that account for a small share of global sales.

3. Parallel trade is expected to have less effect on launch delay in low-price

countries than external referencing because (a) parallel trade only reduces prices

for the traded fraction of sales, (b) manufacturers may be able to limit parallel

trade supply restrictions to exporting countries, two-tier price structures, etc., and

post-launch price differentials.

4. Delay could simply reﬂect launch budget constraints faced by ﬁrms, in which

case a rational strategy may be to launch ﬁrst in the most proﬁtable markets and use

the revenues generated to cover launch in less proﬁtable markets.

5. We treat the lagged average price of competitor products as exogenous because

prices are regulated in all but three countries in our sample; the exceptions are the

United States, Germany (except classes under reference pricing and time periods

with price freezes), and the United Kingdom (except price increases are not

permitted). Although ﬁrms could in theory cut price below the regulated price, this

would not be a rational entry deterrence strategy in pharmaceutical markets for

two reasons. First, subsequent price increases would be barred by regulation, and,

second, patients have little incentive to be price-sensitive in most markets, because

co-payments are generally independent of drug price. We conﬁrmed the assumption

that prices are sticky and are unrelated to anticipated launch of competitors by

comparing several measures of price change prior to launch with price change in

other periods.

6. Estimates for the old subclasses (reported in Danzon & Epstein, 2009) show

that late-launching drugs in older subclasses have more limited launch success

than new drugs in new subclasses, consistent with non-price dynamic competition.

Launch by local corporations accelerates launch timing but does not affect launch

prices. These effects are conﬁned to a few, regulated countries, suggesting that local

ﬁrms beneﬁt from political inﬂuence rather than real experience advantages.

7. Regulating US prices based on external referencing to foreign prices has been

proposed but so far not enacted. Drug importation has been approved but so far is

not implemented, because the Secretary of Health and Human Services has been

unwilling to certify that safety and cost-reduction requirements would be met.

8. Under the mutual recognition procedure, a manufacturer selects one rappor-

teur country to review the application; other countries have 90 days in which to

challenge the approval, otherwise it takes effect automatically.

9. Price approval is generally not required if the drug is launched without

reimbursement, but such unreimbursed launch is rare, except for ‘‘lifestyle’’ drugs.

10. Internal referencing may involve informal negotiations between the manu-

facturer and the regulator, as in France, or a more mechanistic reference price (RP)

reimbursement system as in the Netherlands. RP systems may be (a) therapeutic RP,

which classify drugs based on mechanism of action and indication, regardless of

patent status, and (b) generic RP, which group drugs by molecule, and hence usually

apply only to off-patent drugs. All drugs in an RP group receive the same reim-

bursement (the ‘‘reference price’’). A manufacturer may in theory charge a higher

price, but the patient must pay any excess over the RP.

PATRICIA M. DANZON AND ANDREW J. EPSTEIN66

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

11. Many other countries, including the United Kingdom, Sweden, Italy,

Germany, and Spain, and most US health plans used a form of generic RP

reimbursement for off-patent compounds for at least part of our period. However,

because these generic RP systems apply only to off-patent drugs, they are unlikely to

affect launch decisions for new drugs, given our evidence of no signiﬁcant effect of

prices of generics or older brands on prices of new brand drugs.

12. As of 1998, the EU countries that used external reference pricing include

Denmark (since April 1997, up to 10 EU countries excluding Greece and Italy),

Greece (lowest in the EU), Ireland (lower of United Kingdom or the average in

Denmark, France, Germany, the Netherlands, and the United Kingdom), Italy

(average of up to 12 EU countries, must be on market for 4 countries and at least 2

with direct price controls), the Netherlands (since June 1996, average price in

Belgium, France, Germany, and the United Kingdom sets the maximum internal

reference price for drugs under internal RP reimbursement), and Portugal (lowest in

France, Italy, and Spain) (Burstall, 1998). The speciﬁcs of these explicit referencing

schemes have changed over time. For example, France subsequently included

the United Kingdom and Germany in its referencing basket used as one factor in

setting launch prices. Informal comparisons are also common, for example, the

quinquennial revisions of the UK’s Pharmaceutical Price Regulatory Scheme (PPRS)

frequently compared UK prices to prices in other EU countries to determine across

the board cuts of UK prices.

13. Germany adopted an internal reference price reimbursement system in 1989

but excluded new on-patent drugs, which could be launched and reimbursed without

price approval, until 2005. The United Kingdom permits free pricing of individual

drugs, subject to a rate of return constraint on each ﬁrm’s portfolio of drugs. The

renegotiation of this proﬁt-control scheme every ﬁve years involves international

price comparisons and required cuts in UK prices that are often inﬂuenced by the

price comparisons. Since 1999, the National Institute of Clinical Excellence (NICE)

has reviewed cost-effectiveness as a condition of reimbursement for most new drugs,

which could slow or prevent new drug launch.

14. Footnote 5 explains why we treat average price of branded competitor

products P

bjt

as exogenous. We treat number of generic competitors in the class as

exogenous because generic entry takes several years to plan and obtain regulatory

approval, and entry incentives depend primarily on sales of older molecules facing

patent expiry.

15. For example, France applies spending limits and sets a target volume as well

as regulated price.

16. Clog–log is preferred to logit because the former assumes a continuous launch

decision process (Allison, 1995).

17. Thus, although the ﬁrst-stage results report thousands more observations than

the second-stage results, the number of drug–country launches is the same in the ﬁrst

and second stages.

18. We calculated marginal effects for interactions of dummy variables X

and X

as {[F(Xb|X

¼1, X

¼1) F(Xb|X

¼0, X

¼1)] [F(Xb|X

¼1, X

¼0) F(Xb|

¼0, X

¼0)]}, where Fis the inverse of the clog–log function, 1 exp[exp(Xb)].

Other regressors were held at their means or at zero if they were related to X

or X

Standard errors were calculated using the delta method.

Effects of Regulation on Drug Launch and Pricing 67

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

19. To account for possible selection bias due to correlation between the

propensity to launch and the launch price, we also estimated both a two-stage

Heckman selection correction model that includes as a regressor the inverse Mills’

ratio from a clog–log ﬁrst-stage hazard equation, and a traditional full information

maximum likelihood Heckman model based on a probit ﬁrst-stage equation. The

coefﬁcients on the inverse Mills’ ratios are signiﬁcant only for the newer subclasses

(‘‘superior’’ drugs) and only in the two-stage models. The ﬁrst-stage launch equation

is identiﬁed primarily off functional form. These unconditional results are generally

similar to the conditional estimates, which are reported here. See Danzon and

Epstein (2009) for more detail.

20. Our 15 countries include all 5 large EU markets (United Kingdom, Germany,

France, Spain, and Italy) and a subset of both high- and low-price smaller EU

countries (Sweden, Netherlands, Portugal, and Greece). Omitting the other six small

EU countries (Austria, Belgium, Denmark, Finland, Ireland, and Luxembourg)

should not lead to biased estimates given their small size and that they are not

expected to be systematically different from the included countries.

21. The IMS standard unit is a proxy for a dose for each formulation, for

example, one tablet or capsule, 5 ml for liquids. The IMS price data for the United

States do not reﬂect off-invoice discounts given by manufacturers to health plans and

hence are upward biased for manufacturer net revenues.

22. We combined multiple form-3 level formulations (e.g., tablets and capsules,

possibly of different strength) in a given country and quarter into a single observation

and deﬁned the price as the volume-weighted average price per unit. Identical forms

that were launched by different co-marketing companies were also averaged.

23. Classiﬁcation of molecules as superior or inferior was based mainly on

mechanism of action, with input from several physicians and review of articles in

PubMed (see Berndt, Danzon, & Kruse, 2007). No value judgment is intended by

use of ‘‘superior’’ and ‘‘inferior.’’ Our analysis includes only the ﬁrst launch of each

compound. We exclude follow-on indications or formulations because they typically

face fewer registration requirements and pricing would be based on the price of

existing formulations. We also focus on launch by an originator or licensee

corporation, excluding the few cases of prior launch by copycat products. Five

superior molecules and 20 inferior molecules were diffused to all our countries prior to

our period. These are included as competitor products but are not in the sample

of potential launches.

24. However, we observe 946 country–molecule–product launches, because of

simultaneous launches of multiple distinct formulations (form 2-level, such as an oral

solid and a liquid).

25. Measures of volume by subclass and number of competitors were not

signiﬁcant and were dropped.

26. We estimated speciﬁcations of the launch and price models that measured

spillovers with either the number of countries previously launched in or minimum/

maximum own prices. The speciﬁcations reported here yielded the better model ﬁt

and conceptual ﬁt with the dependent variables.

27. We estimated a model speciﬁcation that included the average price of parallel

imports, and it was not signiﬁcant.

28. Local Originator identiﬁes a molecule’s originator corporation launching in its

country of domicile; Solo Licensee identiﬁes a locally domiciled, licensed corporation

PATRICIA M. DANZON AND ANDREW J. EPSTEIN68

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

that launched the molecule in at least one country by itself; and Local Co-marketer

identiﬁes a locally domiciled, licensee corporation that launched together with

another ﬁrm in its home country and did not launch alone in any country. These

categories are not mutually exclusive; for example, a molecule with a small Local

Originator could also have a Solo Licensee or a Local Co-marketer as a marketing

partner, all from the same country.

29. Danzon and Furukawa (2003, 2008) report weighted price indexes, based on

standardized market baskets, for originator and generic products in 1999 and 2005.

30. Some of the PI, Unbranded Generic, and Other Brand means include very few

products. For unbranded generics, the relatively high mean US price is surprising

and reﬂects its product mix, whereas volume-weighted price indexes for standardized

products show low generic prices in the United States (Danzon & Furukawa, 2003).

31. For the old subclasses of drugs, which experienced a few late launches during

our period, the minimum is 49.0 years. The minimums from the split-population

estimates are 10.9 years for new and 50.3 years for old subclasses.

32. The p-values for Wald tests comparing these marginal effects are as follows:

UK/Germany vs. Italy/France, p¼0.019; UK/Germany vs. high-price non-EU

countries, p¼0.009; Sweden/Netherlands vs. Italy/France, p¼0.062; Sweden/

Netherlands vs. high-price non-EU countries, p¼0.003.

33. We included the average price of PIs in earlier speciﬁcations, but it was not

signiﬁcant.

34. The estimated percent never launching is signiﬁcantly higher once we control

for Local Originator Corporation, suggesting that some countries approve locally

originated drugs that have limited diffusion potential elsewhere.

35. The net impact of time since global launch (a quadratic) was calculated as

btsgl þ2btsgl2

t, where the mean value of time is 2.16 years.

36. If manufacturers priced inferior drugs based on local prices that were ﬂat or

falling in ECUs/Euros, the prices would be rising over time in USD, due to USD

devaluation, thus explaining the positive and signiﬁcant coefﬁcient on the exchange

rate variable.

37. Product characteristics have expected effects. Price per unit is signiﬁcantly

negatively related to pack size, particularly for pack size over 100 units, possibly

indicating economies of scale in packaging and/or the competitive use of large packs

to give discounts to pharmacies in countries such as the United States, that permit

pharmacists to dispense from large packs. Price is unrelated to strength (grams of

active ingredient per unit) in the RE estimates, suggesting that the positive coefﬁcient

in the OLS results for superiors reﬂects between- rather than within-molecule effects.

Compared to the oral solid formulations (the omitted category), price per dose is

signiﬁcantly higher for injectable and non-oral forms (liquids, creams, etc.).

ACKNOWLEDGMENTS

This research was supported in part by a grant from Wyeth Pharmaceu-

ticals and by support from the Merck Foundation to the University of

Effects of Regulation on Drug Launch and Pricing 69

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

Pennsylvania. The opinions expressed are those of the authors and do not

necessarily reﬂect the opinions of the research sponsors.

REFERENCES

Allison, P. D. (1995). Survival analysis using the SAS system: A practical guide. Cary, NC: SAS

Institute.

Berndt, E., Danzon, P. M., & Kruse, G. B. (2007). Dynamic competition in pharmaceuticals:

Cross-national evidence from new drug diffusion. Managerial and Decision Economics,

28(4-5), 231–250. doi:10.1002/mde.1338

Burstall, M. L. (1998). Pricing and reimbursement in Western Europe 1998: A concise guide.

A Pharma Pricing Review Report. Surrey, UK: PPR Communications Ltd.

Danzon, P. M., & Epstein, A. J. (2009). Effects of regulation on drug launch and pricing in

interdependent markets. NBER Working Paper 10536. Retrieved from http://www.nber.

org/papers/w14041

Danzon, P. M., & Furukawa, M. F. (2003). Prices and availability of pharmaceuticals: Evidence

from nine countries. Health Affairs Web Exclusive, October 2003. Retrieved from http://

content.healthaffairs.org/cgi/content/full/hlthaff.w3.521v1/DC1. doi:10.1377/hlthaff.

w3.521

Danzon, P. M., & Furukawa, M. F. (2008). International prices and availability of

pharmaceuticals in 2005. Health Affairs,27(1), 221–233. doi:10.1377/hlthaff.27.1.221

Danzon, P. M., Wang, Y. R., & Wang, L. (2005). The impact of price regulation on the launch

delay of new drugs: A study of twenty-ﬁve major markets in 1990s. Health Economics,

14(3), 269–292. doi:10.1002/hec.931

DiMasi, J. A., Hansen, R. W., & Grabowski, H. G. (2003). The price of innovation:

New estimates of drug development costs. Journal of Health Economics,22(2),

151–185. doi:10.1016/S0167-6296(02)00126-1

Grabowski, H. G., Vernon, J. M., & Thomas, L. G. (1978). Estimating effects of regulation on

innovation – International comparative analysis of pharmaceutical industry. Journal of

Law & Economics,21(1), 133–163. doi:10.2307/725144

Grossman, G. M., & Lai, E. C. (2006). Parallel imports and price controls. NBER Working

Paper 12423. Retrieved from http://www.nber.org/papers/w12423

Heckman, J., & Singer, B. (1984). A method for minimizing the impact of distributional

assumptions in econometric models of duration. Econometrica,52, 271–320. doi:10.2307/

1911491

Jenkins, S. P. (1995). Easy estimation methods for discrete-time duration models. Oxford

Bulletin of Economics and Statistics,57(1), 129–138. doi:10.1111/j.1468-0084.1995.tb00031.x

Kanavos, P., & Costa-Font, J. (2005). Pharmaceutical parallel trade in Europe: Stakeholder

and competition effects. Economic Policy,44, 751–798. doi:10.1111/j.1468-0327.2005.

00150.x

Kyle, M. K. (2006). The role of ﬁrm characteristics in pharmaceutical product launches. RAND

Journal of Economics,37(3), 602–618. doi:10.1111/j.1756-2171.2006.tb00033.x

Kyle, M. K. (2007). Pharmaceutical price controls and entry strategies. Review of Economics and

Statistics,89(1), 88–99. doi:10.1162/rest.89.1.88

Lancaster, T. (1990). The econometric analysis of transition data. Cambridge, England:

Cambridge University Press. doi:10.1017/CCOL0521265967

PATRICIA M. DANZON AND ANDREW J. EPSTEIN70

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

Lanjouw, J. O. (2005). Patents, price controls and access to new drugs: How policy affects global

market entry. NBER Working Paper 11321. http://www.nber.org/papers/w11321

Peltzman, S. (1973). Evaluation of consumer protection legislation – 1962 drug amendments.

Journal of Political Economy,81(5), 1049–1091. doi:10.1086/260107

Schmidt, P., & Witte, A. D. (1989). Predicting criminal recidivism using ‘split population’

survival time models. Journal of Econometrics,40, 141–159. doi:10.1016/0304-4076(89)

90034-1

Effects of Regulation on Drug Launch and Pricing 71

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

The Economics of Medical Technology, Emerald Publishing Limited, 2012. ProQuest Ebook Central,

http://ebookcentral.proquest.com/lib/upenn-ebooks/detail.action?docID=1019357.

Created from upenn-ebooks on 2018-12-03 11:30:40.

The impact of proposed price regulations on new patented medicine launches in Canada: a retrospective cohort study

Article

Full-text available

May 2024
CAN MED ASSOC J

Background: The Patented Medicine Prices Review Board (PMPRB), the agency that regulates the prices of patented medicines in Canada, published proposed amendments to the regulatory framework in December 2017. Because of a series of changes and delays, the revised policy has not yet been finalized. We sought to evaluate the potential early impact of the uncertainty about the PMPRB policy on patented-medicine launches. Methods: We developed a retrospective cohort of patented medicines (molecules) sold in Canada and the 13 countries that the PMPRB currently uses or has proposed to use as price comparators, from sales data from the IQVIA MIDAS database for 2012-2021. The outcome was whether a molecule was launched (i.e., sold) in a specific country within 2 years of its global first launch (2-yr launch). We compared the change of 2-year launch before (2012-2017) and after the proposed amendments were published ("uncertain period," 2018-2021) in Canada with the change in the United States and the other 12 countries as a group ("other-countries group"), using interrupted time series and logistic regressions, respectively. We further conducted analyses for each individual country and subgroups by molecule characteristics, such as therapeutic benefit, separately. Results: We included 242 and 107 new molecules launched before publication of the proposed amendments and during the uncertain period, respectively. The corresponding 2-year launch proportions were 45.0% and 30.8% in Canada, 81.4% and 82.2% in the US, and 83.9% and 70.1% in the other-countries group. All analyses showed changes in 2-year launch during the uncertain period in the US and in the other-countries group that were similar to the changes in Canada. Greater decreases were observed in Norway and Sweden than in Canada. The 2-year launch proportion for molecules with major therapeutic benefit decreased from 45.8% to 31.3% in Canada during the uncertain period and from 87.5% to 62.5% in the other-countries group, but increased from 91.7% to 100% in the US. Interpretation: No negative impact of the PMPRB-policy uncertainty on molecule launches was observed when comparing Canada with price-comparator countries, except for molecules with major therapeutic benefit. The reduction in launches of medicines with major therapeutic benefit in Canada requires continuing investigation.

The impact of changing the reference countries on the list prices for patented medicines in Canada: A policy analysis

Article

Apr 2024
HEALTH POLICY

Pricing, Procurement and Reimbursement Policies for Incentivizing Market Entry of Novel Antibiotics and Diagnostics: Learnings from 10 Countries Globally

Article

Jun 2024
Appl Health Econ Health Pol

Fostering market entry of novel antibiotics and enhanced use of diagnostics to improve the quality of antibiotic prescribing are avenues to tackle antimicrobial resistance (AMR), which is a major public health threat. Pricing, procurement and reimbursement policies may work as AMR ‘pull incentives’ to support these objectives. This paper studies pull incentives in pricing, procurement and reimbursement policies (e.g., additions to, modifications of, and exemptions from standard policies) for novel antibiotics, diagnostics and health products with a similar profile in 10 study countries. It also explores whether incentives for non-AMR health products could be transferred to AMR health products. This research included a review of policies in 10 G20 countries based on literature and unpublished documents, and the production of country fact sheets that were validated by country experts. Initial research was conducted in 2020 and updated in 2023. Identified pull incentives in pricing policies include free pricing, higher prices at launch and price increases over time, managed-entry agreements, and waiving or reducing mandatory discounts. Incentives in procurement comprise value-based procurement, pooled procurement and models that delink prices from volumes (subscription-based schemes), whereas incentives in reimbursement include lower evidence requirements for inclusion in the reimbursement scheme, accelerated reimbursement processes, separate budgets that offer add-on funding, and adapted prescribing conditions. While a few pull incentives have been piloted or implemented for antibiotics in recent years, these mechanisms have been mainly used to incentivize launch of certain non-AMR health products, such as orphan medicines. Given similarities in their product characteristics, transferability of some of these pull incentives appears to be possible; however, it would be essential to conduct impact assessments of these incentives. Trade-offs between incentives to foster market entry and thus potentially improve access and the financial sustainability for payers need to be addressed.

Timely, Cheap, or Risk-Free? The Effect of Regulation on the Price and Availability of New Drugs

Article

Full-text available

Mar 2024

The high level of regulation of innovative drugs on the market, which is necessary to protect consumers, produces important effects on drug availability and innovation. In public healthcare systems, the need to curb prices comes from expenditure considerations. The aim of price regulation is to obtain a more equitable allocation of the value of an innovative drug between industries and patients (by reducing prices to make drugs more affordable), but it may also reduce access. (In the listing process, the industry may find it more convenient to limit commercialisation to profitable subgroups of patients.) Furthermore, with the advent of personalised medicine, there is another important dimension that has to be considered, namely, incentives to invest in drug personalisation. In this paper, we review and discuss the impact of different pricing rules on the expenditure and availability of new drugs.

Harmonizing regulatory market approval of products with high safety requirements: Evidence from the European pharmaceutical market

Article

Full-text available

Mar 2024
HEALTH ECON

We causally analyzed whether being a member of the European Union (EU) and having access to a centralized marketing authorization procedure (centralized procedure [CP]) affects availability and time to launch of new pharmaceuticals. We employed multiple difference‐in‐differences models, exploiting the eastern enlargement of the EU as well as changes in the indications that fall within the compulsory or voluntary scope of the CP. Results showed that countries experienced a mean decrease in launch delay of 10.9 months (p = 0.004) after joining the EU. Effects were higher among pharmaceuticals that belong to indications that might voluntarily participate in the CP but are not obliged to. These are often financially less attractive to manufacturers than pharmaceuticals within the compulsory scope. Availability of new pharmaceuticals launched remained unaffected. We found signs that the magnitude of the country‐specific effect of centralized marketing authorization on launch delay may be influenced by strategic decisions of manufacturers at the national level (e.g., parallel trade or reference pricing).

The design of external reference pricing schemes and the choice of reference countries and pricing rules

Article

Nov 2024

Laura Birg

External reference pricing imposes a price cap for drugs based on prices in other countries. I study the choice of reference countries and pricing rules in a three‐country framework. If the manufacturer sells to all three countries, the minimum price rule yields the lowest drug price. If the referencing country is sufficiently large, the manufacturer may not export to reference countries under the minimum price rule. External reference pricing creates the incentive for the reference countries also to adopt external reference pricing. Thus, external reference pricing results in price convergence.

Low- And Middle-Income Countries Experienced Delays Accessing New Essential Medicines, 1982-2024

Article

Oct 2024

Little is known about how long it takes for new medicines to reach countries with different income levels. We analyzed data, sourced from IQVIA, on the timing of new drug launches in seventy-five low-, middle-, and high-income markets from 1982 to 2024. The sample captured the majority of essential medicines (as designated by the World Health Organization in the twenty-third Model List of Essential Medicines) that first came into medical use anywhere globally from 1982 onward. Kaplan-Meier estimates were used to quantify delays in launches across countries. Our analysis comprised 119 medicines with 6,871 observed launches. Nearly three-quarters (74 percent) of first launches occurred in just eight countries (in order of the most first launches, the US, the Netherlands, Sweden, Switzerland, the United Kingdom, France, Germany, and Japan). From the first launch globally, the median time to availability was 2.7 years for high-income countries, 4.5 years for upper-middle-income countries, 6.9 years for lower-middle-income countries, and 8.0 years for low-income countries. The gap between richer (high- and upper-middle-income) and poorer (lower-middle- and low-income) countries remained largely unchanged over time. Strategies to address the disparities highlighted by this analysis are urgently needed.

Comparing New Prescription Drug Availability and Launch Timing in the United States and Other OECD Countries

Article

Jun 2024

Andrew W Mulcahy

The price of cost-effectiveness thresholds under therapeutic competition in pharmaceutical markets

Article

Jun 2023
J HEALTH ECON

Health systems around world are increasingly adopting cost-effectiveness (CE) analysis to inform decisions about access and reimbursement. We study how CE thresholds imposed by a health plan for granting reimbursement affect drug producers' pricing incentives and patients' access to new drugs. Analysing a sequential pricing game between an incumbent drug producer and a potential entrant with a new drug, we show that CE thresholds may have adverse effects for payers and patients. A stricter CE threshold may induce the incumbent to switch pricing strategy from entry accommodation to entry deterrence, limiting patients' access to the new drug. Otherwise, irrespective of whether entry is deterred or accommodated, a stricter CE threshold is never pro-competitive and may in fact facilitate a collusive outcome with higher prices of both drugs. Compared to a laissez-faire policy, the use of CE thresholds when an incumbent monopolist is challenged by therapeutic substitutes can only increase the surplus of a health plan if it leads to entry deterrence. In this case the price reduction by the incumbent necessary to deter entry outweighs the health loss to patients who do not get access to the new drug.

Reference Pricing as a Deterrent to Entry: Evidence from the European Pharmaceutical Market

Article

May 2023

External reference pricing (ERP), the practice of benchmarking domestic drug prices to foreign prices, generates an incentive for firms to withhold products from low-income countries. Using a novel moment inequality approach, we estimate a structural model to measure how ERP policies affect access to innovative drugs across Europe. We find that ERP increases entry delays in eight low-income European countries by up to one year per drug. The European Union could remove these delays without replacing ERP by compensating firms through lump-sum transfers at the cost of around €18 million per drug. (JEL L13, L51, L65)

Parallel Imports and Price Controls

Article

Full-text available

Jun 2008
RAND J ECON

Price controls create opportunities for international arbitrage. Many have argued that such arbitrage, if tolerated, will undermine intellectual property rights and dull the incentives for investment in research-intensive industries such as pharmaceuticals. We challenge this orthodox view and show, to the contrary, that the pace of innovation often is faster in a world with international exhaustion of intellectual property rights than in one with national exhaustion. The key to our conclusion is to recognize that governments will make different choices of price controls when parallel imports are allowed by their trade partners than they will when they are not.

Patents, Price Controls and Access to New Drugs: How Policy Affects Global Market Entry

Article

Jan 2005

Jean O. Lanjouw

Easy estimation methods for discrete-time duration models

Article

Jan 1995

S.P. Jenkins

A method for minimizing the impact of distributional assumptions in econometric models for duration data

Article

Jan 1984
ECONOMETRICA

The Econometric Analysis of Transition Data

Article

Nov 1992

Survival Analysis Using the SAS System: A Practical Guide

Article

Jan 1995

Paul D Allison

Sample Selection Bias Specification Error

Article

Jan 1979
ECONOMETRICA

James J. Heckman

LIMDEP: User''s Manual and Reference Guide

Article

Jan 1992

William H Greene

Easy estimation for discrete-time duration models

Article

Jan 1995

Jenkins Stephen

The Role of Firm Characteristics in Pharmaceutical Product Launches

Article

Sep 2006
RAND J ECON

Margaret K. Kyle

I examine the determinants of new pharmaceutical launches since 1980 in G7 nations. Both market and firm characteristics, and their interaction, are important in explaining entry. New drugs are 1.5 times more likely to be launched in markets that share a border or a language of a drug company's country of headquarters. The effect of competition depends on the characteristics of both the potential entrant and incumbents: domestic entrants prefer to compete with domestic incumbents. Despite the potential for licensing and low transportation costs, the match between the innovating firm and market conditions remains an important determinant of entry.

Effects of Regulation on Drug Launch and Pricing in Interdependent Markets

Abstract and Figures

Recommended publications

Price Regulation, Industry Structure and Innovation

Launch and Pricing Strategies of Pharmaceuticals in Interdependent Markets

Regulation of Price and Reimbursement for Pharmaceuticals

The Economics of the Biopharmaceutical Industry

Pharmaceutical Pricing in Emerging Markets: Effects of Income, Competition and Procurement