Eur J Health Econ (2015) 16:7382
DOI 10.1007/s10198-013-0555-3
ORIGINAL PAPER
Competitive pricing within pharmaceutical classes: evidence
on follow-on drugs in Germany 19932008
Michael T. Mueller Alexander Frenzel
Received: 15 August 2013 / Accepted: 11 December 2013 / Published online: 27 December 2013
Springer-Verlag Berlin Heidelberg 2013
Abstract
Objectives Competition from follow-on drugs has
been a highly controversial issue. Manufacturers launching
new molecules in existing drug classes have often been
criticized for inflating health systems expenses, but it has
been argued that such drugs increase therapeutic options.
Economic theory suggests that follow-on drugs induce
price competition. We contribute to this discussion by
addressing the topic of pricing at market entry and price
development in the German market.
Methods We measure determinants of price strategies of
follow-on drugs using regression analyses, considering all
new molecules launched in the German market from 1993
to 2008. Prices of products are standardized on defined
daily dosages controlling for sales volumes based on data
from the IMS Health DPM database and for the therapeutic
quality of a new product using ratings by Fricke/Klaus as a
proxy for innovation.
Results We identify prices correlating with therapeutic
value at market entry. While the first two molecules engage
in quality competition, price discounts below the market
price can be observed from the third entrant on. Price
discounts are even more distinct in development races with
several drugs entering the market within 2 years and in
M. T. Mueller (&)
Lehrstuhl fur Institutionenokonomik und
Gesundheitssystemmanagement, Universitat Witten/Herdecke,
Alfred-Herrhausen-Str. 50, 58448 Witten, Germany
e-mail: michael.t.mueller@gmail.com
A. Frenzel
IMS Health GmbH & Co. OHG, Darmstadter Landstr. 108,
60598 Frankfurt, Germany
e-mail: AFrenzel@de.imshealth.com
classes with a low degree of therapeutic differentiation.
Prices remain relatively constant over time.
Conclusion This study contributes to assessments of
competition in pharmaceutical markets focusing on price
strategies of new market entrants. After an initial phase of
market building, further follow-on products induce price
competition. Largely unchanged prices after 4 years may
be interpreted as quality competition and can be attributed
to prices in Germany being anchor points for international
price referencing.
Keywords Pharmaceuticals  Pricing  Me-too drugs 
Therapeutic classes
JEL Classification I11  I18  L11  L13  L52  L65
Introduction
Competition within therapeutic drug classes from follow-
on or me-too drugs has been a highly controversial
issue. Manufacturers launching therapeutic substitutes in
existing drug classes have often been criticized for inflating
health systems expenses and misallocating R&D resources
while providing little or no incremental therapeutic benefit
to patients [13]. On the other hand, it has been argued that
such drugs increase therapeutic options [4, 5] and that the
commercialization is the result of R&D races between
manufacturers in many cases [69]. With regards to market
impact, economic theory suggests that follow-on drugs
induce price competition, which has been confirmed for
unregulated markets in selected analyses [913].
We contribute to this discussion by addressing the topic
of pricing at market entry and price development in the
German market. We adapt a methodology used by Lu and
123
74
Comanor [10], henceforth LC, in the USA as well as
Ekelund and Persson [14], henceforth EP, in the price-
regulated Swedish market. Conducting the analysis with
the distinct focus on competition within pre-existing drug
classes, we aimed to evaluate price strategies of follow-on
drugs and their consequences in the German market.
The paper is structured as follows: First we review
previous research to develop our hypotheses and then
introduce the data and methods used. The results are out-
lined and discussed and the paper concludes with impli-
cations for further research.
Previous research and hypotheses
Previous research
LC conducted a holistic assessment of new molecule
pricing for the period from 1978 to 1987 in the USA,
identifying the therapeutic value and the competitive
environment in the respective markets as main explanatory
variables for price setting. LC confirm previous empirical
findings [1517] on imitative pricing strategies for thera-
peutically equivalent drugs and premiums for innovations.
They also find markups for acute and discounts for chronic
drugs hinting at skimming and penetration strategies
respectively. Intensified competition from an increasing
number of branded substitutes at market entry leads to
additional price discounts, whereas an increasing share of
generic competition, weakening the competitive position of
existing molecules, allows an additional premium.
For the price development of new drugs 48 years after
launch, LC find proof of skimming for innovative as well
as acute and penetration strategies for therapeutically
equivalent as well as chronic drugs. An increasing number
of competitors have an additional depressing effect on
prices, whereas competition from off-patent drugs does not
influence price trends.
In a more recent descriptive assessment with focus on
launches in the USA from 1995 to 1999, DiMasi [18]
confirms LCs results. He finds relative price premiums for
innovations and drugs used in acute conditions as well as
price discounts taken when the number of competitors
increases.
EP adopt LCs approach for an analysis of the Swedish
market in which new drug prices are set by the National
Social Insurance Board in order to be reimbursable within
the health system. For the period from 1987 to 1997, they
find significant premiums for all new drugs. The magnitude
of the markup, however, correlates with therapeutic value.
Variables controlling for the number of competitors and
drugs in acute conditions show negative, yet insignificant,
values. Competition from generics does not have any
123
M. T. Mueller, A. Frenzel
influence on launch prices. Although the regulator is
granting higher premiums at market entry, the price trend
4 years after entry is characterized by a decline across
classes. This result is in line with the price-cap regulation
in Sweden which does not permit price increases across the
overall product portfolio on manufacturer level.
Studies on pricing at market entry in Germany have
focused on selected parameters of competition or the ana-
lysis of specific therapeutic classes so far. Kanavos et al.
[12] compare price effects arising from the entrance of
additional competitors in the class of statins across four
European countries. In this limited sample, they find pre-
miums in Germany for two of the four follow-on drugs
which maintain these price points over time. Haussler et al.
[9] make the first distinct assessment of competition from
follow-on drugs in the German market describing price and
volume effects in 12 therapeutic classes. They find that
new molecules enter the market with price discounts,
thereby limiting price increases and inducing price com-
petition among existing branded substitutes. Price compe-
tition among existing brands is also further reinforced by
the existence of generics.
Hypotheses
Like in the USA, the German Statutory Healthcare Insurance
(SHI) system allows free pricing. Prices actually paid in the
USA may, however, be determined by price regulations in
public programs as well as discounts negotiated with Phar-
macy Benefit Management companies [19]. In Germany, list
prices historically were fully reimbursable within the SHI
system. With the introduction of reference pricing for on-
patent drugs in 2005 and the benefit analysis introduced as
part of the Arzneimittelmarktneuordnungsgesetz (AMNOG)
in 2011, the reimbursement level of new drugs can, however,
be limited. Prices actually paid may also vary depending on
rebate contracts for on-patent drugs that can be negotiated
between individual sickness funds and manufacturers from
2007 on [20].
Yet, for our observation period from 1993 to 2008, we
expect results at market entry to be closely related to LCs
findings for the US market (hypothesis 1). As we focus on
pre-established therapeutic classes only, price premiums
for more innovative drugs should, however, be less
prominent than in the US sample. LC might also consider
competitive situations in which products are compared
across pharmacological classes, whereas this stronger
product differentiation could lead to wider distributed price
points.
The phenomena of skimming strategies for drugs used in
acute and penetration strategies for drugs used in chronic
conditions have only been described in the US market so
far. Price discounts for chronic drugs are backed by
Competitive pricing within pharmaceutical classes
economic theory, since doctors and patients need to build
experience with the application of new drugs. Switching
costs are furthermore relatively higher for drugs that are
used continuously (chronic) than sporadically (acute) [21,
22]. LC, however, also base their explanation of price
discounts for chronic drugs on their stronger pervasion
among more elderly patients who face higher co-payment
rates and hence are more price sensitive. In the German
SHI system there are fixed co-payments, whereas patients
can be exempt from co-payments when a certain level of
financial burden is reached. Physician drug budgets may,
however, be a limiting factor on the selection of new drugs
[23]. Hence, we expect penetration strategies to be less
distinct than in the USA (hypothesis 2).
Intensified competition from an increasing number of
branded substitutes has been described to lead to additional
price discounts taken at market entry [11, 13] (hypothesis
3). In therapeutic classes where substitutability among
drugs is especially high and differentiation is low, physi-
cians may have become experienced with/accustomed to
prescribing the general class drugs [12]. New molecules
might therefore need to enter the market with additional
discounts in such homogeneous classes to win customers
(hypothesis 4).
Apart from the competition between horizontally dif-
ferentiated products, the timing of their market entry could
also play an important role in new molecule pricing
(hypothesis 5). As a result of development races, physicians
may be faced with several new drugs with similar thera-
peutic value within a short period of time. This might
require additional detailing efforts from manufacturers or
larger price discounts taken to motivate doctors to change
prescription habits. Such a behavior would also be in line
with limit pricing theory according to which the entry of
successors may be deterred through aggressive pricing [24].
With reference pricing and aut idem substitution being
in place, a strong reducing effect of generic competition on
the prices of respective branded drugs has been described
for Germany in numerous studies [13, 2527]. Therefore,
new drugs are expected to enter with premiums in situa-
tions where prices of off-patent drugs have been driven
down by generic competition (hypothesis 6).
Concerning the development of new products entrance
prices, we would expect competition from on-patent and
off-patent drugs to be main explanatory variables as iden-
tified by LC for price changes in the first years after
commercialization. However, more recent findings on
competition in selected therapeutic classes in prescription
and hospital drug markets in various countries [12, 28, 29]
suggest that manufacturers engage in quality competition
based on product attributes instead of price competition
after market entry and these quality attributes may be the
dominant factor for a physicians prescription decision
75
[30]. Another influential factor inhibiting price decreases
may be the importance of the German market as a major
anchor point for external price referencing of new drugs in
numerous health care systems [31, 32]. Hence, it may be in
the prime interest of manufacturers to maintain price levels
in Germany instead of reacting to changes in the market
environment in order not to endanger profits in other
markets (hypothesis 7).
Major price changes may, however, be induced by
regulatory measures such as therapeutic reference pricing
for on-patent drugs, which was, however, not implemented
before 2005.
Data and methods
Data
We consider all new molecules launched in the German
market in the period from 1993 to 2008 (n = 458)
according to the publication series Neue Arzneimittel. We
limit our selection of new drugs to the year 2008 to obtain a
consistent data set for the analyses of entry prices and the
price development after 4 years.
Ratings by Fricke/Klaus, a widely known and influential
classification system from the same series Neue Arznei-
mittel from 1982 on, are adapted as a proxy for the ther-
apeutic quality of a new product. The authors rate the grade
of innovation of new drugs along a scale comparable to the
one applied by the US Food and Drug Administration
(FDA):
A Innovative structure or novel mechanism of action
with therapeutic relevance,
B Improvement of pharmacodynamic or pharmacokinetic
features of pre-existing mechanisms of action,
C Me-too drug with no or only marginal differences to
existing molecules,
D Not sufficiently proven mechanism of action or
uncertain therapeutic relevance.
Fricke and Klaus rating has gained widespread atten-
tion. For example, it is used for analyses of potential cost
savings in classes with me-too drugs in the series Ar-
zneiverordnungs-Report and even serves as a basis for
regional mandatory me-too prescription quotas and sub-
stitution lists introduced in the Nordrhein SHI region in
2006. Owing to its similarity to the FDA rating, which,
however, only focuses on therapeutic relevance, we adapt
this German rating scheme as a proxy for new molecule
quality in our empirical analysis.
This rating, however, has been subject to broad criticism.
As assessments are made on the basis of the respective
market situation at market entry [33], the winner of a
123
76
development race can be classified as highly innovative,
whereas the runner-up will be considered a me-too drug.
Also, ratings remain fixed and are not revised later on the
basis of results from phase IV studies or new indications.
Furthermore, the classification has been characterized as
being made from a predominantly pharmacological point of
view. Hence, A ratings may be awarded for new mecha-
nisms even if the clinical value is not yet evident [34]. As a
consequence, Fricke and Klaus have been implementing
mixed ratings (e.g., A/C, B/C) from 1991 on a low scale to
cope with this problem. An overall modification of the
rating scheme differentiating between pharmacologic and
therapeutic levels of innovation for every new molecule
has, however, not been made.
To avoid these controversies, drugs with mixed Fricke/
Klaus ratings (e.g., A/C, C/D) are excluded from the
analyses. Molecules rated A are also not considered as it is
not always obvious whether this grade is awarded for a real
therapeutic innovation or only for a new pharmacological
mechanism [34]. We thereby also avoid the problem of
finding appropriate substitutes of A drugs discussed by EP.
As A drugs always establish a new substance class, our
analysis hence solely focuses on new competition within
pre-existing therapeutic classes.
Next, drugs that are (primarily) used in hospitals are
excluded, as paid prices and resulting pricing strategies in
this segment may differ from the prescription market [35,
36]. In accordance with LC and EP, several therapeutic
classes are also not considered. This applies to antineoplas-
tics as well as systemic antivirals that can be used in com-
bination therapies and hence might react to price changes of
supplements. Furthermore, vaccines, diagnostics, and indi-
vidually dosed medications such as anesthetics are excluded.
Drugs with differing therapeutic and resistance spectrums
such as antibiotics or antimycotics are also not considered
unless they are classified as essentially substitutable. We
conduct this assessment as well as the identification of
therapeutic substitutes considering various evaluations of
substance classes [9, 3739] as well as further clinical lit-
erature and consultation with physicians where necessary.
For the definition of the competition landscape, the rele-
vant, direct therapeutic substitutes of a new drug at the time
of market entry are considered, reflecting the strategic pric-
ing decision the manufacturer must have taken when the
product was launched. This is usually, but not always,
restricted to competitors within the same ATC level 4 sub-
group.1 For instance, we compare the price of valsartan as the
first AT2 antagonist in our sample only with losartan as the
1
We use the ATC classification developed by the European
Pharmaceutical Marketing Research Association (EphMRA), which
categorizes drugs rather by indications than the more pharmacolog-
ically oriented WHO ATC.
123
M. T. Mueller, A. Frenzel
innovator of the substance class (rated A by Fricke/Klaus),
not with related hypertension medication such as ACE
inhibitors or calcium antagonists. If a new drug can be used
for several diseases, only the main indication is considered.
In case of doubt, the main indication is defined by higher
sales of a specific dosage/package combination of a new
product. For substances that can be used in various topical
areas (e.g., steroids as inhalants or dermatological agents),
all forms for which the product has been discussed and rated
by Fricke/Klaus are considered. In accordance with LC and
EP, new products are classified as acute for indications
lasting no more than 3 months and chronic otherwise.
As a result of these selection criteria, our sample is
limited to 108 new molecules from 38 different therapeutic
classes. Our approach differs from LC and EP with regards
to three characteristics: Drugs must be prescribed (pri-
marily) outside hospitals, enter an established drug class as
a successor, and be fundamentally therapeutically substi-
tutable within this class. This provides for great analytical
clarity to identify pricing effects.
We compare prices of products based on the defined
daily dose (DDD) recommendations at the time of market
entry from the World Health Organization (WHO), Ger-
man Institute of Medical Documentation and Information
(DIMDI), and the Rote Liste drug directory where neces-
sary. Similar to the approaches adopted by LC and EP,
prices are compared for the main application form (usually
tablets for systemic preparations). On package level, the
largest package containing 1 DDD per tablet is selected
where possible and reasonable. Otherwise, the most com-
parable dose/package combination, usually the one with
highest sales, across competitors is chosen. In accordance
with EP, competitor prices are weighted with market shares
at brand level in the year before market entry. Observations
of new molecule prices after 4 years are deflated to the
respective base year using the German consumer price
index (CPI).
We use data from the IMS DPM database from IMS
Health to perform our analysis. IMS DPM covers the entire
German pharmaceutical market at the interface of phar-
macy purchasing through a complete coverage of pur-
chases from 16 full-line wholesalers as well as a
representative pharmacy panel covering direct business
from short-line wholesalers, parallel importers, and man-
ufacturers. Prices in the database are adapted from the
Lauer-Taxe and represent net manufacturer sales prices.
Volume and price data have been extracted with annual
observations in the period from 1992 to 2012.
Methods
To estimate pricing at market entry and price development,
we adopt the linear regression models used by LC and EP.
Competitive pricing within pharmaceutical classes
Several alternative variables measuring competition from
on- as well as off-patent drugs are added on in order to
specify effects within therapeutic classes further on.
lrelp b0 b1 rateb b2 acute b3 comp b4 gx e lgs
Model I
lratiot d0 d1 rateb d2 acute d3 dblp d4 dclp
d5 comp d6 gx c
Model II
lrelp Logarithm of the ratio of the new molecule price
per DDD to the volume weighted price of
therapeutic branded substitutes.
lratio(t) Logarithm of the ratio of the CPI deflated price
t years after introduction to its real introduction
price.
General control variables
rateb Dummy variable for therapeutic quality, equals 1
for B and 0 for C ratings.
acute Dummy variable, equals 1 for acute and 0 for
chronic indications.
dblp Interactive variable, product of lrelp and rateb for
rateb = 1.
dclp Interactive variable, product of lrelp and rateb for
rateb = 0.
Variables controlling for on-patent competition (comp):
lns(t) Logarithm of 1 ? the number of branded
substitutes at t years.
launch(y) Dummy variable, equals 1 if another
competitor enters the market within
y years (proxy for a development race).
launch(y)a/p Dummy variable, equals 1 if launch(y) = 1
and competitors enter the market before
(ante) and/or after (post) the new molecule.
me2class Dummy variable, equals 1 if the class has been
characterized in the literature as a substance
class with very high substitution potential and
a low perceived differentiation among
molecules by patients and physicians.
Variables controlling for off-patent competition (gx):
dg Dummy variable, equals 1 if at least one
competitor is off-patent (and has generic
competitors) at the time of market entry of
the new molecule.
lg Logarithm of 1 ? the share of branded
substitutes that are off-patent at market
entry, whereas the share is expressed as a
whole number.
77
bg Dummy variable, equals 1 if at least one
competitor is off-patent at time t when there
was none at market entry.
Logarithm of 1 ? the ratio of the share of
branded substitutes that are off-patent at time
t over the share of branded substitutes that
were off-patent at market entry, whereas the
share is expressed as a whole number.
gs(t)(z) Dummy variable, equals 1 if there are exactly
z off-patent molecules at time t.
gs(t)(z)min Dummy variable, equals 1 if there are at least
z off-patent molecules at time t.
Results
Table 1 summarizes statistics on relative introduction pri-
ces in the German marketplace in comparison to previous
US and Swedish data. As expected, innovative B sub-
stances show a distinct price premium of an average
?182 %, which is higher than in the US, but lower than in
the Swedish market. The median in the German sample is,
however, only at 1.00. This suggests a highly skewed
distribution with predominately imitative pricing strategies
and only a few very highly priced outliers. For C-rated
drugs, median and mean values show proof of pricing at
competitors levels. In comparison to the Swedish and
especially the US data, the standard deviation is consid-
erably smaller, again supporting imitative pricing in most
cases.
In contrast to the other countries data, no differences
between prices of drugs in acute and chronic conditions are
observable. While acute B drugs show a 12 % discount as
median and a 401 % price premium as mean values, sample
size (n = 5) and standard deviation in this subgroup
diminish a possible significance of these observations.
Descriptive statistics on the price development 4 years
after entry in Table 2 reveal a decline of real prices of
25 % for B as well as C drugs. In contrast to the US
market, there is no indication for penetration or skimming
strategies, nor does the price decrease reach levels like the
average minus 1027 % in Sweden. Since the standard
deviation in our sample is very small, this minor average
price change can most likely be attributed to the deflation
of nominal prices exclusively.
The econometric analysis is conducted using Stata 12.0.
To ensure comparability with previous results, we compute
standard ordinary least squares (OLS) regressions for the
regressions used by LC and EP and estimate heteroske-
dasticity robust standard errors only in enhancements of
these basic models.
123
78 M. T. Mueller, A. Frenzel

Table 1 Relative introduction prices at market entry (relp)

Class Type Germany 19932008 USA 19781987 (LC 1998) Sweden 19871999 (EP 2003)
n Median Mean SD Range n Median Mean SD Range n Median Mean SD Range

A Acute 3 2.97 3.12 0.88 2.324.07 5 1.43 1.59 0.56 1.102.55

Chronic 10 2.07 3.11 2.76 0.208.08 3 10.97 9.40 4.88 3.9313.30
Combined 13 2.47 3.11 2.47 0.208.08 8 2.04 4.52 4.83 1.1013.30
B Acute 5 0.88 5.01 5.52 0.6614.20 19 1.72 3.09 3.65 0.4013.60 30 1.64 3.75 4.96 0.1719.28
Chronic 14 1.00 2.04 2.34 0.758.31 29 1.19 1.63 1.14 0.485.97 42 2.55 3.92 4.77 0.0920.32
Combined 19 1.00 2.82 3.71 0.6614.20 48 1.35 2.21 2.53 0.4013.60 72 2.25 3.86 4.85 0.0920.32
C Acute 26 0.98 1.08 0.38 0.402.01 16 1.23 1.37 0.68 0.553.13 63 1.26 2.16 3.09 0.0416.70
Chronic 63 1.00 1.04 0.29 0.651.92 53 0.94 1.07 0.80 0.376.00 75 1.13 2.18 3.54 0.1226.17
Combined 89 1.00 1.05 0.32 0.402.01 69 0.97 1.15 0.78 0.376.00 138 1.16 2.17 3.33 0.0426.17

Table 2 Real price changes 4 years after market entry (ratio4)

Class Type Germany 19932008 USA 19781987 (LC 1998) Sweden 19871999 (EP 2003)
n Median Mean SD Range n Median Mean SD Range n Median Mean SD Range

A Acute 3 1.11 1.06 0.23 0.811.26 16 0.78 0.83 0.28 0.521.65

Chronic 10 0.87 0.92 0.17 0.681.19 12 0.75 0.73 0.12 0.440.91
Combined 13 0.87 0.95 0.19 0.681.26 28 0.77 0.78 0.21 0.441.95
B Acute 5 0.95 0.98 0.11 0.941.03 19 0.95 1.00 0.20 0.791.59 20 0.93 0.90 0.33 0.342.05
Chronic 14 0.97 1.00 0.10 0.821.22 29 1.18 1.13 0.16 0.841.42 26 0.82 0.84 0.13 0.641.10
Combined 19 0.97 0.99 0.11 0.821.22 48 1.09 1.08 0.19 0.791.59 46 0.86 0.86 0.24 0.342.05
C Acute 26 0.98 0.97 0.04 0.611.19 16 1.04 1.11 0.26 0.811.71 42 0.89 0.89 0.29 0.402.51
Chronic 62 0.95 0.98 0.10 0.661.29 53 1.26 1.26 0.30 0.682.01 33 0.79 0.81 0.10 0.641.07
Combined 88 0.95 0.97 0.09 0.611.29 69 1.22 1.23 0.29 0.682.01 75 0.84 0.85 0.23 0.402.51

Table 3 compares determinants of relative prices at
market entry with previous results from LC and EP
respectively. We find a small (yet insignificant) price pre-
mium for new products in the constant term. As expected,
B-rated drugs show a significant markup (hypothesis 1).
The strong effect of ?44 % almost reaching Swedish
levels can, however, most likely be attributed to few rela-
tively high priced outliers we identified in our sample. The
dummy for acute drugs is, however, insignificant and only
shows a negligible effect (hypothesis 2). Hence, we drop
this variable in further regressions.
Like in the two previous studies, price discounts taken at
market entry increase with the number of branded competitors
per substance class (hypothesis 3). Interestingly, competition
from off-patent drugs and respective generics controlled for in
regressions #23 improves the model quality drastically. This
implies that off-patent drugs engage in a price competition
induced by the entry of generics (hypothesis 6).
Regressions with further detailed variables on compe-
tition from patent and generic drugs are displayed in
Table 4, now involving the estimation of heteroskedasticity
robust standard errors. Omitting the dummy for acute drugs
and controlling for on- as well as off-patent competition,
our constant term is still slightly insignificant at the 5 %
level. However, it becomes apparent that this term is
overcompensated by the effect from one additional com-
petitor expressed in lns0. This implies that the first com-
petitor in a substance class is entering the market with a
premium. Pricing below the average competitors level is
first observed when the number of drugs in a substance
class increases from two to three2 (hypothesis 3).
Apart from the number of substitutes, a distinct influence of
development races on the pricing of new molecules becomes
apparent. We find an additional price discount of 18.7 % in
scenarios in which at least one additional competitor enters the
market within a period of up to 2 years (hypothesis 5). There
is, however, no separate effect indicating intensified compe-
tition from development races when controlling for additional
entries within a time frame of 3 years or more. Results on
whether price discounts for parallel launches within 2 years
are only made when another competitor is entering the market
before (ante) and/or after (post) the new molecule are weak
2
We find further evidence on a gradually intensifying price
competition by conducting a sensitivity analysis on the impact from
the number of competitors.

123
Competitive pricing within pharmaceutical classes 79

-0.197 (-1.38)
-0.119 (-1.12)
and insignificant. Hence, this price discount can be charac-

0.486** (2.89)

0.463** (2.98)

-0.077 (0.15)
0.921* (2.44)
terized as both proactive and reactive when facing intensified
competition during parallel product launches.
In addition, we find a negative impact on price when

0.08
0.06
3.82
218
(3)

applying a tightened definition of therapeutic class effects.

Controlling for those classes which are characterized as

-0.197 (-1.35)
-0.124 (-1.14)
0.485** (2.89)

0.463** (2.98)
0.920* (2.48)

having highly interchangeable products with a low degree of

0.004 (0.02)
differentiation perceived by physicians and patients [9, 37],
a price discount of 17.6 % becomes apparent (hypothesis 4).

0.08
0.06
3.82
218
(2)

Finally, effects arising from an increasing number of

Sweden (EP 2003)

off-patent competitors are tested. Price premiums in sce-

-0.197 (-1.37)
-0.124 (-1.24)
0.485** (2.90)

0.463** (2.99)
0.920* (2.45)

narios with only one or multiple generic molecules per

drug class differ only marginally. In a sensitivity analysis,
we find a premium of more than 80 % in classes with at
0.08
0.07
4.79
218
(1)

least five generics; this effect can, however, only be

attributed to observations in the class of nonsteroidal anti-
-0.376*** (-4.41)

inflammatory drugs.
0.335** (2.84)

0.086** (2.96)
(0.221* (2.32)

0.285* (2.34)

Regressions for the price development after 4 years are

0.387 (1.88)

presented in Table 5. Results confirm the indications from

the descriptive analysis that only minimal price movements
0.32
0.29
130

n/a
(3)

take place (hypothesis 7). All independent variables as well

as the overall model are highly insignificant. The relatively
-0.386*** (-4.45)

strongest explanatory value is provided by the number of

0.336** (2.84)

0.332** (2.78)

substitutes (t = -1.13), hinting at a slight price decrease

0.293* (2.39)
0.230 (1.82)
0.388 (1.78)

when competition intensifies. A possibly positive effect can

be attributed to the interactive variable dclp (t = 1.16),
0.32
0.29
130

n/a

suggesting that price decreases are more distinct among C

(2)

drugs that have been priced relatively higher at market entry.

-0.332*** (-3.75)

New regressors introduced to detail effects on branded

USA (LC 1998)

0.334** (2.75)

and generic competition have been tested for collinearity.

0.295* (2.32)

0.267* (2.13)
0.374 (1.76)

Models at market entry demonstrate cluster robustness

when clustering effects by substance groups, manufactur-
0.27
0.25
130

ers, or launch years.

n/a
Table 3 Determinants of relative introduction prices 1/2 (model I: lrelp)

(1)

-0.263** (-2.83)
0.435*** (3.85)

0.141*** (4.87)

Discussion
0.155 (1.18)

0.077 (0.82)

We find strong evidence for characteristics of a functioning

10.12
0.28
0.25
108
(3)

pricing mechanism for new products in accordance with

economic theory: Discounts correlate with competitiveness
-0.291** (-2.91)

of the market, i.e., the number of branded substitutes, and

0.444*** (3.86)

0.482*** (4.45)

t Statistics in parentheses, p values for S.E

premiums correlate with therapeutic value.

0.199 (1.46)

0.045 (0.47)

* p \ 0.05, ** p \ 0.01, *** p \ 0.001

Applying a stricter definition of substitutability than LC

and EP to our data, we only consider molecules whose ther-
0.26
0.23
9.01
108
(2)

apeutic class has already been established by an originator.

The skewed distribution of prices for B-rated drugs with only
-0.030 (-0.34)
0.436*** (3.49)

few highly priced outliers might, however, reveal one of the

0.031 (0.21)

0.082 (0.79)

weaknesses of the classification system by Fricke/Klaus,

Germany

which may reward innovation on pharmacological as well as

0.12
0.09
4.59
108
(1)

therapeutic level with the same grade, ignoring the real per-
ceived value a new drug may bring into the marketplace.
constant

adj.r2

Competition within therapeutic classes seems to be

acute
ratea
rateb

lns0
dg

r2

characterized by a low degree of product heterogeneity and

lg

F
n

123
80 M. T. Mueller, A. Frenzel

Table 4 Determinants of relative introduction prices 2/2 (model I: lrelp)

(4) (5) (6) (7) (8) (9)

constant 0.206 (1.98) 0.299* (2.55) 0.187 (1.62) 0.297* (2.51) 0.345* (2.91) 0.201 (1.91)
rateb 0.444* (2.26) 0.417* (2.29) 0.447* (2.23) 0.423* (2.29) 0.382* (2.23) 0.445* (2.23)
lns0 -0.228** (-3.16) -0.255** (-2.84) -0.294** (-3.07) -0.255** (-2.82) -0.205* (-2.13) -0.284** (-3.23)
dg 0.487*** (3.81) 0.460*** (3.74) 0.492*** (3.70) 0.461*** (3.72) 0.435*** (3.62)
launch2 -0.187* (-2.10) -0.142 (-1.03) -0.185* (-2.16)
launch3 0.033 (0.35)
launch2a -0.012 (-0.12)
launch2p 0.067 (-0.68)
me2class -0.176* (-2.08)
gs01 0.490*** (3.46)
gs02 min 0.470*** (4.38)
n 108 108 108 108 108 108
r2 0.26 0.29 0.26 0.29 0.31 0.26
F 4.99 4.06 3.81 2.66 3.60 5.20
t Statistics in parentheses, p values for heteroskedasticity robust SE
* p \ 0.05, ** p \ 0.01, *** p \ 0.001

spatial agglomeration in the marketplace. In order to
address customers, new entrants must therefore take price
discounts which are even more distinct in classes with little
product differentiation. This also holds for timing of mar-
ket entries as a result of development races.
Price competition at entry, however, first sets in when
the third competitor enters a class. These results are in line
with findings from Lexchin [40] who identifies price dis-
counts in the Canadian market first taken by the fourth
molecule. Earlier entrants therefore have the opportunity to
build a new market and retain customers. As Kanavos et al.
[12] point out, physicians become familiar with prescribing
these general class drugs over time. When markets mature,
the small differentiation of new substitutes does not offset
the level of experience physicians have reached with the
general class. New substitutes therefore need to compen-
sate for switching costs with lower prices.
Competition from generics is an indispensable explan-
atory variable in our model. Off-patent molecules engage
in price competition with generics. This supports the strong
generics culture in the German market with phenomena
such as branded generics and also affirms the functioning
of mechanisms such as reference pricing or aut idem sub-
stitution in the German market.
Whereas price competition gradually sets in at market
entry, quality competition seems to hold for the price
development of new drugs. In line with various recent
studies [12, 28, 29], we find constant prices and no vari-
ables explaining the price development after 4 years in our
data beyond the influence of deflation. Apart from com-
petition based on product attributes, the lack of price
reactions may also be attributed to two other specificities in
Germany: First, the German market serves as a major
anchor point for external price referencing of new drugs in
numerous health care systems [19]. Manufacturers there-
fore pursue strategic anchor pricing by launching drugs first
in the German market and aim to maintain these price
points over time in order not to endanger profits in other
international systems [31, 32]. Second, manufacturers may
be penalized for increasing prices. Apart from mandatory
manufacturer rebates off the list price, the German SHI
system has traditionally implemented price freezes peri-
odically, during which any list price increases must be
given as additional discounts to sickness funds. However,
in a sensitivity analysis of price changes by calendar year,
no effects that could be attributed to price freeze periods3
or other modifications of the Social Insurance Code (SGB
V) become apparent. We also find no influence of rebate
contracts, which have been introduced into the system in
2003 and allow negotiations of (non-disclosed) rebates off
the list price between individual sickness funds and man-
ufacturers. Yet, this option has been exerted rarely for on-
patent drugs in initial years [20].
Conclusion
We have applied an approach used by LC and EP with a
focus on competition within therapeutic classes in the
German marketplace between 1993 and 2008. Our study
confirms previous research identifying the therapeutic
3
In a study of overall price movements in the German market
between 2004 and 2006, Stargardt [13] identifies quarterly price
reductions during a price freeze of 0.38 % compared to an increase of
0.32 % before and after this period. Our model would, however, not
be able to capture such small movements precisely.

123
Competitive pricing within pharmaceutical classes 81

-0.168** (-2.90)
value of a new molecule as a main explanatory variable for

-0.137 (-0.93)

-0.008 (-0.31)
-0.003 (-0.20)
-0.010 (-0.87)
-0.010 (-0.40)

-0.313 (-1.28)
price setting at market entry. While new classes allow

0.005 (0.06)
0.075 (1.90)
market building through quality competition of the first
two molecules, price competition sets in when further

0.10

2.04
(12)

128

.0.5
successors enter the market. The principle of quality
competition holds for the price development of new mol-
-0.174** (-3.00)

ecules, whereas the role of strategic anchor pricing con-

-0.130 (-0.88)

-0.008 (-0.32)
-0.001 (-0.08)
-0.009 (-0.71)
-0.014 (-0.54)
-0.032 (-0.47)
tributes to lacking price reactions in the market.
0.004 (0.05)
0.077 (1.94)

Our study raises various questions for further research.

Although new molecules do not engage in price competi-

0.10
0.04
1.85
(11)

128
tion in initial years, their entry might induce price com-
petition among existing substitutes, especially towards the
Sweden (EP 2003)

-0.175** (-3.04)

end of their patent periods. Previous results on this issue

-0.126 (-0.85)

-0.008 (-0.33)
-0.001 (-0.04)
-0.009 (-0.76)
-0.008 (-0.09)

are mixed, describing price [9] as well as quality compe-

0.003 (0.04)
0.076 (1.93)

tition [12, 29] in selected classes.

Other issues worth exploring may be whether the pro-
0.09
0.05
2.09
(10)

128

cess of market building is limited to a certain time frame

[12], if market shares correspond to therapeutic quality and
-0.339*** (-4.11)

-0.138** (-2.62)
-0.076** (-2.79)
-0.093* (-2.16)

-0.106* (-2.46)

first mover advantages [11, 12], or how price anchoring

-0.054 (-1.29)
-0.007 (-0.13)
0.300*** (7.59)

influences prices and shares in other systems.

0.004 (0.39)

Our paper reflects the old reality in Germany where

freely priced new drugs automatically were fully reim-
0.25
0.20
(12)

130

n/a

bursed. With the benefit analysis as a process having come

into place with the AMNOG in 2011, reimbursement levels
-0.337*** (-4.08)

-0.136** (-2.57)
-0.075** (-2.74)
-0.092* (-2.12)

-0.107* (-2.47)

of new molecules are subject to negotiations with sickness

-0.052 (-1.26)
-0.008 (-0.15)
0.295*** (6.00)

funds. Drugs may also be included in reference price

0.002 (0.05)

clusters when no significant additional therapeutic benefit

over present therapies is shown. As a result, reimbursement
0.25
0.20
(11)

130

n/a

prices should become less skewed, reflecting the thera-

peutic value of a new drug even more strongly. With
-0.339*** (-4.14)

-0.139** (-2.66)
-0.077** (-2.82)
-0.092* (-2.15)

-0.108* (-2.50)

regards to launch prices, we would expect that manufac-

USA (LC 1998)

-0.055 (-1.33)
-0.007 (-0.13)
0.302*** (6.56)

turers will add additional premiums, factoring in possible

price discounts that the negotiation process might bring.
The first empirical results [41, 42] also hint at the selection
Table 5 Deflated price ratios 4 years after launch (model II: lratio4)

0.25
0.21
(10)

130

n/a

of inappropriate competitors in manufacturers dossiers as

a possible means to secure higher reimbursement levels.
-0.001 (-0.07)

-0.006 (-0.25)

-0.023 (-1.13)

With many major market impacts having occurred in the

0.005 (0.15)

0.018 (0.59)

0.048 (1.16)

0.004 (0.48)

past years such as the AMNOG benefit evaluation or the

-0.02

increasing share of rebate contracts for on-patent drugs, we

0.03

0.57
(12)

107

believe that a new holistic empirical investigation will be

feasible in a few years when sufficient data on launches
-0.002 (-0.09)

-0.010 (-0.40)

-0.019 (-0.95)
-0.026 (-0.69)

under these changed market conditions is available.

0.002 (0.07)

0.025 (0.83)

0.045 (1.08)

t Statistics in parentheses, p values for S.E

* p \ 0.05, ** p \ 0.01, *** p \ 0.001
-0.02
0.04

0.62
(11)

107

References
-0.002 (-0.09)

-0.008 (-0.31)

-0.021 (-1.06)
0.004 (0.12)

0.020 (0.68)

0.047 (1.13)

1. Angell, M.: The Truth About the Drug Companies: How They
Germany

Deceive Us and What to Do About It. Random House, New York

-0.02

(2005)
0.03

0.65
(10)

107

2. Hollis, A.: Comment on The economics of follow-on drug

research and development: trends in entry rates and the timing of
constant

development. Pharmacoeconomics 23, 11871192 (2005)

adj.r2
acute
ratea
rateb

3. Glaeske, G.: Mit bitteren Pillen gesunde Geschafte: von man-

dalp
dblp
dclp
lns4

lgs
bg

r2

gelnder Differenzierung des Marktes und den Erfolgen von

N

123
82
Analogprodukten. In: Neubauer, G., Rebscher, H. (eds.) Ges-
undheitsokonomie und Gesundheitspolitik im Spannungsfeld
zwischen Wissenschaft und Politikberatung: [Festschrift fur
Gunter Neubauer], pp. 543559. Economica, Heidelberg (2006)
4. Wertheimer, A., Santella, T.: Pharmocoevolution: the benefits of
incremental innovation. http://www.who.int/intellectualproperty/
submissions/Pharmacoevolution.pdf (2005). Accessed 14 Aug 2013
5. Blochl-Daum, B.: Me-too drugs und das Konzept des Klas-
seneffekts. Wien. Med. Wochenschr. 156, 1718 (2006)
6. DiMasi, J.A., Paquette, C.: The economics of follow-on drug
research and development: trends in entry rates and the timing of
development. Pharmacoeconomics 22, 114 (2004)
7. DiMasi, J.A., Faden, L.B.: Competitiveness in follow-on drug
R&D: a race or imitation? Nat. Rev. Drug Discov. 10, 2327
(2011)
8. Munnich, F.W.: Innovatorischer Wettbewerb auf dem Arznei-
mittelmarkt. In: Bausch, J., Klauber, J., Schroder, H., Selke, G.W.
(eds.) Innovation im Arzneimittelmarkt, pp. 107129. Springer,
Berlin (2000)
9. Haussler, B., et al.: Analog-Wirkstoffe im Arzneimittelmarkt:
Therapeutischer Nutzen und Bedeutung fur die Ausgaben der
Krankenversicherungen. IGES, Berlin (2002)
10. Lu, Z.J., Comanor, W.S.: Strategic pricing of new pharmaceuti-
cals. Rev. Econ. Stat. 80, 108118 (1998)
11. Danzon, P.M., Chao, L.: Does regulation drive out competition in
pharmaceutical markets? J. Law Econ. 43, 311358 (2000)
12. Kanavos, P., Font, J.C., McGuire, A.: Product differentiation,
competition and regulation of new drugs: the case of statins in four
European countries. Manag. Decis. Econ. 28, 455465 (2007)
13. Stargardt, T.: Modelling pharmaceutical price changes in Ger-
many: a function of competition and regulation. Appl. Econ. 43,
45154526 (2011)
14. Ekelund, M., Persson, B.: Pharmaceutical pricing in a regulated
market. Rev. Econ. Stat. 85, 298306 (2003)
15. Reekie, W.D.: Pricing new pharmaceutical products. Croom
Helm, London (1977)
16. Reekie, W.D.: Price and quality competition in the United States
drug industry. J. Ind. Econ. 26, 223237 (1978)
17. Weston, F.: Pricing in the pharmaceutical industry. In: Chien, R.I.
(ed.) Issues in pharmaceutical economics, pp. 7195. Heath,
Lexington, MA (1979)
18. DiMasi, J.A.: Price trends for prescription pharmaceuticals
19951999. http://aspe.hhs.gov/health/reports/drug-papers/
dimassi/dimasi-final.htm (2000). Accessed 25 Oct 2013
19. OECD: Pharmaceutical Pricing Policies in a Global Market.
OECD, Paris (2008)
20. Mandl, J.: Rabattvertrage uber Patentgeschutzte Arzneimittel.
Nomos, BadenBaden (2012)
21. Hellerstein, J.K.: The importance of the physician in the generic
versus trade-name prescription decision. RAND J. Econ. 29,
108136 (1998)
22. Schmalensee, R.: Product differentiation advantages of pioneer-
ing brands. Am. Econ. Rev. 72, 349365 (1982)
23. Schreyogg, J., Busse, R.: Physician drug budgets in Germany and
effects on prescription behaviour. J. Pharm. Finance Econ. Policy
14, 7795 (2005)
24. Bergman, M.A., Rudholm, N.: The relative importance of actual
and potential competition: empirical evidence from the pharma-
ceuticals market. J. Ind. Econ. 51, 455467 (2003)
123
M. T. Mueller, A. Frenzel
25. Pavcnik, N.: Do pharmaceutical prices respond to potential patient
out-of-pocket expenses? Rand J. Econ. 33, 469487 (2002)
26. Kanavos, P., Costa-Font, J., Seeley, E.: Competition in off-patent
drug markets: issues, regulation and evidence. Econ. Policy 23,
499544 (2008)
27. Vandoros, S., Kanavos, P.: The generics paradox revisited:
empirical evidence from regulated markets. Appl. Econ. 45,
32303239 (2013)
28. Jena, A.B., Calfee, J.E., Mansley, E.C., Philipson, T.J.: Me-
Too innovation in pharmaceutical markets. Forum Health Econ.
Policy 12(1), 121 (2009)
29. Hostenkamp, G.: Do follow-on therapeutic substitutes induce
price competition between hospital medicines? Evidence from
the Danish hospital sector. Health Policy 111, 6877 (2013)
30. Serra-Sastre, V.: McGuire, A: Information and diffusion of new
prescription drugs. Appl. Econ. 45, 20492057 (2013)
31. Frenzel, A.: Pricing insights, prescription patterns and average
treatment costs across six therapy areas and 19 European coun-
triesDiscussing the effect of international price referencing on
average treatment cost of pharmaceuticals. Poster presented at
ISPOR 14th Annual European Congress, Madrid (2011)
32. Augurzky, B., Gohlmann, S., Gre, S., Wasem, J.: Effects of the
German reference drug program on ex-factory prices of prescrip-
tion drugs: a panel data approach. Health Econ. 18, 421436 (2009)
33. Fricke, U.: Arzneimittelinnovationen-Neue Wirkstoffe: 19781999:
Eine Bestandsaufnahme. In: Bausch, J., Klauber, J., Schroder, H.,
Selke, G.W. (eds.) Innovation im Arzneimittelmarkt, pp. 8597.
Springer, Berlin (2000)
34. Kaber, G., Twarock, S.: Beurteilungsverfahren neuer Arznei-
mittel: Eine kritische Bewertung. http://www.uni-duesseldorf.de/
kojda-pharmalehrbuch/FortbildungstelegrammPharmazie/Ser
ieApothekenpraxis/Beurteilung_NAM_FORTE_PHARM2009.
pdf (2009). Accessed 14 Aug 2013
35. Frank, R.G.: Prescription drug prices: why do some pay more
than others do? Health Aff. 20, 115128 (2001)
36. Hajen, L., Paetow, H., Schumacher, H.: Gesundheitsokonomie.
Kohlhammer, Stuttgart (2011)
37. Albrecht, M. et. al.: Ausweitung selektivvertraglicher Versor-
gung: Auswirkungen auf die Gesundheitsversorgung und Anfor-
derungen an den zukunftigen regulatorischen Rahmen. Hans-
Bockler-Stiftung, Dusseldorf (2011)
38. Beubler, E.: Kompendium der Pharmakologie: Gebrauchliche
Arzneimittel in der Praxis. Springer, Wien (2011)
39. Derendorf, H., Schulz, M., Wemhoner, R.: Arzneimittelkunde.
Deutscher Apotheker Verlag, Stuttgart (2011)
40. Lexchin, J.: Do manufacturers of brand-name drugs engage in
price competition? An analysis of introductory prices. CMAJ
174, 11201121 (2006)
41. McBride, D., Kilburg, A., Mauskopf, J.: An Early Assessment of
the German Federal Joint Commission Resolutions of Additional
Clinical Benefit. Poster presented at ISPOR 15th Annual Euro-
pean Congress, Berlin (2012)
42. Wild, L., Forster, L.: AMNOG: pricing reforms in action. Poster
presented at ISPOR 15th Annual European Congress, Berlin
(2012)