International price discrimination: The pharmaceutical industry

~t’orld Dewhprr1mr. Printed

Vol. Il. No. Y. pp. 1111-1150. IYSh.

in Great Britain

1)3)_~73),Wh

S3.M) + o.cn,

~9 IYS6 Pergamon Journals

Ltd.

International Price Discrimination: The Pharmaceutical Industry

FREDERICK

T. SCHUT and PETER A. G. VAN BERGEIJK* Erasmus University, Rotterdarn

Summary. It is generally asserted that price discrimination is a common feature of the international pharmaceutical market. resulting in unnecessarily high medical costs to developing countries. since it is pharmaceuticals that are the largest component of their health care expenditures. However, little comprehensive empirical research has been carried out to test this hypothesis. This article compares the prices of identical packages of pharmaceutical products for 32 countries for the year 1975 and examines which factors contribute to the hu!e price differences. A strong positive relationship between price level and per capita GDP is tound. a 10% increase in per capita income being associated with on average 8% higher drug prices. The implementation of direct price control measures by the government results on average in a 3)“~ price reduction. u hile governmenf policies such as bulk purchasing through a centralized uencrics and. to a lesser extent. excluding patent government agency. promotion of the use of ~ protection seem to be successful in lowering the general price level of pharmaceuticals. These results suggest that the ph;~rmaceutical industry charges what the market “will bar.”

1. INTRODUCTION For years the pharmaceutical industry has been severely criticized for charging too hiph ;I price for its products, prices Lvhich held no relation to “true” scarcity or production costs. “Excessive” profits, absence of significant or “abnormal” price competition. high concentration in production. exceptional expenditure on marketing and promotion as well as price discrepancies for identical products are said to be common features of the pharmaceutical market. Therefore, the pharmaceutical industry is often characterized as highly oligopolistic, effectively exercising market power and charging whatever price the market will bear. For developing countries, which as a group spend some 1% of their Gross Domestic Product (GDP) on ph3rm~!c~,:ticals. an oligopolistic markup implies high opportunity costs due to a loss of foreign exchange nnd investment opportunities foregone. Several authors have suggested policies to reduce these costs for developing countries. However. evidence for the existence of international price discrimination by the pharmnceutical industry and for the effectiveness of the II-II

suggested policies has been based on random studies typically considering only ;I few drugs and a limited number of countries. Systematic empirical research into the causes of the observed price discrepancies is hardly available.’ The purpose of this paper is lo test the hypothesis that the pharmaceutical industry practices international price discrimination. This will be done by means of a cross-country analysis based on prices of packages of pharmaccutlcal products. First. the relationship between structure and pricing behavior of the pharmaceutical industrk will be discussed. Next. the price levels o-f pharmaceuticals in 32 countries will be compared. Finally, the international price discrepancies will be explained by means of a cross-countr) analysis.

‘The authors would like lo thank Dr. W. P. L1. Xl. van dc Ven. Drs. J. G. Vecnbergcn. Prof. Dr. J. &loll and Prof. Dr. f-1. C. Bos for their useful comments.

2. STRUCTL’RE

OFTHE

IYTERN.ATION.AL

Since drugs art‘ by thar very nature rather heterqcnrous (;I gastric ulcer should not he healed Lvith aspirin). the pharmaceutical market can be divided into ;I large number of indqxndent sub-markets (charactcrircd by lo\v cro\s \\ hich corrapond to ela5ticitirs of demand). .A low It‘vt‘l of certain therapeutic cla~cs. concentration for the industr) a ;I \vhc>lc (home S”,, market share for the largest tlrus firm) thus conct’als the real market power which is t‘sercised at sub-market Ie\~cl. \k here “the four firm concentration ratio i5 quite high. (. .) There arc several in the ‘90 percent range. and the unweiphtcd average \V;IS hS pcrccnt.‘.’ The resulting monopolistic power leads to 2 behavior in Lvhich “competitive forces do not play ;I major role in competition is det~rniining prices.“-’ . so that mainly pursued “1; nit‘:in\; of innovation and product differentiation. aipportcd by trademarks and heavy promotion. Bring one of the most innovating industries. the pli;iriii;lc”itic~iI industry is confronted with txtrcnit’ uncertainty on the supply side of the market and seeks to protect itself +inst this phrnomcnon through intensive ilst’ ot patents (protcctin~ the innovation), trademarks (protecting the product against meimprcssivc marketing too-cluplic~ition) and budgets. thus reducing the intlucnce and possibilities of price competition t‘vt’n further. In addition the drug niarkct is characterized by low price elasticity of dcmm~l, due to the intransparency of the market and the nature of the I’ti;lr”i;l~“ltic~II product. which should be considered ;I\ ;I commodity often purchased out of necessity. Moreover, in most dt3eloped countries the decision to buy is taken by ;I doctor who has little or no inccntivc to economize’. since he does not have to pay for the ~oocls he orders and seldom has knowltxlge over the costs of drug\.’ while consumers too art not very cost-conscious bcc;iuse drugs costs are often paid by ;I third party. According to several authors the described structure: of the pharmaceutical industry has caused st‘vert‘ price dixrimination within and among countries, the latter bring the subject of this article. bluller for instance ascertains: ..What is significant is not just the huge variations between the quotations of different companies. but the fact that the same companies asked prices which varied by -100 per cent from one country to another.“’ (See Table I.)

Source:

UNCTi\D

(IWO).

3. TliE COhlPARISOS OF PfIARMACEUTICAL PRICE LEVELS BETWEEN COUNTRIES As ;I co~~qt~~ncc of the high deyree of product diffcrcntiation (kadiri~ to ;I vast amount of often almost identical drugs) and the reluctance of the pharmaceutical inclu>try to publi an important problem.” It is widely recognized that comparing domestic price5 of various countries by simply converting them at market exchange rata into ;I coiiiinoii currency unit is a very unsatisfactor> method, since rschangc rates u~u:~lly do not rctlect the relative purchasinp-paver of the currencies.’ First. the market exchange rate onI> applia to internationally traded good5. so comparing non-traded and nun-tradabk poods. converted at official exchange rata. can produce misleading results. It is indeed a urll-kno\vn fact that relative prices of non-tradabk goods and services tend to be substantially loucr in louincome countries. Second. the official exchange rate is often ;I poor indication of the relative scarcity of a currency. when it is not freeI> convertible against other currencies. In man) developing countries ovexduation of the curattributed to rapid rates of domestic rency. inflation not compensated for by de\nluation. is a common phenomenon. In the case of Ghana. for instance. it has been ascertained that:

INTERNATIONAL

PRICE

DISCRIXIIS:\TION

is not free remarks.

Finally. the big fluctuations in currency values since the early 1970s uould result in unrrasonably large fluctuations in the relative price levels of pharmaceuticals. if prices \vere simply converted at market eschange rates.” Fortunately. the recently published report on the United Nations International Comparisons Project (ICP) by Krnvis. Heston and Summers”’ (19%) offers a unique opportunity to overcome both the problems of availability and comparability of pharmaceutical prices. The ICP provides an alternative yardstick for international price and income comparisons. the so-called PurchasingPower Parity (PPP). A PPP is essentially an international price indes.” In order 10 compare the price levels of the pharmaceuticals between the 37 considered countries” we used the PPP of each country for the category “Drugs and Medical Preparations” to construct price indices \vith rqard to the United States as base-country (Table 7). According to Kravis, -the “organization and execution of the price collection is easily the most resourcr-absorbino ha\e of the work, 21swell as 3 ?I .. the most critical.” For binary comparisons Kravis er (11. (I%?) strictly apply the criteria of common usage and product comparability to select relevani drugs. liowever. if a large number of countries. widely differing with respect to income level and culture. are included in multilateral comparisons (as in this article) an effort to produce a common list of drugs will result either in too short a list or in pricing items that are not representative for all countries. In the case of multilateral comparisons therefore. Kravis Ed (11. followed an alternative approach, inviting each country to choose particular drugs from ;I ILst that is larger than the target number desired. This method produced an incomplete price matrix. The approach followed by Kravis cf crl. in the case of missing prices is a weighted form of Least Squares. the weight of each country being selected so as to correct for the number of drugs with missing price data in that country.“’ Because of the large number of methodological problems and conceptual ambiguities that needed to be resolved to arrive at ;L set of relative international prices. it is not surprising that the study of Kravis PI al. (19%)

In data

our of

from

judpernrnt pharmaceutical

II-!.?

critic&m. ”

the

“Yet.”

reliabilitv products

of

as Theil

the

price

.is prcsunwbly

relatively good: when one compares pharmaceutical commodities with services like health care or education it is clear that the criteria of common usage and product comparability are relativeI> easily applicable to drugs. which can be defined with rather high accuracy, (e.g. by chemical formula or by patent description). In any case. the price indices of Table 2 ought to be regarded as indicative. rather than precise measures of the relative price differences among countries.

4. SCOPE

AND METHOD OF THE COUNTRY STUDY

CKOSS-

In the next section relative price discrepancies in the pharm~iceutic~il world market. as reprcsented by fhe price indices in Table 2, will be explained by mans of a cross-section study in which several demand and institutional variables serve as explanatory variables. The relationship between price level and explanatory variables. expressed in a linear reduced form equation. will be estimated by the Ordinary Least Square (OLS) method. The choice of the explanatory variables is based on those factors of demand and those government policies, which are generally considered to be determinants. for the pricing behavior of the drug industry. Variables describing the supply side of the market are left out of consideration. since the available data are predominantly of an extremely aggregated nature. thereby prohibiting the exposure ot monopolistic power, which is, as stated before, exercised at sub-market level. With regard to our previous discussion on the structure of the pharmaceutical market. the variables are distinfollowing explanatory guished.

Proceeding from the oligopolistic features of the drug industry, which sets prices according lo

Ill-l

WORLD

DEVELOPhlENT

\‘UlUlllC Price index pharmacruticals No.

Country’

01 blalawi 02 03 0-l 05 06 07 OS 09 10

13 I4 15 I6 17 IS

Kenya India Pakistan Sri Lanka Zambia Thailand Philippines South Korea blalavsia

Brazil Mexico Yugoslavia Iran Uruguay Ireland

II Ital) 72 Spain 13 United Kingdom ‘4 Japan 15 Austria 16 Ncthcrl;tnds 77 Belgium 18 France ?Y Luxembourg 30 Denmark 31 Germany. Fed. Rep. 32 United Stata hlcan SD

(0

GDP per capita (index) (GDf’..V)

consumption (index)

((,‘I’) 0.0 1-l

9.30 IO.3 13.0 13.2 20.7 21.5 22.4 Z-l.0 75.2 34.7 36. I 37.7 39.6 12.5 4Y.h 50. I 53,s 55.0

10.4s 7.77 3.x3 -3.-77 _ 0.44 0.57 2.36 s.os 17.0 Y.01

71.1’) s I ss 130.53 137.2Y 101.73 01.56 100.27 157.56

63.Y 6S.J hY.6 75.2 77.7 SI.0 S7.0 SZ.4

Y.Yh 2S.M I.24 I.54 3.4Y 25.l.l 0. IO 0.70

74.00 31.x

6.M 6.56 s.23

83.0

I00 47.SI 28. I6

of

Of

60.S3 50.63 31.71 3X.76 15.22 96.58 JS.01 51.14 35. IO 70.74 4x.07 46. I3 63.S3 hY.hS JS.24 70.4’ 65.05 73.53 s7.25 53.YS hY.01 hY.hS

152.52 100.00

4.90

\‘OlUtIl~

0.(17 IS.66 3.42 0.42 0 05

2.21 0.77 7 70 _._ 0.50

1.Sh 0.2I

Population (index)

(IL’) 1.36 6.37 282.76 v y __. 6.32 Z .3 3 I 0.60 10.70 16.52 5.5s 1I .oo O.Y6 50.17 2x. 16 9.41 15.33

I .30 I .4Y 4.Y-l 15.93 26. I4 16.63

24.29 100 S.81 1s. 13

25.7s 50.27

Indlrcct

per

capita consumption (index)

(C\?.l’) 0.6 1.1 h.6 10.4 6.7 7.’ Il.3 3.‘) 13.3 S .9 1-l.I 72.0 21.6 ‘7.6 40.6 ‘1.3 33.s 3s.o 47.8 50.7 -15.‘) 54.2

Patent protection (dummy) (PP)? I I 0 0 I

I II I 0 I 0 I 0 0 0 0 0

I 0 0 0 0

price control (dummy) (If’C‘Jf

0

0 0 I I 0 0 0 0 0

I 0 I 0 I 0 0 0 I I 0 0

Direct pr,ce control (dummy) (DPc‘)S 0 0

I I I 0 0 0 0 0 0 0 0 0 I 0 0 0 I

I I 0

I

.?S.O 54.7 35.1 74. I 76.0 I0I.S 60.5 2Y.5

I 0 0 I I I

I I

I 0 0 0 0 0 0 0

S3.Y I00

I I

0 I

0 0

0.50 0.50

0.x 0.4s

0.3s

33.Y.i ‘7.49

I 0 0 I I

1 0

0.4s

Sources: Chudnovsky (IYS3). p. ISY: Kravis eral. (192). pp. 10. 17. X)&i-201). 312-213.21&217. Xb711: hlclrosc 153. 157, 1%; O‘Brien (lY79); UNCTAD (1975); UNIDO (1978). pp. l(!-15. 23. (IYSZ). pp. IS’, ‘Two of the 34 ICP countries, Romania and Syria, arc left out sinw data on several explanatory variahlcs arc lacking. +Pf’ = 1. if patents for pharmaceutical products arc recognized; = 0. if not. $fPC = 1, if the govcrnmcnt stimulates price competition by mc‘ans of ;I centralized purchasing policy or promotion of the USC’ of generics: = 0. if not. :DPC = I, if the govcrnmcnt applies strict price controls; = 0. if nut.

INTERZATIONAL

PRICE

“what the market will bear.“” a positive relation between price level of pharmaceuticals (P) and the purchasing-power of a country is expected. where GDP..V (converted by Kravis’s PPP”) serves as a standard. This relation is entorced by the fact that the price consciousness of a consumer vvith great purchasing-power. or. for that matter. of a doctor with relatively many patients with great purchasing-power. w11l probably be low: when income rises. the incentive to ransack the market for the cheapest type of a drus needed diminishes and hence demand becomes more price-inelastic.

(b)

Volrrr,re ofcot~swnprior~

(CV)

Total drug consumption is considered as an indication for the size of the actual market of a specific country. The size of the market is of particular interest when economies of scale can be reaped. Given the extant importance of economies of scale in marketing as well as in research and development to the drug industry and the resultin 1 possibilities to reduce average ‘1 negative relation between overhead costs!’ price level (P) and actual size of the market (0’) is expected.

Attractiveness of a market is not only based on the actual size of the market. but also on potcntiat demand. The number of inhabitants is chosen as a prosy for the number of potential consumers and. consequently, of future sates. Within attractive markets more competition is likely.“’ Hence we expect a negative relation between price level (I’) and population (N).

(d)

Volurw of cotulmptior~

per ccrpittr (CV/N)

This quantity indicates the extent to which a market is developed. An increase in quantity demanded by a consumer. generally implies that the composttion of the consumption outlays becomes more differentiated: in general the need for necessities (i.e. essential drugs) tends to be satisfied first before one passes on to the purchase of luxuries (e.g. tranquilizers). which are of more interest to the pharmaceutical industry. Based on the premise that a vvell-developed market attracts more firms and thus shows more competition. a negative relation between price level (P) and average level of consumption (CVIN) is expected.

II-Ii

DISCRISlIS.-\TION

(c‘) Prltetlrprort,c~ion (PP) Patents are an important way to strengthen market power: the producer of a patented drup is more or less safeguarded against both indigenous and foreign competition and certainly will take advantape of the resulting monopolistic power by commendins a higher price. Hence a positive relation is espected between price Iev,sl (P) and strong patent protection (PP = 1. here defined as recognition of product patents).”

( f) It~tlirccr p-ice cotltrol (I PC) Patent protection is limited to the legal duration of the patent. but its effect can be extended if brand loyalty is built up during the patent period. Statman (1981) concluded for the United States that ..generic and brand name competitors were not successful in capturing a significant market share from the original (patented) brand.“” Reducing the use of trademarks in favor of generics may thus lower barriers to entry and stimulate price competition. Price competition is also stimulated by the establishment of a centralized import policy. which offers a country the possibility to ransack the market by tenders for the cheapest supplies of sufficient quality. Thus a negative relation betvveen price level (P) and government policies that stimulate price competition (IPC = 1) is expected.

(6) Direct price control

DPC)

Finally. oovernment policies aiming to set maximum crices for either ph: tr maceuticals or raw materials are expected to tower the general price level in the drug industry. Consequently, a negative relationship between price level (P) and direct price control policies (DPC = 1) is expected. Summarizing, the following three linear equations are estimated (ordinary teast squares):

P = P (GDP/A’. CVIN, PP, DPC. /PC)

(1)

P = P (GDPIX,

CVIN. CV. PP. DPC. IPC)

(7)

P = P (GDP/h’.

CVIN. N, PP. DPC. /PC).

(3)

With

N priori

expected signs:

dPld(GDPl,L’) dPId(CVIN)

dPldDPC

>

0, dPldCV

0. dPldPP

<

<

0, dPldN

> 0. dPldlPC

<

0,

<

0.

< 0.

Data sources

are to be found

in Table

2.

Il4h

K’ORLD

5.

RESULTS STUDY

OFTHE ON

DE\‘ELOP\IES-l

CROSS-COUNTR\’

PRICE

LE\‘ELS

OF

PHAR.\lACEUTICALS

are reported in Table 3. all abwe 0.M. mdicute ;I firm relationship between dependent and esplanator! wriablss. The signs of the estimated regression coefficients in each of the three equations conform to our predictions. In the dkussion of the resrrssion results \\t’ will concenrratt‘ on the first equation. The

The

empirical

coefficients

rcdts

of determination.

The estimated cwfficicnt of per capita GDP is highly significant (at 1% level) indicating a positive relationship between the price Icvcl of drugs in a country and the pur~h~~sing-pc,\v~r of the population. A IO?,, incrc‘ast‘ af piirch;iGngpower is associattd with about S’!;, higher drug prices on average. This result implies that the pharmaceutical industry is sucwssfully creaming off fhc international cimsurmr surplus and thus supports the widely held view that the pharmaceutical

world

market

The fact that the total drugs in ;L country (0’)

is highly

VOIUIIIC‘

oligopolistic.

(2) (3)

I .J.T$ (6.W) 1.41: (6.4s) I .a; (6.50)

-O.bOS (-7.6s) -0.10 (-1.01) -0.60~ (-2.61)

per

mination therrbv

capita

conwnxd

bc of ampIt

to

of the price suggesting

quantiry

signitknce level

that

of

of

drugs

for the deterph~~rmaceuticals.

;I well-dc~rloped pharattracts more price the effect of an increase consumption (Cl//h’) is

nixuu~ic;rl ,_niarkrt cornpetition.--’ I lowcvcr. in the per apita drug quite mo&rattz (a IO% iiicrt’xc is associated with only about a 3% fall in drug prices). which is consistent with the assertion that competition in the pharm~lcrutic~~l industry does not play a major role in the tlctsrmination of prices and occur4 mainly through innovation and product differentiation.

of concurnetl

(adclcd in the saxmd equation). which stands for the actual mxkt‘t size, has hardly any significant effect on the price level. >ugg!ests that ccononiics of scale iii the pharmaceutical market (mainly concerning

(1)

The seems

-0.20 (-0.6s) -0.001 (-0.01)

(e) Of price in

the

considcrccl

control

lowering

7.0s

- 15.77$

(1.16) 7. I!, (1.17) 7.07 (1.12)

(-2.X)) - 17.w (-2.44) - 15.708 (-2.22)

‘SW Table Z for the meaning of the ued svmbols. ?R’ is the adjusted R’. corrected for dcgrces offrrcdom. 10 the btter. it pcnalizrs the inclusion of \ariablcb tD~not~s significant I test at 1”!4, Iwcl. liDcnottx significant I test at 5% Icvcl.

(DPC.

(;o~~crr~/trcr~rpoli~ic~.s

government

(U/‘c’) the

policies

PP) direct

seems to bc most successful

price

-11.1’ (- I.50) -S.h7 (- l.O
which i\ often prcfcrrcd which csplain very littIc:

IPC.

level

3s.5.y (6.0-l) 37.9.:: (5.N) .3s.56; (5.61)

of

~~harmaceutic~~ls.

I).SI

0.7s

0.2

0.77

O.Sl

0.77

ahovc the R’, because. (i.e. when the r-ratio

contrary < I).

INTERS.ATION.-\L

PRICE

Implsmrntation of direct price control measures implies. on averqs. ;I more than 20”,, decrease in drug prices. Xlthoush the other two government policies - indlrect price control measures ([PC) and escludins pharmaceutical products from patent protection (PP) 3lw appear to influence the price level of drugs in a downward direction (hv about 15 and l(l”0. respectively). their effect *is less convincing because of the rather low significance of thrlr estimated coefficients. This may be partly due to the fact that dummy variables are a very crude measure of the oovernment policies considered. because the c “a II -or-nothing approach” of a dummv variable implies an unavoidable simplification of realitv in which usually a broad range of measures within a policy can be discerned. In the cake of patent protection, for example. not only the exclusion from patentability of pharmaceutical products may be of importance for determining the price level of drugs. but also the exclusion from patent protection of pharmnceutical production processes (although considerably Lveaker patents). the duration of the granted patent protection and the possibilities to evade the patent law. Ilowever. the modest effect of excluding patent protection on the price level of pharmaceuticals. as set‘n before, supports Chudnovsky‘s expcctation. when he arpues:

Chudnovsky therefore considers a mere in patent legislation insufficient to modify the structure of the pharmaceutical industry in developing countries or to reduce the prices of l,h~irmaccutical pr0duct.s and imports.” Instead he suggests that ;I combined policy of excluding patent protection and promoting the use of tlenerics (or reducing the Iqal protection of c trademarks) will he more successful in reducing the price level of drugs.“’ Unfortunately, howt‘vcr. policies on patents and trademarks have not yet been applied together. so there exists no concrete evidence to evaluate this hypothesis. Nevertheless. the supposition is supported b> Pate1 (IYS3) who estimated up to 60% potential savings in total drugs expenditures through the implementation of ;I combination of five policy options.” Indeed. Sri Lanka has proved that a comprehensive drug polic): can tower price level of drugs dramatically. Durlnp the period 1977-76 Sri Lanka reformed the structure of production, change

DISCRI\IIS.-\I-ION

1147

importation and distribution of pharmaceuticals. The introduction of a svbtem of centralized purchaslns of imports ot’ finkhcd drug and intermediate chemical5 (tc) aLoiJ tramfer pricing) by compctiti\s tender through the State Pharmaceutical Corporation (SPC) achieved a -1O”,, overall savinp in just one year (some TNCs - Hoechst and Beecham - reduced their prices by 51k70’:0 from one vear to the next).‘” This centralized import poiicv in combination with other policv measures &s responsible for the fact that S;i Lanka had by far the Iou.est price level of druq in 1975 amongst the 33 countries under discuxkn (Table 2).

6. DISCUSSION sonsitivitv of the price level of The pharmaceuticals kith respect to the purchasingpower of the population and the implementatiorl of ;I national drug policy. strongly supports the assertion that international price discrimination is ;I common feature of the pharmaceutical world market. Of course. international price discrimination often results in lower prices for poorer countries. because of their lack of sufficient purchasilig-po\ver. f lowever. the fact that drug prices vary arbitrarily. depending on the esistence and degree of success of ii national drug policy,“’ implies much higher prices than nt‘cessary for many poor countries with little or no possibilities to establish ;I successful drug policy. Although the drug prices in developing countries are often lower than in developed countries, the real costs of these products, relative to the purchasing-power of the population. are COW siderably higher (the real costs of drugs in klalawi art‘ more than 12 times higher than in the United States 2nd even Sri Lanka experiences about 00% higher real costs). Moreover. 40% of the total health care expenditures in developing countries \vith

consists

only

thereby, other

putting

wtal

in relation importance

UNIDO

about health

of pharmaceuticals. 8%

in

developed

a relatively

to the production to

(197s)

severe

care needs. developing

compared countries?’ constraint

Hcncr costs

a fair

on price

is of extreme

countries

or.

as

states:

The ohjcctlvc of devcluping ruuntricb should bc tu set the lowest puaiblc prices Cur the deirrd number of drug5 cunsistcnt with rhc cncuuragsmcnt ul production and rclcvant rcsearch. This objective cannot lx achieved with a free rn;dx~. which allo\vs too many drugs to bc wld. with cnormuus price variations on identical products (and so a large rent accruing to mow heavily promoted. brand-named or patent-pruttxtcd products). profits often in

11-s

\VORLD

7.

CONCLKDINC

DE\‘ELOPIIEN-f

REXIXRKS

Our empirical wearch. based on the data of the International Comparison Project (ICP). provides evidence for th e esistcnce of considerable price discrimination in the pharmaceutical world market. Drus prices set’m to be highlq influenced hk the purchasinp-pwvtr of ;I countrv ;I 1O”;, incrrae in per capita GDP being amxY;cted and

by

tion

art‘

with

on

;iwrage

S”;,

higher

prices

-

the implementation of ;I successful national drug policy (e.s. direct price control mt’;wrcs taken bv the povcrnmcnt result on average in a 3)‘:1e price reduction). This perceivtxi price discrimin~~tion implies ;I severe burden to poor countries with little institutional and market power. The method usd to explain the price kvel of drugs. which is based on aggregated rlatn. means. of course. ;i simplification of the compks nature of the pliarrn~ic~uti~~ll world mxkct. I lcncc the results of our study should be interpreted with caution. Bearing this rewrvation in mind. the established price discrimination points out to dcvcloping countries the nwd for ;i rational drug policy. especially when future trends in drug consump-

Our finding support the i&n that the appl~cation of price control policies. especially those which dirrctlv rrpulatr the price It’\el of drugs. with an exclusion of possibly in combination (some) drug5 from patent protection. might be succtxful to rc’verse this trend. Of c‘ourse price reduction is onlv one side of the coin. For a successful conta;nment of pharmaceutical t’yprnditures ;I more rational purchase of drug5 i\ also imperative. In IYSZ. 5 years aftr’r the WHO produced the first mo&l list of essential drugs.” 70 countries already had restricted lists of chst’ntial drugs based on local need~.~’ hlany of thew lists. however, represent little more than poliq w;~l~ and art: all intendd tc> applv onlv to the > public sector (with the esceptions bf Sri Lanka. Mozambique. Afghanistan and Bangladesh which have also taken mt’asurt’s to restrict the private market to ;I selection of essential drugs). Drug policies for developing countries should focus on the provision of essential drugs dt to Mt2lrose rt2asonablc prices, hut. according (lYS3): Thctr adoption tlcpcn& on rhc political will of Third World gowrnmcnts to put the health ncd\ of the majority hcforc the commercial interests of A minority (and) on aclivc support from govcrnmcnt\ of the major drug-producing nations and a rcatllns\\ amongst manufacturers to allopt more farsight markcling politics.”

concernccl:

Total pharnl;lc~utic;II cspcndirurc by dcvcloping countries incrsilw5 hy ;llmos1 four time. from IYSI 10 the year 2000. It incrcascs from USSIY hillion, 111 IYSI. to over USS7l billion in the year WOO (in IYSI

I. So far, the onlv comprchcnsive empirical rexarch into diffcrcnccs in pharmaceutical prices bctwccn countries, has twcn carrid out by the BEUC (Bureau EuropAan dcs Unions dc Con~ommat~llrs). The BEUC Report (I9S-i) comp;ira the price Icvc’i of drugs bctwecn scvcn EC-countries (i.c. Belgium. Fcdrrnl Republic of Germany, France. Ireland. Italy. the Ncthcrland~ and the United Kingdom). In or&r to do w, the BEUC sclcctcd S-l drupb, togcthcr representing at ICM N’X, of the ph;lrmaccuticaI ~lcs value in each country. The BEUC Report cstahlished huge priw discwpancics. cvr‘n within such a rclativcly homugc‘ncous croup of cconomic;illy coopcratine: countries. For Ice‘than 20’1: of fhc SJ considcrcd drugs (hc dlffcrcncc bctuwn the loa.cst and highest price is Icss than lW%. The gcncral price Icvcl of drugs is about two timcs as high in rhc most cxpcnsivt: country (Germany) as compared IO the cheapest country (France). See BEUC (1YS-l. pp. 3OY-375).

2.

Fclcl~~cin (IYS3).

3.

.hlotc C’I d.

(1077).

4.

Lull

pp.

5.

hlullcr

(1971).

(IYSZ),

p. -MY.

p. l36Y

IN-151 p. SS

6. The comparison might further bc complicated b) the fact that usually three partIc‘s arc in\olvcd in the pricing of :i drug: th: manufk[urcr. the aholesakr and the chemist. However. the BEUC Report concludc~ that the different margins for wholcdcr and chemist In each country arc of minor importance in cuplaining the pcrceivcd price discrepancies hctwwn countries (BEUC. IYSJ, pp. 316-320. 314). 7. Comparing manufacturer’s prices of cipht knding multinational pharmaceutical firms brrucen Europe.

INTERN.ATION.AL

PRICE

United State> and Japan. Rcrkie (IYS?. p. 176) remarks that “the simple purchasing-power of money in its domestic environment duo not nrcess.xtiy explain it5 furcign cxchanpe baluc.” IIt tries to wercoms thlr problem by rvsmlning the price5 in trrmz of minutes of work rcqulred to purchase the products. b’hethsr this method rckxs the rc’lativc purchabins-pwer of the population rather than only that of the wxlrl;erb in the the

pharmaceutical industry the case of developin: S.

Barnctt

sr‘cms dubious. countries.

er rrl. (IYSO).

czpeci,llly

in

p. 4X1.

9. Rrrkie shows how wnsttivc international price comparisons are to exchange rate Fluctuations by comparing the same domestic drug prices convcrtcd at the ruling eschangc rates in IYSII and IYSI. The appreciation of the yen and the dcprcciation of the European currt’ncit’s (except the pound sterling) against thr US dollar causes an incrrnsc in the pharmaceutical price level of Japan rt’lativc to that of Europe of nearly 35%. when prices arc measured in US dollars (Reckie. 1’353. pp. 17Y-180).

Exchange rate movements will probably lx of minor inllucnce on the results of the BEUC Report because of the European Monetary System countries only the United Kingdom in the EMS).

(of the considered did not participate

10. Kravis L’I (I(. (10X7). The IYSZ volume reports on phase III of the ICI’. covering dota collsctcd from 34 countries for the year lY75. (In phac I of the ICP IO

countries wcrc considered. which were eutcndcd to I6 in phase II.)

DISCRI\lIN,ATION

16.

The11 (IvS.3).

17.

Xlrlro5r

IS.

U\ing

11-I’)

p. 515.

(IYS2J.

p

cxch,insc

-IS.

riltc\

tend\

to

o\r‘rstcltc

the

puvcrtv of poor nationa ~W;IUW of the fact that services. \+hich arc chrapcr rclativc to commodities in arc mudrstly rcprescnted in intcrother countric\. national trade and hcncc in cxchangc rate. Bcsidcs. the volatility of many csas mentioned hcforr, chanec rates makes them unluitahls 3s convcrbion factors. For IY7.i. one of the mat scnsiti\c comparisons among the 34 sampled countries is between the per capita GDPs for Denmark and India: by the ICP measure Denmark‘s per capita GDP is about I.! timch higher than India’s, whcrcas Denmark’s cuchange r;ltc converted per capita GDP U’;IS more than 50 timcb as high as that of India. 19.

Lali

(1974).

pp.

lJf%I47.

Muller (1982). p. 87: “Indeed. the first impression 20. in many countries. especially those. which by virtue of their population or wealth arc regarded as Important future markets. is one of chaotic csccss of competition.” Since established production processes arc easily 21. to be circumvented in the drug industry. process patents (which only protect the process of production) arc inferior to product patents. which protect the formulation of the drug. 22.

Statman

(IYSI),

p.

115.

I I.

Starting with data on prices I>,, in domestic currency units and quantities c,,, for ii list of products i and countricsj. the procedure involves solving ;I system of equations for an international price X, for each product and simultaneously a purchasing power parity l’l’l’, for each country’5 currency. l’l’f’, equa1.sthe ratio of the cost of country i’-\ total bill of goods at national price> to the cost at international prices and ,U, is a quantity wcightcd sum of purchasing power adjusted prices. 12. Two ICI’ countries. Romania out bccaux of imufficicnt data cuplanatory variahlcs. 13.

Kravis

(IYSJ).

p.

and Syria. concerning

wcrc‘ left several

See e.g.

Marris

24.

25. for

Chudnovsky

(IYS3).

p.

IYO.,

This view is shared by several other instance: Lall (lY7-1). pp. lSY-160.

26.

Chudnovsky

(IYS3).

27.

PlllCl

2s.

1.211 er trl. (lY77).

p.

authors,

SW

193.

I?.

14. The BEUC Report (19%) trtcd to solve the problem of absent prices in ii diffcrcnt (less satisfactory) way. The sum of the prices (in ECU) of the uvallablc drugs in the considcrcd country is compared \vith the sum of the average prices of the same drugs in all countries. The price of the ahscnt drug in the considered country is then obtained by assuming that it is equal to the average price kvc’l of that drug in the other countries corrcctcd for the deviation bctwccn the two price totals. 15.

This straightforward interprctxtkm howcvor must 23. bc considcrrd with some caution hccausc of the rothcr high collinearity bctwccn CVlA’ and C‘L)P/IV. The R’ bctwcen CV/N and GDP/N varies bctwcen 0.67 and 0.75 for the three concerned equations. For the rest. the collinearity among the regressors is very low, the R’ being in all CIISCSbelow 0.X.

(IYS-I).

(IYH3).

pp.

IYS-204.

Their empirical analyses Ied the BEUC ah well as 29. Rtxkic to .\imilar conclusions. The fir&t carcfully stated that “it is more than clear that absent or minor price control vy the government certainly doa not Icad to the provision of the cheapest drugs” (BEUC. IYSJ. p. 345); whereas the latter argues that the consistently higher drug prices in Japan seem to be the rault of the fact “that price controls arc in many countries strict and rigid except in Japan” (Recbie. lYS3. p. ISO).

1 Ii0

U’ORLD

30.

Patcl (1’)s.:).

?I

l_.NIDO

33

Pate1 ( IYS3).

p.

(lY7S).

DE’.‘ELOP\IEX-I

196. TdhlC I

33.

\VHO

(1977).

p. 2.;.

34.

\i’HO

( IYS2).

.3S

>lclros~

p. 3HI.

I IYS3).

p. 30. p.

IS5

REFERENCES

Patel. hl. S.. .‘Drug costs in dcvelopins countries and policie to reduce them.” It’orki L)L’\ doprrwrrf. i.01. I 1. No. 3 (IYS3). pp. I’).%70-b. Reckic. W. D., .‘Pritx comparisons oi Identical products in Jap;m. the Unitcd States and Europe,” in .I. Finsinger. (Ed.). Ecwrfmic Atwl~.~i~ of Rqd~~frrf lCltrrkcl.s (London and Basingstoke: Mxxiillan. IYS3). Statman. hf.. “The Ellcct of patent espirntion on the market position of drugs.” in R. Helmj. Drys crrrtl /s.src~~s trrrtl Polic: Ol)fcctir~cr tlcnlrh: Ecotrorrrit (Washington D.C.: American Entcrprlse lnstitutc. IXSI). Thcil. f-1.. “World product and incomi’: A rcvIct$ oJ f’olirictrl Ecotwv~~. \‘ol. Y I. No. 3 iirticlc.” Jowrd (I’M). pp. 505-517. UNI DO, “The growth 01‘ the pharmaceurical industrb countries: Prohlcni~ and prosptxts” in developing (New York: United Nation>. lY7S). UNCTr\D. ‘.hlonographic study of the pharmaceutical industry” (Geneva: UNCTAD. lY711 UNCTAD. .‘Economic cooperation among dewloping countries in pharmacc~lticals.” Thr APEC/TTIExperwncc, CAlUI’HIWPl7 (Gcncva: UNCTAD. I’M)). WIIO. “The sclcction of csxntial &up\.” Technical Report Series 615 (Gcncva: WHO. lY77). wr 10. “National politics and prixtice in rsgnrd to nicdic;il products and rclatcd intsrnJ[ion;d prohkm\.” Background Document (Gcncva: WHO, March lY7S). WtIO. “Action programmc on essential drugs.” Kcport hy the Euccutivc Board Ad Hoc Committee on Drug Politics on Bchalt’ ol ths Escativc Bo.lrd (A?-7). Thirty-Fifth World Hcdth Asscmhl> April IYS1). (G c‘neva: WHO,