LIBERALISATION, FIRM SIZE AND R&D PERFORMANCE: A FIRM LEVEL STUDY OF INDIAN PHARMACEUTICAL INDUSTRY JayaPrakashPradhan In the presentpaper, an attempt is made to empirically verfu the impact of economicliberalisationonthe R&D hehaviourofIndianpharmaceuticalfirms,controllingforthe fficts of severaL firm specificcharacteristicsincludingfirm size.The resultsfrom the Tobit analysisfor a iampte of Jirms over the period 1989-90to 2000-01indicate that competitivepressuregeneratetlbl,tiberalisation has worked eJfectivelyin pushing Indian pharmaceuticalfirmsinto R&D activiy-.A host oJ Jirm characteristicslike age, size,profitability, intangibLeassets,erport orientation ancl outwartl foreigndirect investmentof thefirmare alsofoundto be importantdeterminantsof innovativeactivi6, inthe industry.Thestudysuggestsseverolpolicy measurestofurther incligenoustechnologicalefforts of pharmaceuticalfirms, which includeremovingobstaclesthat inhibit outn-ardorientationoJfirms, providing special schemefor small sizefirms in the overall technologypolicy for the iniustn,, intensifiing collctborativeresearchefforts betweenprivate sectorsand governmentresearch instttution,and utilisirtgflexibilities in the TRIMsagreementsto persuade foreign firms to relocatethetr R&D units into the countrN. I. Introduction

asenvisagedin thePatentAct1910 was a turning point in the growth of indigenouspharmaceutical India's pharmaceuticalindustry today stands industry.The provisionsof processpatentswith amongthetechnologically mostvibrantsegments a maximum duration of patenting reduced to of Indianmanufacturing. It is well understood in sevenyearsand the compulsorylicensingafter the literaturethat the level of growth and tech- three years from the time of grant of the patent nologicaldevelopmentexhibitedby the industry had boostedlocal innovation,mainly in process rs a successof strategicpolicy interventions and formulationdevelopment.2 The availability consciouslyundertaken sincelate 1960swith the oflife savingandotherdrugsin India at a fraction specific objectsof self-sufficiencyin drugs pro- of the pricesprevailinginternationallyand sigduction, self-reliancein drugs technology and nificantly at a lower time gap betweenits introaccessibilityof quality drugs at reasonable ductionin the domesticmarketand introduction prices.rTheseinterventionsincludedencourag- in the world market underscorethe successof ing indigenous production and technological favorablepolicy interventions.s At the dawn of developments throughlocal contentand linkage Independence, the industry hardly had any techrequirements, incentivesto local R&D, encour- nologicalbaseto start local productionand was aging genericsover brandedproducts,subsidis- only processing imported bulk drugs into ing small-scale sectors, Drug Prices Control formulations. By the eightiesthe industry had Order (DPCO) and containing the activities of accumulatedtechnologicalcapabilityto produce foreign multinational enterprises (MNEs) bulk drugs from as basic a stageas possibleand through Foreign Exchange Regulation Act achieveda high degreeof self-sufficiency con(FERA)anddiscriminatory licensingsystem.The cerningrequirements of basicraw materialsand softIntellectualPropertyProtection(IPR)regime intermediates. This risingdomestictechnological Jaya PrakashPradhanis ConsultantResearchand InformationSystemfor DevelopingCountries(RIS), Zone 48, India H a b i t a tc e n t r e ,L o d i R o a d ,N e w D e l h i - l 1 0 0 0 3 ,I n d i aT e l : + 9 1 - l t - 2 4 6 s 2 1 7 5 12 1 7 7 ; F a x 4 s s z g n q Email : pradhanjayaprakash @hotmail.com ^ Acknowledgements:The author thanks Researchand Information.systemfor the Non-alignedand Other Developrng countries (RIS), New Delhi, for giving_accessto the clatabase usedin the"presentstudy. tn pani.:utati *""iJ-iL"i" i.L.o my slncereappreclationto Dr. NageshKumar, Director Ceneral,RIS, for his jnvaluableguidance,supportand inspiruii,onrn preparati.on of the paper.I am also thankful to Vinoj Abraham for many useful discussi6nsand AIka'Chadhafor assistausc in cornpilingthe data set.The authorgratefullyacknowledgesthe valuablecommentsof an anonymousreferee of theloumal. However,the author aloneis responsiblefor views expresledin the paper.

JOURNAL OF INDIAN SCHOOL OF POLITICAL ECONOMY

capabllity in the industry is also reflectedin the favorable trade balance that the country is enjoyingin pharmaceuticalproductssincethelate eightiesascomparedto thehugedeficitsof sixties and seventies.

OCT-DEC2OO2

size, age, knowledgeacquisitionfrom abroad, export orientation,outward investment,multinational affiliations, etc., which literature on R&D hadidentifiedas importantdeterminantsof R&D behaviourat the firm level. The main purpose of such a quantitativeanalysisis to derive some strategicpolicy options that can help to strengthenthe technological life-blood of the industryto maintainits competitiveness in a liberalisingregime coupled with product patent system.

However,as a part of the ongoingeconomic reforms,many of the favorablepolicies that had nurturedthis industry through the decadesafter Independenceare radically changing. TRIPs agreementsseek to completelyunderminethe existing processpatent regime - the heart of growth impetusof the industry.The country has The paperis structuredas fbllows: SectionII a 10-yeartransitionperiodto implementa 20-year presentsrecent trends and patternsof R&D in patentprotectionfor an innovation,irrespective Indian pharmaceuticalindustry. SectionIII forof the fact thattheproductis locally manufactured mulates the empirical framework and the or imported. With the amendmentsof Indian hypotheses on thedeterminants of R&D activity. PatentAct,1970in December1999in Parliament. It also discussesmethodologicalissues.The the mechanismof exclusivemarketing rights empiricalresultsand discussionarepresented in (EMRs) and a mailbox system of accepting Section IV. Section V provides concluding product patent are alreadyin place as transitory remarkswith some underlyingpolicy implicameasuresto shift to the productpatentregime.As tions. per rule, Indian companieswill not be able to reverseengineerany patentedproductin the post II. R&D Activity in Indian Pharmaceutical 2005 scenario.Even though they have the free- Industn-: RecentTrendsand Patterns domto do soin thecaseof all molecules registered until December1994,their scopefor adaptations R&D activity in Indian pharmaceutical and process developmentswill progressively industry has increasedsubstantiallyin the latter reducein the long run. Therefore,this emerging half of the nineties,both in absoluteamountin policy regimehassignificantimplicationsfor the rupees spent and as a proportion of the total futuretechnologicaldevelopmentin the industry. turnover.The estimatedR&D expenditureby the samplefirms hasrisen from mere Rs 8 croresin The pharmaceuticalindustryis a researchand 1990to an impressivefigure of Rs 515 crorein developmentintensive industry. Therefore,a 2001 (Table-l). The trend in R&D intensity continuousflow of R&D effortsis essentialfbr indicatethatthe samplefirms havespentaround the developmentof pharmaceuticalindustry. 2 . 2 p e rc e n t o ft h e i r s a l e isn 2 0 0 l a s c o m p a r e d t o Againstthebackdropof therecentpolicy reforms, 0,2per centin 1990.In termsof R&D intensity, the most important question is, how has the the performanceof foreign firms is, however, indigenoustechnologicalactivity of the industry observedto be contraryto the expectation,combeen affected by the new policy regime. The pared to domestic firms. The observed R&D primary objective of the present study is to intensityof domesticfirms, 2.6 per cent,is three empiricallyexaminethe impactof liberalisation and half times higher than that of foreign firms, on the innovativeactivity in Indianpharmaceu- which is low at 0.74percent.The R&D intensity tical industry. It will also analysethe role of curveof thedomesticfirms is lying continuously several firm-specific characteristicslike firm abovethesampleaverage sinceI 994andhasbeen

voL. t4 No. 1

LIBERALIZATION, FIRM SIZE AND R&D PERFORMANCE

more or less rising (Figure-l), while that of foreign firms is continually lying below the sample average after 1994 and appear to be decliningsince 1997.

firm-specifictechnology.The counrryhad bitter experience with thePatentandDesignsAct, 191I wherea strongpatentregimeled foreign firms to merelyengagein tradingactivitiesby processing imported bulk drugs into formulations and virThe advocatesof strictpatentregimegenerally tually holding back indigenousefforts towards argued that product patent would lead to an t e c h n o l o g i csael l fs u f f i c i e n c y .Eam p i r i c asl t u d i e s increasein the internationaltechnologytransfer on the relationshipbetweenpatentprotectionand to India by encouragingforeignfirms to introduce locationof R&D activity by MNCs fails to derecr theirnew productsandrelocatingtheirR&D units any significantcorrelationin the caseof develinto the countrybecauseof its cheappersonnel oping countries.5 Therefbre, the low R&D costs.The trends in R&D intensity,however, appearto be not supportiveof this view. Foreign intensityof foreignfirms ascomparedto domestic firms, given their captiveaccessto the laborato- firms should not surpriseus. Nor should we ries of their parents,areincurringminimal R&D expectthat their R&D intensityis going ro be expenditurein the nature of local adaptationof changed substantiallyafter the product patent theirproductin thecountry.This is in accordance regime comesinto force. Given their monopoly with the trend in R&D activity of MNEs to be statusenjoyedunderTRIPsandalsotheprovision concentratedin the home country becauseof the that imports of the product is akin to local proeconomies of scale in innovative aotivities. duction,the hopeon fbreignfirms as a sourceof agglomeration economies,and a needto protect R&D activitymay be unrealistic. Table 1. R&D Intensity in Indian Pharmaceutical Industry, 1990-2001

SampleFirms

DomesticFirms

Foreign Firms

R&D N u m Inten- ber of s l t y firms (Vol

(l)

Num- NumVa R & D R & D N u m - N u m % R&D ber of ber of Share (in Rs Inten- ber of ber of Share (in Rs flrms firms of crore) sity firms firms of crore) (Vo) rncur- R&D incur- R&D ring firrns ring firms R&D R&D (2) (3) (4) (5) (6) (1) (8) (e) ( l 0 )

1990 l99l 1992 1993 1994

6t 82 101 124 n5

4 6 2l

0.12 0.02 0.l8 o.74 |.49

l6

1995 1996 t997 1998 t999

18 t9 t9 19 2l

l5 I6 l6

3 3 . 5 264

I. 8 0 r.55 1.80 t.57 1.75

2000 2001

3 4 . 1 305 3 7 . 4 419

1.83 2.60

2l t7

t 3 6l.9 l 3 76.5

67

6.6 8 7.3 5 20.8 13 31.9 51 354 |13

0.20 0 . ll 0.21 0.77 t.23

452 654 84 16 106 33 r57 50

4.4 6.2 l9.0 3l.l 318

215 234 221 220 221

19 90 94 85 82

36.7 38.5 42.s 38.6 3 7. l

1.48 1.38 1.59 1.39 1.50

t91 2t5 202 201 200

64 14 18 69 67

.tl.)

229 188

84 17

36.7 340 I .55 208 4 l . 0 5 1 5 2 . 2 1 t'7|

'7 | 64

t74 t92 260 248 298

3 I 7 't:) 90 149

34.4 t62 38.6 )14 2t0

(l l)

Num7o R&D R&D ber of Share ( r n R s I n t e n finns of crore) slty (7o) lncur- R & D n n g flrms R&D (12) 03) (14) (15) (16)

11

t'7 l8 l8

2 125 2 ll.8 5 294 1 1 7 7. 8 t2 66.1

t6 t5

83 84 84 84 '7 t.4

5 5 o

22

25 30 36 38

0.30 0.25 0.26 0.84 0.13

1:)

0.72 088 091 0.86 0.71

34 36

0.65 0.'74

NoIe: Data are for flscal year ending March 3 I of the year shownSource:Authors' computationbasedon RIS-DSIR database (2002\

r

JOURNAL OF INDIAN SCHOOL OF POLITICAL ECONOMY

OCT-DEC.2OO2

Figure l R&D in Indian Pharmaceutical Industrv. 1990-2001

C)

Io

qoo

o

g

s

o) ! eoo

)

1q'6

E c o) X o

o

c)

-lno- -

1, t€

.= o

od E.

--|FDomestic Firms

-tsForeign Firms

Note: Bars representR&D expenditure;Iines representR&D intensity.

It is encouragingto find that R&D intensityof CSIR, a public fundedresearchinstitution.6 the industry has risen substantiallyin the latter part of the nineties.However, it is very low Table 2 gives the distributionof firms over comparedto the cxisting internationallevel of different size classes of R&D intensity in l 0 - 1 5 p e r c e n to f s a l e s .T h e f a c tt h a to n l y o n e - 1999-2000(also see Figure 2). The number of thrrds of sample firms were incurring R&D firms is unevenly distributed across different expenses in the industryneedsattention.Further, classeswith a strong concentrationin the lower most of the researcheffbrts are confined to pro- end.Thereare 139 firms in the industrywho did cess improvementsand, to a limited extent, to notundertake anyR&D activity(0.0-0.0size)and researchon drug dclivery system.Barring a few another47 firms who engagedin R&D but it flrms, the industryhas not yet madeprogressin amounted to lessthan I percentof theirtotaisales channelingresearchactivity into basic research ( 0 . 0 - 1 . 0s i z e ) . O n l y i n c a s e o f l 6 f i r m s t h e wherein the goal is to invent new drugs. The observed R&D intensity was found to be resource constraints appear to explain this respectableat 3 per cent and above.Therefore, inability of private sectorfirms to meet the huge the pattern of R&D activity in Indian pharmacost entailedin developinga new drug. This is ceuticalindustryrevealsthatmajority of firms do clear from the fact that from Independenceto not engage in innovative activities and the 2001, only 14 new drugshavebeendevelopedin majority of thoseengagedspentvery small prothe country out of which 11 have come from portionof their sales.

voL. t1 No. 1

LIBERALIZATION, FIRM SIZE AND R&D PERFORMANCE

Table 2. Distribution of Firms According to R&D Intensity, 1999-2000 R & D I n t e n s i t y{ 7 c ) (l)

00-00 00-l0 1.0-3.0 3.0-5.0 5.0-above

Number of Firms (2)

Per cent (3)

CumulativePer cent (4)

r39

62.3 2t.l 9.4 4.0 3.1

62.3 834 928 96.9 100.0

/1

2l 9 7

Source:Authors'computation basedon RIS-DSIR database(2002). Figure 2. Distribution of Firms based on R&D intensity in Indian Pharmaceutical Industry, 1999-2000

160 140

120

o tr .! 1oo

40 rdU o

o an9 0)

z

o-

40

20

20

10 0.0-0.0

0.0-1.0

1.0-3.0 R&Dintensity(7d

3.0-5.0

5.0 - above

A list of twenty firms, with largest R&D strong presencein antibioticsand analgesics. expenses incurredduringtheperiodI 999-2000is Even though the company stood second in given in Table 3. Ranbaxy LaboratoriesLtd., absoluteamountof R&D, it is at the top considspent around Rs 55 crore in R&D activity and ering R&D expensesin relation to sales.There ranksat thetop. It is oneof the few researchbased areonly two foreignfirms, namelyNovartisIndia domesticpharmaceutical companiesdriving the Ltd. and GlaxosmithklinePharmaceuticals Ltd. competitiveness of the industryin international whichmakeintothelist by virtueof theirabsolute marketwith subsidiaries in more than 20 coun- amountof R&D expenditure. It is importantto triesacrossthe globc.The companyhasa strong notethatthesetwo foreignfirmsspentsubstantial presencern the anti-infectivesegmentwith l2 amounton R&D in absoluteterms;but it is in fact brandsin the top 250 in the domesticmarket.The very nominalin termsof R&D intensity.These Indian companythat has rankedsecondin terms two firms stood last in the rank basedon R&D of R&D expensesis WockhardtLtd.; it has very intensity.

JOURNAL OI' INDIAN SCHOOL OF POL]TICAI, ECONOMY

UCT.DEC 2UO1

Table 3. List of Twenty Firms with Highest R&D Expenses,1999-2000 Name of company

Ownership

R&D

(2)

R&D (in Rs Crores) (3)

RanbaxyLaboratoriesLtd. WockhardtLtd. Cipla Ltd. Cadila HealthcareLtd. S u n P h a r m a c e u t i c Ianl d s .L r d .

Domestic Domestic Domestic Domestic Domestic

55.39 40.25 30.02 2t.27 18.8

I 2 3 4 5

2.93 1.21 3.89 4.45 3.92

Aurobindo PharmaLtd. USVLtd. Pfizer Ltd. D r . R e d d ys L r b o r r t o r i e sL t d . PanaceaBiotec Ltd.

Domestic Dornestic Domestic Domestic Domestic

t4.34 14.23 t3.47 13.27 12.59

6 7 U 9 l0

1.92 5.93 4.08 2.69 6.48

ll

Lupin LaboratoriesLtd. IMergedl Nicholas Piramal India Ltd. N o v a r t i sI n d i aL t d . Ipca LaboratoriesLtd. Glenmark Pharmaceuticals Ltd.

Domestic Domestic Foreign Domestic Domestic

9.32 9.26 6.17 6.44 :\.2

ll l2 l3 t4 l5

t . 1| 1.89 0.81 l.75 3 . 58

l5 t2 t9 13.5 6

Orchid Chemicals& Pharrnaceuticals Ltd GlaxosmithklinePhannaceuticals Ltd. Unichem LaboratoriesLtd. Cheminor Drugs Ltd. IMerged] R P G Lif-eSciencesLtd.

Domestic Foreign Dornestic Domestic Domestic

4.54 4.t4 3.62 3.51 3.1

l6 t7 l8 l9 20

|.26 0.43 | .15 1.53 t.47

18 20 r3 . 5 t6 t7

(l)

intensity (5)

(b)

9 I 7 4 6

3 5 l0 2

Source:Author's computationbasedon RIS-DSIR database(2002).

III. Determinants of R&D F ramework and Ht,potheses

behaviour: The and uiris a normally distributederror term.

The important reasonfor estimatinga Tobit The R&D behaviour of a firm is generally modelis thefact thatthedependenrvariableR&D conceptualised into two importantdecisionsthat intensity takes on the value of zero for a large it hasto make: ( I ) whetherit will engagein R&D proportionof casesand hencesimple OLS estiactivityor not,and,if yes,(2) how muchresource it will devoteto this.The first questionboilsdown mationwill producebiasedestimate.As thereare to estimatingthe probabilityto do R&D and the two types of effects associated with each variablein theTobit model- ( I ) the secondone to estimatingR&D intensityregres- independent on the valueof R&D intensityfor casesat sion.In this casetheobviouschoiceis to estimate eff'ects a Tobit modelfor Indianpharmaceutical firms of thelimit value(i.e.,zero)and(2) anotherfor cases the following form: abovethe limit, the singleordinaryTobit coeffiR&D,, = X,,B+ u,, -0

ifX,,B+ u,,> 0 ifX,,B+u,<0

(l.l)

cient is not directly interpretable.Researchers often make mistake by interpreting Tobit coefficientsas the eff'ectsof independentvariableson the dependentvariablefor casesabove

WhereX,, is a vectorof k (k= 1...k)factorsthat t h el i m i t .M c D o n a l da n dM o f f i t t ' s [ 1 9 8 0 p , . 318] decomposition is therefbrehighly usefuldue to expiain the R&D intensity (R&Di,) of irh ( i = 1 . . . 2 2 1 ) f i r m i n t t h t i m e ( t = 1 9 8 9 - the fact that it disaggregatesTobit effects into 90...2000-01). B is thevectorof Tobit coefficients thesetwo types of effects:

voL.t1 No.4

LIBERALIZATION, FIRM SIZE AND R&D PERFORMANCE

D E ( R &=D-r ;t ' tlla e tn a o 't) a\ D\ )

positive relationshipbetween firm size and innovation. Larger the firm size the larger its market power and larger its capacity to appro- (aprrt) priateeconomicrentfrom innovativeactivity.By +E(R&D,[ ...(1.2) nature R&D activities involve huge financial Ur,_,J resources,contain considerablerisks and the ILall, 1992].Firm size. Where F(z) is the cumulativenormal distri- outcomeis unpredictable which is proxy considered to fbr theresource base bution function fbr the proportionof casesabove of the flrm, risk perception and scale economies, the limit. E(R&D) is the expectedvalueof R&D intensity for all cases(firms with and without is thus predictedto be favorably affecting the R&D). E(R&D') is the expectedvalue of R&D R&D behaviourof firms. The empiricalfindings for cases above the limit (firms with R&D). on the role of firm size,however,is observedto Lall aE(R&D.yax* is thechangein rheexpectedvalue bemixedin thecaseof Indianmanufacturing. o1'R&D intensityfor casesabovethe limit (wirh [983], {br a sampleof 100 Indian engineering R&D). dF(z)/dXuis thechangein rhecumulative firms for the year 1978,found that R&D intensity probability of being above the limit (having of the samplefirms dependpositivelyon rheir of industriesfor theyear R&D) associated with an independent variable. size.For a cross-section 1918-19, Katrak il9851 reported a less than Thus,equation( L2) statesthatthetotalchange proportionateincreasein R&D expenditurewith in R&D consistsof two interesting (l ) the an increasein firm size. There is anothergroup eff'ects: changein R&D intensityof firmsincurringR&D, of studies,which detecteda non-linearrelationweightedby the probabilityof doing R&D; and ship between firm size and R&D behaviour. (2) the changein the probabilityof doing R&D, Siddharthan [1988], for a sample of 166 weightedby the expectedvalue of R&D of firms manufacturingfirms over the period 1982-85, incurring R&D. The study will estimatethis found that the relationshipbetweenR&D intenTheR&D intensity decomposition for deriving more information sityandfirm sizeis U-shaped. until firm size,asmeasuredby than what ordinary Tobit coefficient commonly of firms decreases sales,reacheda thresholdlimit of Rs 600 million provides. and thereafterit increaseswith sales volume. Followingtheearliertheoretical andempirical Kumar and Saqib [1996] have esrimaredborh literatureon the determinantsof R&D activity at Probit and Tobit models for a sample of 291 firm-level for India and othercountries,the study Indian manuf'acturing firms for the period envisagethat R&D activity of pharmaceutical 1911-18to 1980-81to examinethedererminanrs firms may dependupona numberof factors(X ,,) of probabilityand intensityof R&D expendirure as discussed respectively. below. They foundan inverted-'U'shaped relationship betweenfirm sizeandprobabilityto Firm Size undertake R&D activity whereas the R&D intensityof firms is positivelyandlinearlyrelated Most of the empirical literatureon the deter- to firm size.In arecentstudy,KumarandAgarwal minants of R&D following the Schumpeterian [2000] for a much larger sampleof Indian manperspectiveof innovationstressesfirm sizeas an ufacturing firms over the period 1992-93 to importantfactor influencingR&D behaviourof 1998-99have reporteda horizontal S-shaped firms [fbr recent surveys, see Cohen, 1995; relationshipbetweenfirm size and R&D intenKumar and Siddharthan, 1991). The basic sity.In the pooledOLS estimation,flrm sizeand Schumpeterianhypothesisvisualisesa direct its cubic term have a significant negative

654

JOURNALOF INDIANSCHOOLOF POLITICALECONOMY

coefficientwhereasquadraticterm hasa significant positivecoefficient. In view of the inconclusive findings on the role of firm size in innovative activity in Indian manufacturingthe present study will also examine for possible non-linear relationship. Specifically firm size ( S I Z E )a s w e l l a s i t s q u a d r a t i cr e r m( S I Z E ' : w ) ill b e i n c l u d e di n t h ee s t i m a t i o o n f m o d e l( 1 . 1 ) .

OCT-DEC2OO2

variables-DISTECH (royaltiesand technicalf'ee paid abroadby the firm as a percentageof sales) and EMTECH (imports of capital goods as a percentageof sales)as two measuresof technology imports. OutwarclOrientation

R&D perfbrmanceof firms may alsodepend upon whetherthe firm is outwardorientedor not and if yes the degree and mode of outward As firms of developingcountriestend to have orientation. An outward oriented firm is one limited research capabilities to develop their which seesnot only domestic market but also indigenoustechnologicalcapabilities,they resort externalmarket as an important avenuefor its to imports of technologies from abroad. A growth and expansion.It can serve the external domestic firm can import technologicalinputs market throughexport or outward direct investlike plantandmachineryandfurtherit canacquire ment.In a knowledge-intensive segmentof global knowledgethrough technologyand know-hov" market like pharmaceutical, the export agreements.How are these embodied and dis- competitiveness increasinglylies in consciously embodiedchannelsof technologyimportsrelated created firm-specific knowledge like better to own in-houseR&D activityof thefirm? To the quality, innovative design and marketing by extentthat imports of foreign technologyrequire incurring greaterR&D expenses.Therefore.the further R&D on the part of importing entity to exporrinrensity(EXPOINT)of a flrm is expecred absorb. adapt and assimilate the imported to affect favorably its R&D activity. Braga and knowledgeto local conditions,it may stimulate Willmore [ 991] for Brazil andKumar and Saqib local knowledge-creating activities.It is also [1996]and Kumar andAgarwal[2000]for India possiblethat the relationshipwill be dominated have found that diversification of firms into by substitutionwhen availabilityand useof for- internationalmarketssignificantlyincreases both eign technologydiscourageandhencesubstitute theirprobabilityto do R&D andability to do R&D R&D activity of receiving firms. The natureof moreoutof totalsales.Whentheoutwardoriented R&D determineswhetherthe relationshipwill be firm choosesto servetheexternalmarketthrough complementaryor a substitutingtype. If R&D the mode of foreign direct investment, the activity is mainly of an adaptivetype as assumed industrialorganisationtheory suggeststhat such by Lall [1983] and Katrak [1985] for R&D internationaloperationof firms can be possible activity in Indian manufacturingthen a comple- only when it possessedsome monopolistic mentaryrelationshipcanbe postulated. Previous advantages conf'erring on it somesuperiorityover studieson Indian manufacturingpredominantly l o c a l r i v a l s i n t h a r m a l k e r . 'T h e R & D i s a n indicatea complementaryrelationshipbetween importantchannelof accumulating monopolistic importsof foreigntechnologyand R&D activit./ advantages and thereforefirms aspiringto go for of domestic firms [Lall, 1983; Katrak, 1985, internationalproductionare likely to undertake 1990;Kumar, 1981;Siddharthan,1988;Deolal- R&D activity.Lall [1983] documentedrhar rhe ikarandEvenson, 1989;Basant,1997;Kumarand proprietaryadvantages of Indian firms operating Agarwal, 20001. To test the impact of foreign overseas activitymainlydependupontheirability technologyon local R&D activity of Indian to reproduce a giventechnology, assimilating and pharmaceutical firms,the studyhasincludedtwo adaptingto local raw materialsor operating Imports of Foreign TechnoLogt

t

voL. l1 No. I

LIBERALIZATION, FIRM SIZE AND R&D PERFORMANCE.

655

conditionsfatherthanpushingbackthe frontiers large proportionof R&D, accountinglbr more of knowledge. Several other studies on the thanone-thirdof Singapore'stotal R&D spendthird-worldMNEs (TWMNET)suchason Korean ing. However,even in the caseof Singaporeit MNEs [Kumar and Kim, 1984; Euh and Min, was found that the R&D activitiesconductedby 1 9 8 6 1o, n H o n g K o n g M N E s l C h e n , 1 9 8 3 ] ,o n foreign companiesarerarely of basicresearchor ArgentineMNEs [Katz and Kosacoff,1983]and even applied researchand are generally less on BrazilianMNEs [Villela, 1983] suggestthar advancedthan at corporateheadquarters[Amsthe technologicalstrengthof developingcoun- denet al.,200I I. Therefore, a negativecoefficient tries' MNEs lies in their ability in local adapra- fbr the foreign dummy (FDUM) has been tions and modificationsand sometimeslittle postulated in themodel. improvements ol' imported technologies. Therefbre,literature on TWMNEs indicate that IntangibLeAssetsof Firms flrms undertakingdirect investmentabroadfrom developing countries have strengthenedtheir R&D activity of a firm canbe arguedto depend technologicalcapabilitiesby undertakingR&D positively on the intangibleassets(1NASSE7)of mainly in the natureof adaptation,assimilation the firm. Firms with superiorintangibleassetsin and improvementsof foreign technologies. The the form of trademarks,brands,copy-rightsand study thus postulateda positive relationship consumergoodwill are likely to invesi more in between the variable of outward investmenl R&D as their brandsuperiorityenablethem to (OINV) and R&D performance. better appropriatereturns from their innovative activity. Brand loyalty gives the firm required Ownership monopolypowerto undertakeR&D andmeetthe pref'erences of a more informed consumertoday. In the case of ownership of the firm the working hypothesisis that theforeignfirms spend Firru Age relativelylower thanwhat domesticfirms spend on R&D. It is arguedthat fbreign affiliates tend Technologicalcapacitybuilding by a firm is to do little R&D becausethey havecaptiveaccess an incrementaland cumulativeprocess,which to the laboratoriesof their parents situated in requiresthat the firm must accumulateknowlhomecountry.This hypothesis hasbeentestedby edge,skills, learning,operatingknow-how and severalstudiesin India IKumar,1987;Kumarand experiencethat supportcontinuouschangesand Saqib, 1996; Kumar and Agarwal, 20001and improvementsin productionprocess,products overwhelmingevidence suggeststhat foreign and procedures[Bell and Pavitt, 1992;Aw and firms in Indian manufacturinghave done signif- Batra,19981.A firm learnsfiom pastproduction icantly less R&D than their domesticcounter- experiences andusestheseaccumulated learning parts.Many studieson the internationalisation of for further technological improvement.ThereinnovativeactivitiesalsosuggestthatMNEs tend fore, firm age(ACE) as a proxy for accumulated to conductlittle R&D outsidetheir home base experience and technological learning is fPateland Pavitt, 1995;Pateland Vega, 19991. hypothesised to affect R&D performanceposiAmsden [2001] in a study on major developing tively. countriesof EastAsia and Latin Americafound that themore the fbreign ownershipthe lessis the Profit Margins depthand breadthof R&D. Among developing countriesSingaporestandsout to be an outlierin GiventhefactthatR&D activityinvolveshuge the sense that MNE affiliates had undertaken resouroe capabiiityon thepartof innovatingfirm,

JOURNAL OF INDIAN SCHOOL OF POLITICAL ECONOMY

OCT-DEC, 2OO2

a higherprofit margin indicatinginternalresource generationis likely to have f'avorableimpact on R&D decisionof the firm [Kumar and Saqib, 1996;Kumar and Agarwal,20001.This variable also capturesthe impact of fiscal measureslike tax exemption ofl'ered by the government for firms with recognisedR&D units. Other things being constantit is expectedthat a higher profit margin (PMRG) is likely to induce firm to undertakeR&D and spentmore as a proportion of sales.

policy, the hold of price controlon pharmaceuticalindustryhasbeensignificantlyreduced.The domesticflrms no longercan counton domestic marketsfor their growth and survival.In the face of stiffbr competitionfrom free imports as well as entry of new foreign firms they are forced to utilisetheirresourcesandconstantlyupgradeand improve their technologicalcapabilities.To the extent liberalisation forces firms to undertake R&D on accountof foreign competitionfor their survival,a positiverelationshipbetweenLiberalisationandR&D can be expected.The eff'ectof liberalisationhas beencapturedby includinga Liberalisation dummy variable(LIBDUM) taking valueo1'I for reform period (1993-94 to 2000-01) and 0 for Therehas beena radical shilt in the country's pre-reformperiod ( I 989-90 to 1992-93). policy frameworkgoverningproductionandtrade in 1991.Along with severalregulatorychanges After discussingthe probabledeterminantsof in the Indian economy including abolition of R&D, we shall include them in our model mandatorylicensingsystemand liberalisingFDI explicitlyto obtainthe following forrn: R&Dt,=80+BlAGE,,+B2SIZE,,+B3SIZEI+BaDISTECH,,+B5EMTECH,, +B6INASSET,, +B7OINV,, + B8EXPOINTI, + B9PMRGi, + pl0FDUM+ BI IILIBDUM+ u,, =0

ifX,,B+ u,,> 0 ilX,,B+u,,<0

(1.3)

Fitting a regressionequationlike equation effects. Although the simultaneousequations (1.3)for the searchof the determinants of firms' approachhasnot beenpursued,the singleequaR&D bchaviourhasbeenthe standardpracticein tion Tobit estimationadoptedin the studyserves the literature.However,regressing R&D expen- as a useful exploratoryestimation.As a result of diture on its supposeddeterminantsin a con- ignoring the problem of simultaneousrelationtemporaneoussettlng,as pursuedby the majority ship,theestimatesof all the parameterspresented ofexisting studiesand the presentstudy,suffers in the studyarelikely to be biasedto an unknown from the problem of simultaneity. The R&D extent. behaviourof firms is a complexphenomenonand the lines of causationoften run from supposed IV. Resultsand discussions determinantsto R&D and from R&D to its supposeddeterminants.For example,foreign techThe model (1.3) has been estimatedfor a nology purchaseby frrms may dependon their sampleof 2ll lndian pharmaceuticalfirms over initial indigenous technological capabilities the period 1989-90ro 2000-01.The srudydraws [Katrak,1997]or high profit marginsof the firm uponan exclusiveR/S-DS1R database to conduct may itselfhaveresultedfrom its successful R&D the quantitativeanalysis.Detailsaboutthe dataactivitiesIKumar and Saqib,1996].A few of the base used and measurements of variablehave previousstudieshave usedlaggedindependent beenprovidedin AppendixA. Table4 reportsthe variablesin theestimationbut precedence in time maximumlikelihoodestimationof pooledTobit does not necessarilydistinguishcausesfrom model as well as panel data random-eff'ectsof

r

VOL. l,l NO. I

LIBERALIZATION, FIRM SIZEANDR&D PERFORMANCE,

657

Tobit estimation.The pooled estimationresults The reported Wald Chi-squarestatisticsfor given under the heading column-A have been pooledandrandom-effectsTobit modelindicates provided with robust standardenors. STATA - that the estimatedmodels are statisticallysignifthe statisticalpackageused for the estimation icant. That meanstaken togetherall our indepurpose- produccsrobuststandarderrorsusing pendentvariables explaina significantproportion theHuber-Whitesandwichestimators which can of variation in the dependentvariable. It is effectivelydeal with a collectionof minor probremarkablethat the overallconclusionsderived lems of not meeting the classicalregression assumptions, namely about normality, from pooled Tobit model is same as those proheteroscedasticityor some observations that vided by the random-effectsTobit model. This thatobtainedresulton the exhibitlargeresiduals,leverageor influence.In similaritythussuggests determinants of R&D activity is robust to altercolumn-B we have providedfully standardised coefficientsof independentvariableswhich are nativeestimationprocedures, at leastbetweenthe by constructionscalefiee and henceareusefulin pooled and random-effects model. The comparingthe relativestrengthof the indepen- performance of individualindependent variables dentvariablesin termsofeffect on the dependent areas discussed below. variable.As discussedbefore,theordinaryoutput aspresentedundercolumn - A providesonly one Age; Therole of firm agein theR&D performance unstandardisedTobit coefficient for each indeof firms in Indian pharmaceuticalindustry is penden{. variable,notwithstanding thepresence of found to be favorable. Both the pooled and two typesof cases- thosewith zerovalueof R&D random-effectsmodel indicatesthat the variable intensity(firms not incurring R&D) and those wrth non-zerovalueof R&D (l'irrnsdoingR&D). has a positivecoefficient,which is statistically Therefore,thescsingleTobit coefficientsale nol significantat I per cent level. Keepingall else useful for effective interpretations.We have c o n s t a n t sa. o n e - y e a ri n c r e a s ei n a g e . o n a n provided two types of marginal effects in average,producesabout0.012increasein R&D McDonald-Moffitt Decomposition framework, intensityof samplefirms andabout0.002increase which are directly and effectively interpretable in their probability to undertakeR&D activity. (Column- C & D). In view of thepanelstructure This stronglysupportsour hypothesisthat older of our dataset,we also have estimatedrandom- firms in the industry have the competitive eff'ectsTobit model and the results obtained advantages of technologicallearningand expethereof have been presentedin column-E. As riencein doing R&D as comparedto start-ups. theoretrcal development on the conditional The vectorof standardised coefficients,however. fixed-effectsTobit model is still in infancy and indicatethat the relativecontributionoffirm age theredoesnot exist a sufficient statisticallowing in the explanationof R&D behaviourof pharthe fixed effects to be conditioned out of the likelihood,we arenot ableto provideresultsfiom maceutical firms is less dominant than other fixed effects.However,it is possibleto estimate factors like PMRG, SIZE, INASSET, erc. In unconditionalfixed eff'ectsmodel by including particular,fbr a standarddeviationincreasein firm-specificdummiesin the estimation,but the age, R&D intensityis expectedto increaseby resultsobtainedwill be biasedand henceinfer- 0.117 standarddeviations,holdins all other encesdrawnfrom thoseresultwill be misleadine. variablesconstant.

658

TOURNAL OF INDIAN SCHOOL OF POLITICAL ECONOMY

OCT.DEC.2OO2

Table 4. Tobit Estimation of R&D Intensitv Dependentvanable: R&D intensity (70) PooledTobit Estimation lndependenl Variable

(l) Firm Age SIZE SIZE, DISTECH EMTECH INASSET OINV EXPOINT PMRG FDUM LIBI)UM Constant

Sigma Sigma_e Sigma_u Log likelihood Waldchi2(ll) Prob > chi2 Observations Number of group

Coefficients ( R o b u s tZ value)

Fully Standardised coeffictents

( C o l u m n -A ) (2)

( C o l u r n n -B ) (3)

xx 0.0486098x (3.22) 8* 0.0r25460* (5.4e) -0.(x)ool s9*** (4 30) -0 0089174 (0.70) 0.002| 737 ( 1. 3 1 ) 0.0037849* (1.75) x* 0.0032283x ( 3 .r4 ) ** 0.0636728" (3.09) 0.0t21921** (2.30) o -s857572 (l2l) 3 3509366+** 13.17) l0 94662-50*** (4.15)

McDonald-Moffitt Decomposirion M a r g r n c lE f f e c r sa r M e a n s dF(z)/dk, dEy'/dk, ( C o l u m n -C ) (4)

( C o l u m n -D ) (5)

0 .I l 7 l

.01161679

.00200.5 13

0.4320

.00538806

.0009300 I

"0.3260

-3.79le-06

-6.543e-07

- 0 . 0 l18

-.00213108

- 00036784

-o.0226

-.0005 I 948

-.00008967

0.1912

00090453

.00015613

0.07'72

.00077t49

.000133 l6

0.I 769

.0t52t654

.00262646

1.2505

.00305707

.00052767

0.0231

.14256t04

.02465791

0.1624

.73797651

t2311236

-2.6161)2'79

-45I54167

Random-effects Tobit Estimation Coefficients (Z-value) ( C o l u m n -E ) (6)

0.0461297xxx (3.67) o o2Ios77*** (8.09) +* _0.00001428 (5.72) -0.0t't37 47 (0.,19) -0.00l4 I 54 (o.27) + 0.0036426x

(2.s s)

* 0.0027093* (2.04) 0.0602249*xx (6.2s) 0.0I 20649x** (3.87) 0 5873535 (0.82) I 19t4808*** (s33) -l04l3t003xx* (.t4.41J

7.6015t6 - 3 0 0 .15t 4 l 60.1 8 0.0000 I998 217

7.049t86 t.20t745 -2969.850 I 214.37 0 0000 I 998 271

A b s o l u t ev a l u eo f z s t a t i s t i c isn p a r e n t h e s e s . * Significantat 1070,*+ significantat 5Vc..*** significantat I %. Note: L DEy-/dx'is the changein the expectedvalue of dependentvariable for casesabove the lirnit (i.e., R&D rntensity> 0) and dF(z)/Jx,is the changein cumulativeprobability of being abovethe limit associatedwith an independent variable.2. Marginal effects is for discretechangeof dummy variablefrom 0 to L

Firm Size: According to the vector of standardisedcoefficientsthe effect of firm size on R&D behaviourof Indian pharmaceuticalfirms stood as the seconddominantfhctor after the eff'ectsof profit margin(PMRG).Not only it is the second most importantlactor influencingR&D but it is alsoobservedto possessnon-lineareffects.The flrm size and its squared terms turn out with statisticallysignificant positive and negative

coefficientsrespectively.Apparently, firm size has a positive effect on R&D perfbrmanceof firms but after some threshold the effect decreases with increasinglevelsof firm size(see Figure-3).This finding of inverted U-shaped relationshipbetweenR&D and firm size lend supportto the earlierfinding of Kumar and Saqib (1996) for a sample of Indian manufacturing firms.

vot. 14No 1

LIBERAI,IZATION,FIRM SIZE AND R&D PERT,ORMANCE

It should be noted thar majority of earlier studies suggesting that firm size and R&D behaviouris characterised by non-linearityindicate only the shapeof the relationship,falling shortof providingany exactfigure of threshold effect.In ouropinionresearchers shouldcalculate and presentthe value of threshold,as such a quantitymay be of directsubstantive interestfbr useful policy purposesand academicinterest alike. For Indian pharmaceuticalindustry this

information has been furnishedin Table 5. The numericallypreciseestimateof the turningpoint after which extra size affectsR&D negativelyis estimatedto be Rs 710.1 crore.Followine the delta methodN the standarderror of rhe tuining point is computedto be 69.9. The 95 per cenr confidenceinterval fbrmed on the assumption that the turning point is normally distributed clearlyoverlapswith the relevantrangeof firm size.

Ftgto'e 3 Fiued quudrutit. elJectoJ'.firm siz.eon R&D intensit,-" Totaleffect of Size with 95% value-wise confidence band

10.8239

.! o .! U)

u

-32.945 1983.89

S ize

Table 5. Analysis of the Non-linear Effect of Firm Size Statistics (l) R a n g eo f F i z e ( R s C r o r e ) S i z e + s i z eh' a sm a x i m u mi n t h e t u r n i n gp o i n t Std Error of turning point (delta method) 957oconfidenceinterval for the turning point

As we know now that firms sizesonly up to Rs 710.7crorehavea positiveimpacton theR&D performance,it will be useful to look at the size wise distributionof the total samDleobservat r o n s . rF" r o m T a h l e 6 r t c a n b e s e e nt h a tn e a r l y halfoftheobservations fall in thelowestsizeclass of Rs 0-20 crore.By the time sizereachRs 200 crore, 90 per cent of the sample has been exhausted.There are only 25 observationsthat fall in thesizeclass700-above range.Thisfinding

Value (2)

[.0],1983.89] 7 t0.6994 69 9656 (573.5693,847.8295)

only verifies the often emphasizedfeature of Indian pharmaceuticalindustry as highly fragmentedwith more than 20,000 firms competing fbr aroundRs 19,737croremarket.' ' The bulk of these20,000firms are small-scalefirms that are activein theindustrynow.Therefore.majorityof Indian pharmaceuticalfirms are f'ar below the turningpointandsuggests thatsmallfirm sizehas beena foremostfactorresponsiblefor keepingthe R&D performanceof the industryat a low level.

660

JOURNAI, OF INDIAN SCHOOL OF POLITICAI, ECONOMY

OCT-DEC. 2OO2

Table 6. Distribution of Sample Observation According to SalesRange SalesSize (Rs Crores) (t)

Number of observations (2)

(3)

0-20 20-50 50-l 00 I 00-200 200-400 400-700 700-above

r,015 359 246 238 143 45

49.0

2-J

Per cent

n3

I1.9 lt.5 6.9 2.2 t.2

Cumulative Per cent (4)

49.0 663 18.2 89.7 96.6 988 1 0 00

Source:Author's compurationbasedon RIS-DSIR database(2002)

The government policy in thepasthadactively healthsecurity.''The governmentprotectionof encouragedsmall-scalesector in the pharma- small-scalesector coupled with low level of ceuticalindustryas a part of the overall industrial patentprotectionfinally hasresultedin the larger development strategy of protecting and role that small firms are playing in the growth promotingsmall-scalesectorto achievea multi- performanceof the industry. Another upshot of ple of socio-economicobjectives such as thispolicy is thegeneration and strengthening of employmentgeneration andequity,decentralisedinter-firms linkages between small and large industrialdevelopment,tappingnew sourcesof enterprisesin the industry.Many largefirms who entrepreneurial capabilitiesand so on. However, formerly used to undertakeall slages of drug the two most important objects that marked the productionwith their integratedproductionprogovernmentpolicy in the caseof pharmaceutical cess started subcontracting work on several industry were the objects of self-reliancein the intermediatestageof productionto varioussmall productionofbasic drugsandensuringsupplyof firms to take advantageof governmentsubsidies cheapdrugs to the poor. A numberof drugs like t o t h es m a l l - s c a lsee c t o r . Paracetamol,Parabenes,Calcium Gluconate, Benzyl Benzoate,Pyrazolones,Lanolin AnhyAs the small size firms do not have huge drous, Citrates,HalogenatedHydroxy Quino- resourcesnecessaryfor developing any new Iines, etc., have been reservedfbr exclusive chemicalcompounds, theirsurvivalin theproduct developmentin the small scaleenterprises. The patent regime without government support is small-scalefirms were kept outsidethe purview unthinkable.Even their small size do not permit of DPCO andwereexemptedfiom thedrugpolicy them to undertake adaptive innovation as parameters.They were providedwith substantial reflectedin the large numberof firms not doing shareof the market in the GovernmentHealth anyR&D at all andin theverysmallshareof R&D Care Programme. in the total sizeof the majorityof thosewho do some R&D. The f'actthat competitionin pharThis policy of encouragingsmall-scale enter- maceuticalindustryis basedon technologyand prises has significantly influencedthe structure that small sizefirms lack resourcesto strengthen and development of Indian pharmaceutical their technologicalcapabilities warrant approindustry. It led to the emergenceof a strong priatepolicy responsespecificallyfocusingon the small-scale sector in Indian pharmaceuticaltechnologicalneedsof small scalesector.Just industryengagedin themanufacture ofdrugsand becausesmall sizefirms do not havethe required pharmaceuticals. Perhapsmoreimportanteffects technologicalstrengthto survive in a market are f'elt on the productionof bulk drugs and driven regime,the countrycan ill-affbrd to see consequentlyon the accessibilityof peopleto the witheringof its small-scalesectorthat is so

voL. 14NO.4

LIBERALIZA'TION, FIRM SIZE AND R&D PERFORMANCE

instrumentalin keepingthe pricesof many lii'e savingdrugsaffbrdableto the poorpeople.What the governmentat leastcould do is to strengthen the technology support and training for smallscalesectors.

Out*-ard Orientation: Both the measures of outwardorientation,viz., OINV signifyingserving of theforeignmarketthroughoutwardforeign direct investment and EXPOINT indicating servingof theforeignmarketvia exportsturnsout with positivecoefficientsand aresignificantat I TechnologyImports: None of the two measures per centlevel.ObviouslyIndianpharmaceutical of technologyimports,viz.,DISTECHmeasuring firms that are branchingout into foreign markets disembodiedtechnologyimports and EMTECH whethervia FDI or via exportsexhibit higher measuringembodiedtechnologyimports have probabilityto undertake R&D andinvestmorein comeup with significanteffect.The sign of both R&D as a proportion of total sales. In a these variablesare observedto be negativebut knowledge-based industrylike pharmaceuticals, statistically not differentfrom zero.This suggests the globalcompetitiveness of a firm is drivenby the relationshipbetweentechnologyimportsand hightechnology, highskill,qualityandreliability. R&D efforts of firms is neithermarked by com- Therefore,entry into global market requlres a plementaritynor substitution.The impact of strongtechnological backupon thepartofentrant. technologyimportstendsto vary aorossfirms and and intense competitive pressure.based on on the averagedoesnot possessany systematic technological dynamism,ensuresthatthe firm is effect on the technologicaleffbrts of importing continuously innovativein orderto be ableto stay firms.This findingsis consistent with the earlier in the market. findingsof KumarandSaqibI I 996]thattheR&D activity of Indian manufacturingfirms is neither Profit Margins: The link betweenprofit margins, complemented by technologyimportmeasured as PMRG, and R&D activity has been lbund to be technologylicensingpaymentsnor is substituted positive.PMRG hascomeup with a positivesign by it. and significantat 5 per cent level. In particulara 1 per cent increasein the profit marginsof firms Intangible Asse/s.' INASSEI representing the on an averagelncreasesabout 0.00053in the intangiblcassetsof the firm turnsup with a pos- probabilityof firms to undertake R&D andabout itive sign and is statisticallysignificantat l0 per 0.0031 in the R&D intensilyof firms keeping cent level. In terms of the strengthof relative othervariablesconstant.The efl'ectof this varicontributionas indicatedby standardised coeffi- ableis the most significanton R&D performance clents vector, intangibleassetsof the firm stood as shown by the vector of standardised as the third dominating factor. A I percentage coefficients.Therefore,the result suggeststhat increasein the intangibleassetsof thefirm, on an internal resourcegenerationof the firm signifiaverage,brings about 0.0009 increasein R&D cantly increasesthe R&D activity of Indian rntensityof firms engagingin R&D activity, pharmaceutical companies. keepingother variablesconstant.The marginal impact of I per cent incleasein the intangible Ownership:The FDUM capturingrhe effect of assetson the probability of firms engaging in foreign ownershipon the performanceof R&D R&D activity is, on an average,estimatedto be emergeswith a positivecoefflcientthat is statisabout0.00016.The findingweaklylendssupporr tically not different from zero. Therefore,there to our contentionthat firms with high brand val- is no evidenceto suggestthatR&D behaviourof uation are inclined to do R&D as they are better firms differs on having majority fbreign ownerplaced to appropriatereturns from their R&D ship as opposedto havingdomesticownership. actlvlty. This finding is particularlysignificantand at

662

JOURNAL OF INDIAN SCHOOL OF POLITICAL ECONOMY

OCT.DEC.2OO2

variantwith the view that liberalFDI policy and and probabilityto do R&D are also quite constrengthening of patentsystemwill leadto a spurr siderable.This suggeststhat liberalisationof in innovativeactivitiesof l'oreignfirmsandhence industrial,tradepolicieswith impendingproduct will lead to an increasein the internationaltech- patentregime have made Indian pharmaceutical nology transferto India. It is arguedthat foreign firms more consciousof the need to undertake firms will introduce their new products in the R&D activity, and indeedthey had devotedsubcountryandmay relocatetheir R&D unitsin India stantialresourcesin thatdirection.Remembering becauseof its cheappersonnelcosts.However, the structureof industrywheremajority of firms theview thatMNEs may actasan engineof R&D areessentially smallsizeimply thattheimproved R&D perfbrmance performance in the reform periodmay well doesnot inspiremuchconfidencein the f-aceof many MNEs like Ciba Geigy,Boots, havecomefrom theperformanceof a small group Hoechstand RhonePoulenceclosingdown their of largesizefirms.Small-scale sector,dueto scale R&D unitsat a time when the countryis moving and resourceconstraint,are not in the positionof towardsa productpatentregime.If experiences venturing into R&D-led growth as a few large areany indication,the monopolystatusof MNCs Indian pharmaceuticalfirms are doing. The may even lead to contractionof innovativeacti- governmentincentivepackageoftenwas of little vities, as happenedin the caseof Patentsand help to small sectoras comparedto largeenterDesignsAct, 19I I . Giventheprovisionof TRIps prisesbecauselatterare betterplacedto obtain thatimportsis akinto localproductionit mayeven import permits fbr capital goods, intermediate result in shifting of existingR&D units in the rnputs and raw materialsand have preferential country to the home country of foreign firm accessin the domesticcredit market.In many concerns.TRIMs, which prohibitthe imposition cases,small firms were ignorant of available or wereunableto handlethe proceof performance requirements like, export. concessions dural and administrativecomplexity involved in obligations,local content requirements,local the relevantoffice work. The fact that small-scale manufacturingrequirements, etc.,by host counsectoris instrumental in ensuringtheaccess of the tries further underminethe capabilityof develpoor to quality drugs calls for greater role of oprngcourtiersto induceforeignfirms to do R&D governmentto directlystrengthening its techncllocally.r3 logical capabilitiesso that it can survive in a liberalisedbusiness environmenr. Liberalisation: The variable, LIBDUM, which capturethepossibleeff'ects ofliberalisation on the V. Conclusions and Implications R&D performance of Indian pharmaceutical flrms has come out with a positivecoefficient Along with the implementationof macroecostatisticallydifferent from zero at I per cent signomic liberalisationin the countrythe nineties nificance level. This suggeststhat R&D perhad witnessedsignificantchangesin the policy formanceof pharmaceuticalfirms has increased regimegoverning Indian pharmaceutical indussubstantiallyin the refbrm period (1993-94to try. The progressive dilutionsof DpCO, liberal 2000-01) as compared ro pre-reform perioc FDI policy, and transitorymeasuresof TRIps (1989-90to 1992-93).The standardised coeffi- haveinducedintensecompetitionin the market. cient indicatethat in the postreform periodR&D The above empirical exercise finds that this intensity of Indian pharmaceutical firms is competitivepressurehas worked effectively in expectedto increaseby 0.1624 standarddevi- pushingIndian pharmaceuticalfirms into R&D ations,holding all other variablesconstant.The activity. However, it is inferred that this impact marginal effects of LIMDUM on R&D intensity of liberalisationis likely to be limited to a f'ew

voL. I't NO 1

LIBERALIZATION, I.'IRM SIZE AND R&D PERFORMANCE.

largeandmediumsizefirms, aslargesegmentof with equityvalueup to $2 millions of which up small srzefirms lack the hugeresources that are to $ 500,000could be in cashand rest by capirequiredfor productdevelopment. The impactof talisation of Indian exports of machinery. firm size is also observedto have strong non- equipment,know-how or other services.These linearimpacton the R&D performance. Recently procedures havebeenfurtherliberalisedin 1999 government has taken some initiatives like and 2002 Guidelines.These outward oriented establishment of a Drug DevelopmentPromotion policiesarelikely to improvethecompetitiveness Foundation (DDPF) and a Pharmaceutical of Indian pharmaceuticalfirms and hencethere Research and Development Support Fund needto undertakelarge-scaleR&D activities. (PRDSF)in order to promoteR&D activity in the rndustry.Thesegovernmentmeasuresare steps Anothersignificantobservation of thestudvis in the right directionsbut also needto be target thattheR&D behaviourof Indianpharmaceutical orientated towardssmallsizefirms asthesefirms firms crucially dependson theirintangibleassets are instrumentalin keepingdrugs pricesaccesmainly brandvaluation.Firms that arepromoting sibleto thepoor.Also, at thesametime we should and creatingbrandsare found to be doing more promotesomenationalchampionsas is doneby R&D activity astheseintangiblesstrengthentheir developedcountries under their strategictrade power to appropriaterentsfrom their innovative policies. activity.In addition,profit marginsand firm age are other two importantdeterminantsof R&D The R&D behaviourof Indianfirms appears behaviourof Indian pharmaceutical firms. The to be not systematicallyaffectedby the availR&D behaviourof foreign firms is found to be ability of foreign technologythroughlicensing not different from domesticenterprises. and imports of capital goods. However, the outward orientationof an enterpriseis a signifiThe policy implicationsfrom the aboveanalcant determinant of in-house R&D. Therefore ysis are obvious.In order to enhanceR&D pergovernmentpoliciesthatencouragesIndianfirms formance of Indian pharmaceuticalfirms the to export and to undertake outward direct government shouldfocuson removingobstacles lnvestmentare very crucialin inducingfirms to that inhibit Indian firm's participationin inrerfocus more on the developmentof indigenous national marketsvia exportsor via outward fortechnologies.For a long time the government policy with respectto outward foreign direct eign direct investment. Recognising the investmenthas beenrestlictivedue to the insuf- importantrole of firm size in R&D performance ficientforeignexchangereserves and precarious policymustcontainspecialschemefbr smallsize BOPposition.Onlyjoint ventureswerepromoted firms in the overall technologypolicy fbr the with minority Indian ownershipand even that industry.Giventhehugecostinvolvedin thebasic thepathofcollaborativeresearchefforts minor equity participationwas requiredto be in research, the form of exportsof Indian madecapitalgoods, betweenprivatesectorsand governmentresearch equipmentsand know-how.It is encouragingto institution appearsto be an important strategic note that recentlytheserestrictionson outward option that needs to be promoted seriously. direct investment have been liberalised. In Technoiogytransferrequirementsfor foreign October1992governmenthad issuedthe modi- flrms or otherperformancerequirementsthat are fied Guidelinesfor Indian Joint Ventures(JVs) permitted under TRIMs agreementscan be utiandWholly OwnedSubsidiaries Abroad(WOSs) lisedto thefullestextentto persuadeforeignfirms which provided for automaticapprovalfor cases to relocatetheir R&D units into the country.

661

JOURNAL OF INDIAN SCHOOL OT'POLTICAL ECONOMY

NOTES

OCT-DEC2002

I I . The productionfigure is for the year I 999-2000taken from Organisation of PharmaceuticalProducersof India (OPPI).

l. See Kumar and Pradhan[2002] for details of policy changesandits impacton thegrowth of Indianpharmaceutical 12. The shareof small-scalesector in the productionof rndustry. b u l k d r u g sh a si n c r e a s efdr o m 7 . 7 p e r c e n ti n 1 9 7 5 - 7 6t o 2 0 . 9 2 . F i k k e r r ! 9 9 3 1 , H a k s a r[ 9 9 5 ] a n d K u m a r a n d S a q i b per cent in 1985-86.The conespondingshareof MNE affiliI I 9961have arguedin their quantitativeexplorationsinto the ateshasdecreasedfrom 40 per cent in 1915-T6 to I 8 per cent R&D activity of lndian finns that the innovativeactivity of in 1985-86[see,Table I in Kumar and Pradhan,2002]. 13.SeeUNCTAD [2001] for an illustrativelist of 39 host theseenterpriseswas stimulated by the soft patent regime country operational measures,Pp 8-9. Historically both underthe 1970 PatentAct. developedanddevelopinghostcountriesalike haveusedthese 3. For example the prices of Ranitidine, Famotidine, measures asa developmental tool to ensuremaximumbenefits Astemizole,Ondansetronin the US market are at about 50 from foreign capital while keepingat minimum its negative times the Indian prices and most of these drugs had been impact. However, the use of thesemeasuresis increasingly introducedin the domestic market within 4-5 years of their under attack from developedcountries led by the United i n t r o d u c t i o ni n t h e w o r l d m a r k e t( s e eT a b l e2 i n K u m a r a n d States.The agreementon TRIMs in the 1994 Uruguayround P r a d h a n2. 0 0 2 ) . G A T T n e g o t i a t r ocno v e r e d( i ) l o c a lc o n t e n rt e q u i r e m e n t (sr.i ) 4 . D e s a r l l 9 8 0 l d o c u m e n t e dt w o c a s e sw h e r e f o r e i g n export performancerequirements,(iii) local manufacturing patent owner neither had used therr patents for domestic requirements,(iv) trade balancing requirements,and (v) manufacturingnor allowed theln to be used by local firms. foreignexchangerestrictions. Theseare:( I ) HoeshstpreventingUnichemLaboratoriesfrom REFER.ENCES producing tolbutamide, and (2) thereuponExcel Industries being prevented from producing the fumrgant by another Amsden,A.H., 2001: Tlrc Riseof the Rest:Challengesto the foreign firrn. West.fron Lute-lndustriuLrsing Eutnomies.Oxford Uni5 . K u n i a r[ 1 9 9 6 ]f o u n dt h a tR & D i n t e n s i t yo f U S a f f i l i a t e s versity Press,New York. is posrtivelyand significantlydependentupon the strengthof A m s d e n .A . H . , T e d T s c h a n ga n d A k i r a C o t o . 2 0 0 1 : ' D o patent protection (Rapp and Rozek index) in the case of Foreign CompaniesConductR&D in DevelopingCoundevelopedcountriesbut not statisticallydifferent from zero tries?A New Approachto Analysing the Level of R&D. with an Analysis of Singapore',ADB Institute Working fordevelopingcountries.Kumar [2001] in a morerecentstudy Paper 14, Tokyo. confirmed that the strengthof patentprotection(Gina(e and 'Technologrcal CapaPark index) is not a significant factor in explaining R&D Aw, Bee Yan and Ceeta Batra, 1998; bility andFirm Efficiency in Taiwan lChina)','[he World intensityof US and Japaneseaffiliates. BtrnkEcononic Review, 12 ( I ), Pp. -59-79. 6 . G O I , 2 0 0 1 ,P p . 1 4 0 . Basant, Rakesh, 19971 'Technology Strategies of Large 7. The industrial organisationtheory of FDI as proposed Enterprisesin IndianIndustry:SomeExplorations',World by Hymer I I 960] and later extendedby KindlebergerI I 969] D e v e k t p m e n 2t ,5 ( 1 0 ) ,P p . 1 6 8 3 -1 7 0 0 . and CavesI I 97 I ] hasbeenthe mostdominantexplanationfor Bell, Martin and Keith Pavitr, 1992; 'Accumulating Techforeign operationof nationalfirms. This rpproach tracesthe nological Capability in Developing Countries'. Proexistenceand growth of the internationaloperationof firms ceedings ol the WorLd Bunk Annuul Conlerente on in the phenornenonof market imperfections.According to D e v e k t p m e nEt c o n o m i c s1, 9 9 2 ,P p . 2 5 7 - 5 8 1 . Hymer, firn.rsundertakinginvestmentabroad must possess l B r a g a ,H . a n dL . W i l l m o r e ,l 9 9 l : ' T e c h n o l o g r c al m p o n sa n d some monopolistic advantageslike product differentiation, TechnologicalEfforts:An Analysisof Their Deternrinants managementskill, patenrsand superiortechnology,controlof in BrazilianFirms' , TheJournul of lndustrnl Economrcs. the supply of key raw materials,economiesof scale.etc., 39(a), Pp. 421-432. which they can profitably expioit abroad by internalisrng C a v e sR , . E . ,1 9 7 1i ' l n t e m a t i o n a C l o r p o r a t i o n st h: e I n d u s t n a l production rather than exporting from home country or Economicsof Foreign Investment',Economicu,38, Pp. licensingthoseadvantagesto a third party abroad. t - 2 1. 8. Linear approximationof the nonlinearfunction of the C h e n ,E . K . Y . , 1 9 8 3 ; ' M u l t i n a t i o n a l sf r o m Hong Kong', in turning point in the regressioncoefficients. Lall (ed.) The New Multinationuls: The Spreud ol Third 9 . T h e g r a p h h a s p l o u e d S I Z E a g a i n s t0 . 0 2 2 5 5 * S I Z E lUorld Enterprrse.r, John Wiley & Sons.New York. Pp. 0 . 0 0 0 0 16 x s t z B z . 8 8 -l 2 6 . I 0. The numberof sampleobservationsin the presentcase C o h e n . W . M . , 1 9 9 5 ; ' E m p i r i c a l S t u d i e s o f I n n o v a r i v e may not be equal to what was reportedin the estimatronas Activity', rn Paul Stoneman (ed.), Handbook ol the STATA had dropped some observationsowing to missing Economics of Innovatbn und Technobgical Cltunge, valuesof indeoendentvariables. Oxford: Blackwell, Pp 182-264.

voL. l1 No.1

LIBERALIZATION, FIRM SIZE AND R&D PERFORMANCE

665

Deolalikar,Anil B. and l{obert E. Evenson, 1989t 'Technol- Kumar,Nagesh,200 I ; 'DeterminantsofLocation ofOverseas ogy Production and Technology Purchase in Indian R&D Activity of MultinationalEnterprises:The Caseof Industry: An Econometric Analysis', The Review ol US and Japanese Corporations',Reseu'chPollc.r.30, Pp. Econonticsond Stuistics,7 I (4), Pp. 687-691. I 5 9 -I 7 4 . D e s a i .A . V . , 1 9 8 0 ; ' T h e O r i g i n a n d D i r e c t i o no f I n d u s t r i a l Kumar, Nageshand AradhnaAgarwal,2000; 'Liberalisation, R & D i n I n d i a ' . R e s e u r t ' hP o l i c t , ,9 ( J a n u a r y1 9 8 0 ) ,P p Outward Orientation and ln-house R&D Activity of Multinational and Local Firms: A QuantitativeExplora14-96 t i o n f o r l n d i a n M a n u f a c t u r i n g ' .R I S D i s t ' u s s i o tpt u p e r E u h .Y . a n dS . H .M r n , 1 9 8 6 1' F o r e i g nD i r e c tl n v e s r m e nftr o m #07/20002. DevelopingCountnes: The Case of Korean Firms', Zfte Kumar, Nageshand Mohamrned Saqib, 1996; 'Firm Size. D e v e k t p t n gE c o r t o n t L exsx. x i v - 2 , P p . 1 4 9 - 1 6 8 . Opportunitiesfor Adaptation,andIn-houseR&D Activrty Fikke(, Brian, 1993,An Openor CktsedTechnokryyPolicl"l in Developing Countries:The Case of Indian ManufacThe Elfects oJ Technobgy Licensing, Foreign Direct turing' , ResettrchPolicy, 25(5), Pp. 1 12-22 Investment, ond Technobgl' Spilbvers in Indiun IndusKumar, Nagesh and N.S. Siddharthan,1997 Technoktgt. ttial Sector Firms, Ph.D. Dissertation,Yale University. Markel Structure und Internationulisution: Issuesund COI, 2001; /NDIA, PublicationDivision, Ministry of InforPolicies.fbrDevektping Countrie,s.Routledgeand UNU mation and Broadcasting,New Delhi. . Press.London and New York. H a k s a r ,V . . 1 9 9 5 ; ' E x t e r n a l i t i e sG rowth and Technology Kumar. Nagesh.and JayaPrakashPradhan.2002. 'Econornic Transf'eA r : p p l i c a t r o nt o t h e I n d i a nM a n u f a c t u r i n S g ector. Reforms, WTO and lndian Drugs and Pharmaceuticals 1 9 7 5 - 9 0 'W . a s h i n g t o nD C : I M F , m i m e o . Industry: Implicatronsof ErnergrngTrends', A Report H y m c r .S . H . , l 9 l 6 : T h e l n t e r n a t r o n oO Lp e r a t i o n o s l Nutionul preparedfor the Centre for Multidisciplinurl, DevelopFirm.s;a Studlof Direct Foreign Inrestment,Boston,MIT nent Reseorch, Dharwad. Press. 'Determinants Lall,S., 1983; o f R & D i n a nL D C : T h e I n d i a n Katrak, Homi, 198-5:'lmported Technology,EnterpriseSize Engineering Industry', Econonic Letters. 31 (3\. Pp. and R&D in a Newly IndustrialisingCountry:The Indian 379-383. Experience',Oxlord BuLletinol Eutnomicsund Stutistics. 'Multinationals Lall, S.. 1983; f r o m l n d i a ' . i n L a l l ( e d . )? r e 4 7. P p . 2 1 3 - 2 3 0 . NeryMuLtinationuls;The Spread ol T'hirtl Wttrld Enter'lmpons K a t r a k ,H o m i , 1 9 9 0 : o f T e c h n o l o g ya n d t h e T e c h p r l . i e s J, o h nW i l e y & S o n s ,N e w Y o r k . P p . 2 l - 8 7 . nologicalEffbrts of Indian Enterprises'. World Det'ekry'Technological Lall, S., 1992; C a p a b i l i t i e sa n d I n d u s t n a l i n r c n t ,l 8 ( 3 ) , P p . 3 7 I - 3 8l . W o r k lD e v e l o p m e n t , 2 0P, p . 1 6 5 - 1 8 6 . s a t i o n ' , Katrak, Horni, 1997: 'Developing Countries' lmports of McDonald, John F. and Robert A. Moffitt, 19801'The Uses Technology, In-House Technological Capabilities and of Tobit Analysis', The Review ol Econontics und Str:Efforts: An Analysis of the Indian Expeience', Journal t i s t i c s6, 2 , P p 3 l 8 - 3 2 1 . ol DevekrymentEcorunnics,53,Pp. 67-83. Patel, P. and K. Pavitt, 1995; 'The Localised Creation of 'Multinationals Katz, J. and B. Kosacoff, 1983; fron Global TechnologicalAdvantage',J. Molero (ed.) TechArgentina', in Lall (ed.) The New Muhinutionuls; T'he nobgicttl Innovotiotr,Multinationul Corporutions und Spreudol Third World Enterprise.r,John Wiley & Sons, New Interrufiional Competitiveness; The Case ol New York, Pp. 131-219. IntennedioteCountries,Harwood Acadeniic Publishers. Kindleberger,C.P., 1969;Anericun BusinessAbntad, Yale G e r m a n yP , p.59-74. U n i v e r s i t yP r e s sN . ew Haven. Patel,P. and M. Vega, 1999; 'Patternsof Internationalisation 'The Kumar, K. and K.Y. Kirn, 1984; KoreanManufacturing of CorporateTechnology: Location vs. Home Country Multinationals'. JournuL ol lnternational Business StuA d v a n t a g e sR ' , e s e u r c hP o l i c t ' ,2 8 , P p . 1 4 5 - 1 5 5 . d l e s ,S p r i n g / S u m m e rP, p .4 5 - 6 1 . S i d d h a r t h a nN. . S . , 1 9 8 8 : ' l n - h o u s eR & D . I r n p o r t e dT e c h Kumar, Nagesh, 19871'TechnologyImports and Local nology.and Firm Size:Lessonsfron.rIndianExperience'. ResearchandDevelopmentin IndianManufacturing',7he The DevektpingEutnomies.XXVI-1" Pp. 212-221. Devektping Econont Les, xxv-3, Pp. 220-233. UNCTAD, 2002: Host Country Operutittnul Meusures. Kumar, Nagesh, 1996: 'lntellectual Property Protection United Natrons,New York and Geneva. Market Onentation and Location of Overseas R&D V i l l e l a ,A . , 1 9 8 3 ;' M u l t i n a t i o n a l sf r o m B r a z i l ' , i n L a l l ( e d . ) Actrvities by Multinational Enterprises', World DevelThe Nev' Multinatiottttls; The Spreud ol Third World ] p m e n t ,2 4 ( 4 ) ,P p . 6 7 3 - 6 8 8 . Enterprises,JohnWiley & Sons,New York,Pp.220-249.

r

666

JOURNAL OF INDIAN SCHOOL OF POLITICAL ECONOMY

OCT.DEC.2OO2

Appendix A: Dataset and Measurements of Variables The datasetused1nthe presentstudyis a sub-sampleof a largerdataset,R IS-DSIR database, constructedfiorn differenr sourcasat the Researchand Inforrnation System for the Non-alignedand Other Developing Countnes, as a part of the Departrnentof Scientiflc and Industnal Research(DSIR) researchproject 'A Strategic Approach to Strengtheningthe lnternationalCompetitivenessin Knowledge-basedIndustnes:Some Explorationsinto the Role of FDI Inflows, Outward lnvestments,and EnterpriseLevel TechnologicalEfforr in Promotionof lndia's KnowledgeIntensiveExports'. The dataset. which covers firm-level data on variousflnancial variableslike exports,imports,sales,R&D, outward investments,etc. of more than 500 Indian manufacturingcompanies.has been compiled fiom the PROWESS database(2002). rhe Ministry of Commercc,the Ministry of Finance,and the lndia InvestmentCentre. M eosurentents A l. DependentVariable. R&D,,: Total R&D expenditureas a percentageof total salesof ith firm in tth year. A2. IndependentVanables. AGE,,:The age of lth firm in number of years. SIZEII: Total salesof lth firrn in fih year. S I Z E , , : f h e s q u r r c dr c n n o f r h es a l e so f r t h f i r n r i n n h y e a r DISTECH,,:Royalties.technicaland other prof'essional f'eesrcmitted abroadby ith firrn as a percenrageo1'salesin the year t. EMTECH,,: Imports of capital goods by lth firm as a percentageof salesin nh year. INASSET,,:Intangibleassetofthe ith firm as a percentageofsales rn the year t. This is the brand valuationas grven in tne balancesheetof the company. OINV,,: Defined as the stock of outward direct investmentof the ith firm as a percentageof salesmultiplied by the age of multrnationality. EXPOINT,,:Exports of ith firm as a percentageof salesin the year t. PMRG,,:Profit before tax (PBT) as a percentageof sales. FDUM: Dummy variablefor foreignowned firm taking value I for firms with 25 per centor more foreignequity participatron and 0 otherwise. LIBDUM: Lrberalisationdummy taking I for reforn period I 993-94 to 2000-01and 0 for the pre-reformperiod I 989-90 ro t992-93