Innovation, Formal vs.
Informal R&D, and Firm Size: Some Evidence from Italian
Manufacturing Firms
Author(s): Enrico Santarelli and Alessandro Sterlacchini
Source: Small Business Economics, Vol. 2, No. 3 (1990), pp. 223-228
Published by: Springer
Stable URL: [Link] .
Accessed: 18/06/2014 11:40
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
[Link]
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@[Link].
Springer is collaborating with JSTOR to digitize, preserve and extend access to Small Business Economics.
[Link]
This content downloaded from [Link] on Wed, 18 Jun 2014 [Link] AM
All use subject to JSTOR Terms and Conditions
[Link]
Innovation, R&D,
andFirmSize:SomeEvidencefrom
EnricoSantarelli
Firms*
ItalianManufacturing Alessandro
Sterlacchini
[Link] thebasisofdatafrom tworecentsurveyson
(mainly R&D expenditures) failto capturethein-
innovationdiffusion in Italianmanufacturing novative
this
industry, performance ofsmall firms
R&D is an important
adequately.1
papershowsthatinformal partoftheWhen small and mediumsize enterprises
totalR&D undertaken by smalland mediumsized firms.
Nevertheless,whenan outputindicator (henceforth
suchas thenumber SMEs) carryout theirinnovative
and the natureof the innovations introduced activities
by firmsof theyoftendo so without specificfinan-
different
size is used,it emergesthatsmallfirms cial and managerial
haveintro- resourcesand,in particular,
ducedmainly incremental rather
[Link]
The formalised procedures. ThusSMEs tend
papertherefore suggests thatsystematic
R&D undertakento undertake
by a significantamountof innovative
largefirmswithin structured is moreeffective
laboratories (in
activitiesin theirdesign,productionand sales
termsof productinnovations) thanoccasionalR&D carried
outbysmallfirms. departments ratherthanin theirR&D depart-
ments(whichoftendo notexistatall).
Theproblem withtheR&D figures provided by
officialsurveys is thattheydo notincludethese
I. Introduction R&D activities. inthisjournal,
informal Recently,
In spiteoftheoverwhelming importanceoflarge Alfred Kleinknecht (1989) has presentedthe
firmsin innovation, it is widelyacknowledged results of a of
survey 3,000 Dutch firms,and
thattraditional indicators of innovative
activities stressed therole ofsmall firmsinindustrial R&D.2
In his article,Kleinknecht showsthatifinformal
R&D is takenintoaccounttheR&D commitment
FinalversionacceptedonJanuary16,1990 of SMEs (with less than 500 employees)is
EnricoSantarelli considerably higherthan thatreportedby the
SciencePolicyResearchUnit [Link] main difference between
ofSussex
University Kleinknecht's questionnaire and theofficial oneis
- Brighton
Fainter BN1 9RF thattheformer includesa questionon informal
EastSussex
UK. R&D (askingformanyearsof R&D insteadof
and R&D expenditures) undertakenoutside R&D
Universitd
"G.D'Annunzio" laboratories.
Institute
ofEconomicsandStatistics In this paper we show that Kleinknecht's
VialeCrucioli are quiteconsistent withthosethathave
64100Teramo findings
emerged fromtwoverysimilar surveys carriedout
Italy
in [Link] fact,informalR&D represents an
important partof thetotalR&D carriedout by
ItalianSMEs. However,if an outputindicator
Alessandro Sterlacchini suchas thenumberand thenatureoftheinnova-
UniversitddiAncona
diEconomia tions introduced byfirms ofdifferent sizeis used,
Dipartimento
ViaBarellill theItaliansurveys showthatsmallfirms are able
60121Ancona to developincremental innovations ratherthan
Italy majorinnovations, whereasthe mostsignificant
SmallBusiness
Economics2: 223-228, 1990.
€>1990 Kluwer
AcademicPublishers.
PrintedintheNetherlands.
This content downloaded from [Link] on Wed, 18 Jun 2014 [Link] AM
All use subject to JSTOR Terms and Conditions
224 EnricoSantarelli Sterlacchini
andAlessandro
productinnovations
aremorefrequently achieved over,as Acs andAudretsch (1988) haveshown,a
by large [Link] findingsuggeststhat the much more disaggregated analysisis neededfor
systematicR&D activitiescarriedout by large rigoroustesting of the relationship betweenfirm
firmswithinstructured laboratoriesare more size and innovation. This is not possiblewith
effectivethan the occasional R&D activities [Link], itshouldbe
undertakenbysmallfirms. stressedthataccordingto theItaliansurveythe
innovative ofSMEs (withlessthan500
capability
employees) is not strikingly lowerthanthatof
II. Innovationand firmsize
firms havingmore than 500 employees.
The firstItaliansurveyon innovation diffusion When the sourcesof innovations are con-
(ISTAT-CNR,1987) wasconducted by interview- sidered,however, significant differences between
ing35,000manufacturing firmswith more than 20 SMEs and largefirms emerge. In thefirst survey,
employees:24,104 firmsreturned the question- innovating firmswereaskedtoidentify thesources
[Link] I showsthattheshareofthosefirms oftheirinnovations: in Table II theseare divided
thatintroducedinnovations during the period between internalsources of innovation ("R&D,"
1981-85 [Link],this and
"Design industrialization," and"Patents") and
resultcannotbyitself lendsupport totheso-called oneofitsmajorexternal sources("Intermediate &
Schumpeterian hypothesis sinceinnovative activi- capitalgoods"), which can be taken as a measure
tiesdid [Link]- ofembodiedtechnical progress. As faras internal
sourcesare concerned,SMEs appearedto be
TABLEI
relatively more engagedin innovative activities
carriedout in the designand industrialization
Firmsize Firms Firmswhichhave
(employees) responding introduced innovations stagesthanwithinR&D departments; onlylarge
tothe firms withmore than 500 employeesinstead
questionnaire assigneda highpriority to formal R&D activities.
On theotherhand,SMEs resorted tointermediate
Number % oftheclass
andcapitalgoodsmorefrequently thantointernal
sources.3
20-49 14885 9419 63.3 in
50-99 4673 3490 74.7 By applyingthe Z-testto the differences
100-199 2553 2090 81.9 means between SMEs and large firms we found
200-499 1298 1084 83.5 thatthesignificance levelofsuchdifferences was
500 andmore 695 618 88.9 99% for all internalsourcesof innovation; by
Total 24104 16701 69.3 contrast, theywerenot significant forthemajor
external sourceconsidered(i.e.,"Intermediate &
Sourceofdata:ISTAT-CNR(1987). capitalgoods").In conclusion, therefore, thisfirst
TABLED
Recoursetodifferent
sourcesofinnovation*
Firmsizes R&D Designand Patents Intermediate
(employees) industrialization & capitalgoods
Number% oftheclass Number % oftheclass Number % oftheclass Number% oftheclass
20 to49 1073 11.39 3354 35.61 1085 11.52 6519 69.21
50 to99 564 16.16 1533 43.93 602 17.25 2402 68.82
100 to 199 449 21.48 1101 52.68 416 19.90 1444 69.09
200 to499 358 33.03 622 57.38 248 22.88 770 71.03
500 andmore 270 43.69 429 68.93 233 37.70 419 67.80
Sourceofdata:ISTAT-CNR(1987).
* Firmscouldindicatemorethanonesourceofinnovation.
firms.
onthetotalofinnovating
arecalculated
Percentages
This content downloaded from [Link] on Wed, 18 Jun 2014 [Link] AM
All use subject to JSTOR Terms and Conditions
Innovation, R&D, andFirmSize
Informal 225
surveyofinnovation diffusion
showeda consider- innovationwere analyzed,with a varietyof
able degree of innovativeness among Italian comprehensive information.
SMEs: onehigher, forinstance,
thanthatrevealed As faras R&D is concerned, Table HI shows
bytheofficial R&D figures(Archibugi,Cesaratto,that the share of firmspossessingan R&D
and Sirilli,1988; Santarelliand Sterlacchini, department in 1985 wasverylowintheclasswith
1989). less than50 employees(4.39%) and markedly
Moreover, SMEs gaveR&D activity higher
relatively fortheclasswithmorethan500 employees
low priority as a sourceof innovation. In this(56.75%).The figures changesignificantlyifonly
regard,it shouldbe pointedoutthatthequestion the R&D performed in other departments is
on R&D in thefirstsurveywas too generical, soconsidered. Suchinformal R&D wasperformed in
thattheabove resultcan onlyserveas a rough 1985 by 11.5% of firmswith less than 50
approximation oftheR&D effort madebyItalian employeesand by 18.42% of thosewithmore
manufacturing firms.
As we shallsee in thenext than500 employees; theshareof informal R&D
section,thedifferentextentofR&D commitment washigher thanthatofformalR&D fortheclasses
betweensize classesemergesmuchmoreclearly of firmswithbetween50 and 199 employees.
whenbothformaland informal R&D activities Whentotal(formalplus informal) R&D is con-
aretakenintoaccount. sidered,theshareoffirms undertaking R&D still
A secondweaknessofthefirst surveylayinits increasedwithfirmsize,butthecommitment of
to provideanyinformation
failure concerning theSMEs was higherthanthefigureobtainedwhen
number ofinnovations adoptedby eachfirm. ThisonlyformalR&D wasexamined. TheZ-testyields
obviouslygave riseto an overestimation of thea 99% significance level for the differences in
roleoffirms thathadintroduced onlyone or a fewmeans betweenSMEs and large firmsin the
innovations duringtherelevant
period. category"Firmshavingan R&D department"
whilesuchdifferences are not significantin the
case of "Firmsperforming R&D in otherdepart-
ID. R&D activitiesand typesofinnovation
mentsonly."
The secondsurveywasbasedon a moredetailed This finding is consistent withthe resultsof
questionnairesent
or administeredbyinterviewto Kleinknechf s study, since it shows thata signifi-
the16,701manufacturing firmswhich,according cant number of small and medium sizedfirms are
to the firstsurvey,had introduced innovations indeedactivein R&D, eventhough, quiteoften,
duringthe period 1981-1985: 8,220 firms suchinnovative efforts are not capturedby the
returnedthe questionnaire. In particular,
both [Link],Mangano,and Sirilli
R&D activitiesand the numberand type of (1988) reportthatin theItaliancase 1,100firms
table m
Formalandinformal
R&D bysizeofinnovating
firms
(1985)
Firmsizes Firmsresponding Firmsperforming Firmshaving
an Firmsperforming
(employees) tothequestionnaire R&D (total) R&D department* R&D inother
departments
onlyb
Number% oftheclass Number% oftheclass Number% oftheclass
20 to49 3939 626 15.89 173 4.39 453 11.50
50 to99 1789 581 32.48 194 10.84 387 21.63
100tol99 1210 503 41.57 207 17.11 296 24.46
200 to499 815 477 58.53 261 32.02 216 26.50
500 andmore 467 351 75.16 265 56.75 86 18.42
Sourceofdata:ISTAT-CNR(1988).
■ Whichcan outR&D eveninotherdepartments.
carry
b andotherdepartments.
Design,production,
This content downloaded from [Link] on Wed, 18 Jun 2014 [Link] AM
All use subject to JSTOR Terms and Conditions
226 andAlessandro
EnricoSantarelli Sterlacchini
were included in the 1985 officialR&D survey ployees, the number of R&D personnel fall
while,forthe same year,the ISTAT-CNR survey between 8 and 20 units,while in the class with
revealed some 2,874 firms engaged in R&D morethan500 employeesthereare,on average,at
activities(formal and informal).4Althoughit is least 21 R&D employeesper [Link] numbers
not possible, at this stage, to make accurate are large enough to indicatethe presence of an
comparisonsin termsof R&D expendituresor R&D laboratoryor, at any rate,the existenceof
equivalentman years,this surveyshows thatthe structured R&D activitiesin firmswithmorethan
officialfiguresconcerningItalian manufacturing 200 employees.
firmsare probablyseriouslyunderestimated. In other words, the average numberof R&D
Table IV gives two measures of total R&D employeesin small firmsis verylow when con-
intensityby size class: the firstbeing the ratio of sidered in absolute [Link] suggeststhatthe
R&D man years (in 1985) to the number of R&D carried out by small firmsis more often
employeesin all firms,the second the same ratio organizedon an occasional,unstructured and less
to die numberof employeesof thefirmsperform- systematic basis than it is in medium and large
ing some R&D. The firstindicatorincreaseswith firms. Accordingly,even if the ISTAT-CNR
size class, thus confirmingthat the number of surveyrevealsthatthereis moreR&D undertaken
SMEs whichperformno R&D at all is higherthan by small firmsthanthatmeasuredby the official
that of large firms; by contrast,the second R&D survey,it mustbe stressedthatthisR&D is
[Link] Kleinknecht(1989) points less significant(in both a technologicaland an
out, this latter measure probably overestimates economic sense) than the R&D carried out by
the R&D activitiesof SMEs, since those which mediumand [Link] our view,itis thelevel
returnedthe questionnairecan be assumed to be of systematicR&D, both formal and informal,
themostinnovativefirmswithintheclass. thatconstitutesthe crucialindicatorof innovative
This,however,is not thewhole [Link] fact,if [Link] fact,systematicR&D enables major
withinthe firmswithless than 50 employeesthe innovationsto be introduced,and it is around
mean R&D intensity is 9.6%, thismeansthatthey thesethatincrementalinnovationsare developed.
have, on average, between 2 and 4 R&D em- These lattermay also be implementedby firms
ployees. Analogously, the number of R&D per- whichdo nothave structured R&D, buttheyoften
sonnelrangesbetween3 and 6 unitsin firmswith cannotoccurwithouttheformer.6
50 to 99 employees,and between4 and 9 unitsfor Supportforthisassertionis providedwhenwe
the class with 100 to 199 employees.5This casts examine the number and type of technological
some doubt on whetherfirmswithless than 200 innovationsintroducedby Italian manufacturing
employees can undertakeR&D in a systematic firmsduringtheperiod 1981-85. Table V shows
way. Conversely,in firmswith 200 to 499 em- the average numberof product innovationsper
firmin each size class. Product innovationsare
divided into three groups on the basis of their
TABLEIV
R&D performance
bysizeclass(1985)
TABLE V
Firmsizes MeanR&D intensityMeanR&D intensity Type of productinnovations
by size of innovating
firms
(averagenumberperfirm)
(employees) ofallfirms" offirms R&Db
having
Firmsize Productsnew Products Improvements
20 to49 1.72 9.64
(employees) forthesector newfor ofexisting
50 to99 2.16 6.70 orforItaly thefirm products
100to 199 1.94 4.61
200 to499 2.48 4.10
500 andmore 3.87 4.30 20 to49 1.24 3.02 2.74
50 to99 1.75 3.68 2.92
100 to 199 2.40 3.53 3.32
Sourceofdata:ISTAT-CNR(1988). 200 to499
a Man 2.53 3.52 4.82
yearsof R&D on the numberof employeesof all 500 andmore 6.55 5.48 6.62
firms.
b Man
yearsof R&D on thenumberof employees
offirms
R&D.
having Sourceofdata:ISTAT-CNR(1988).
This content downloaded from [Link] on Wed, 18 Jun 2014 [Link] AM
All use subject to JSTOR Terms and Conditions
Innovation, R&D, andFirmSize
Informal 227
technological andeconomicimportance. Themost withless than500 employees appearto be more
significantinnovations are identified as products innovative thanis usuallybelievedto be thecase.
thatare neweitherforItalyor forthesectorof This finding has emergedfromthetwoISTAT-
activityofthefirm. According tothisindicator, the CNR surveyson innovation diffusionin Italian
significance of innovations increaseswithfirm manufacturing firms;surveyswhichgivea more
[Link] particular, whereasthereare smalldiffer- reassuring profileof the technological level of
encesbetweenthe size classesfrom20 to 499 Italianindustry thanthatprovidedbytheofficial
employees, firms withmorethan500 employees survey.
introducedan average numberof significant However,thereis stilla significant difference
productinnovations (6.55) whichwas farhigher betweenItalianSMEs and largefirms, sincethe
thanthefigureforSMEs (ranging between1.24 formerare likelyto undertakeR&D activities
and2.53).7 whicharelesssystematic thanthosecarriedoutby
SMEs insteadperformcomparatively better [Link] corresponding innovations reflecta
whenproducts newforthefirm andimprovements similardifference in termsof "quality"of R&D
of the existing productsare considered.8 When andthecontribution
activities, oflargefirms tothe
Table V is used as a matrixfortheanalysisof of
development major innovationsis more signifi-
variance(Spiegel,1975), the differences in the cantthanthatofSMEs.
rowmeansare notsignificant (according theF
to In conclusion, whilstthecommitment ofItalian
testat 99% level)ifonlythelasttwocolumnsare SMEs to informal R&D is evidencethatthey
considered("Productsnew for the firm"and devote considerableeffortto improvingtheir
"Improvements of existingproducts"), although internallevelsofinnovative activity,suchfirms do
they turn out to be significant whenall thethree notseemto haveengagedina systematic research
columns [Link] that [Link], largefirms engagein institu-
whentheproducts newforthesectoror forItaly tionalizedsearchfor innovationwhich,in the
are considered, thedifferences in innovativeness Italiancase, is associatedwithhigherinnovative
of Italianfirmsare significantly associatedwith performance.
firmsize.
Thesefindings are consistent withour above
the nature Notes
assumption concerning non-systematic
ofR&D activities carriedoutbysmallfirms. The * We wishto thankD.
Archibugiand S. Cesaratto(Istitutodi
innovations implemented by small firms are often Studi sulla Ricerca e Documentazione Scientifica,C. N. R. -
ofincremental type,andtheirintroduction andde- NationalResearch Council) who providedus withsuggestions
do notrequireanyparticular and some unpublisheddata fromtheISTAT-CNR surveyson
velopment generally [Link] commentsfrom Alfred Klein-
commitment tostructured R&D. knecht and an anonymous refereeare gratefullyacknowl-
However,thisdoes notentitleus to conclude [Link] usual disclaimersapply.
thatthereis a close cause-effect relationship
1
See, amongothers,Freeman(1982) and Pavitt(1982V
2 See also Kleinknecht
betweenthe"quality" of R&D activities and the (1987) fora shortnote on the same
"quality"of innovations. In effect,no direct subject.
3 This is not to
between innovations andR&D datais say thatsmallfirmshave a purelypassive role
comparison in the process of [Link] the speed and
possible,because the formerrepresent innova- theeffectiveness of innovationdiffusionis based upon a close
tionsintroduced in theperiod1981-1985 while interaction between producers and users of innovative
thelatterrelateto 1985 [Link], it is intermediateand capital goods. Moreover, withoutfurther
reasonable to assumethatthetotalR&D commit- changesin thearea of organizationalprocedures,forinstance,
it is difficultfor small "supplier-dominated"firms(Pavitt,
mentamongfirms ofdifferent sizesdidnotchange
1984) to make significant improvementsin theirproductsor
duringtherelevant period. processes simplybyintroducing new machinery.
4 These
figuresalso include firmsbelongingto the service
[Link] therefore are higherthanthose reportedin Table
IV. Conclusion HI, whichrefersonlyto manufacturing [Link] Cesaratto et
al. (1988, p. 24) point out: "This large differenceis mainly
Thepaperhasshownthattheinnovative capability due to the differentsurveysystemsused: in the surveyon
of firmsof differentsizes does not necessarily R&D a complex questionnaire is used which requires a
[Link] particular,
firms considerableamountof effortto fillin; althoughreferring to
This content downloaded from [Link] on Wed, 18 Jun 2014 [Link] AM
All use subject to JSTOR Terms and Conditions
228 EnricoSantarelli Sterlacchini
andAlessandro
the same definitions as the previousones, the innovation Archibugi, D., S. Cesaratto,and G. Sirilli,1987, 'Attivita
questionnnaireaskeda fewsimplequestions whichcaneasily innovativa,R&S e brevetti:Un'analisi dei risultati
be answered bythefirm's management. Therespondent'stask dell'indagineCNR-ISTAT sulla diffusione dell'innova-
is thusmuchsimplified." zione',Ulndustria Vm(4), 497-514.
5 In thecase ofsmallfirms itis likelythattheactualaverage Cesaratto, S., S. Mangano,and G. Sirilli,1988,SomeResults
number ofR&D employees (in termsofmanyears)is closer of the Surveyon Technological Innovationin Italian
to the lowerratherthanto the [Link] Industry, Paris,OECD, Groupof NationalExpertson
(1989) showsthatin Dutchindustry theaveragenumberof ScienceandTechnology Indicators, RoomDocumentNo.
R&D employeesis 1.5 for the firmswithless than 50 10.
employeesand about 3 for the firmswith 50 to 199 Freeman, C, 1982, TheEconomicsofIndustrial Innovation,
employees. Cambridge, MA,MTTPress.
6 Thisis notto do not ISTAT-CNR, 1987, Indaginesulla diffusione delVinnova-
implythatincremental innovations
deservean important (sometimes autonomous) role,bothin zione tecnologica nella industria manifdtturieraitaliana,
theprocessof diffusion and in termsof feedbackson the Rome.
majorinnovations theyoriginate from. ISTAT-CNR, 1988, Indagine statisticasulVinnovazione
7 Mediumsizedfirmswith200 to 499 neWindustria
italiana,Rome.
employeesperform tecnologica
betterthan small firmsonly in the case of
significantly Kleinknecht, A., 1987,'Measuring R&D inSmallFirms:How
incrementalinnovations. MuchAreWe Missing?', Journal ofIndustrialEconomics
8 The
averagenumberofprocessinnovations introduced by 36(2), 253-256.
SMEs is evenhigher, butit mustbe noticedthatthisis the Kleinknecht, A., 1989,'FirmSize and Innovation: Observa-
mostfrequent typeof innovation introduced byfirms in all tionsin DutchManufacturing Industry',Small Business
thesize [Link] effect,
processinnovations, on average, Economics1 (3), 215-222.
accountfor38.9% of totalinnovations introducedbyfirms Pavitt,K., 1982,'R&D, Patenting andInnovative A
Activity:
withlessthan50 employees and,forinstance, 34.3%ofthose Statistical Research
Exploration', Policy11(1), 33-51.
introducedbyfirms withmorethan500 employees. Pavitt,K., 1984, 'SectoralPatternsof TechnicalChange:
Towardsa Taxonomyand a Theory',ResearchPolicy
13(6), 343-373.
Santarelli,E., and A. Sterlacchini, 1989, 'Patternsregionah
References di innovazionetecnologicain Italia: R&S, brevetti
sull'estero, innovative',
imprese L'IndustriaX(l),25-56.
Acs,Z. [Link]. B. Audretsch, andLarge
1988,'Innovation Spiegel,R. M., 1975, Probability New York:
and Statistics,
and Small Firms:An EmpiricalAnalysis',American McGraw-Hill.
EconomicReview78 (4), 678-690.
This content downloaded from [Link] on Wed, 18 Jun 2014 [Link] AM
All use subject to JSTOR Terms and Conditions
