View
2
Download
0
Category
Preview:
Citation preview
AREROBOTSSTEALINGOURJOBS?
SocialSituationMonitor-ResearchSeminar
Brussels,10November2017
MarcoVivarelli(UniversitàCattolicadelSacroCuore,Milano;IZA,Bonn;MERIT,Maastricht)
Todayalarm…
• Recently, the arrival of 3D printing, self-driving autonomous cars (Tesla, Apple,Google) and widespread robots has raised again a fear of a new wave of‘technologicalunemployment’.
• Brynjolfsson and McAfee, A., 2011, “Race Against the Machine: : the root of thecurrent employment problems is not the Great Recession, but rather a “GreatRestructuring” characterized by an exponential growth in computers’ processingspeedhavinganever-biggerimpactonjobs,skills,andthewholeeconomy.
• Moreover,notonlyagriculturalandmanufacturingemploymentappearsatrisk,butemployeesinservices-includingcognitiveskills-arenolongersafe(see,forinstance,how Uber - just a software tool - is fully crowding out taxi companies). Frey andOsborne (2013) predict that 47% of the occupational categories are at high risk ofbeingautomated,includingawiderangeofservice/white-collar/cognitivetaskssuchasaccountancy,logistics,legalworks,translationandtechnicalwriting,etc.
Itisanatavisticfear:NedLuddandCaptainSwingagainsttextile
andthreshingmachineries
CaptainSwingbyEricHobsbawmandGeorgeRudé,Newyork,PantheonBooks,1968
RICARDO’SSURPRISE
However, technological unemployment is considered an
exception, occurring only when output does not grow,otherwisea“compensation”(throughpriceandincomemarketmechanisms)alwaysoccurs.
“….the opinion, entertained by the labouring class, that theemployment of machinery is frequently detrimental to theirinterests,isnotfoundedonprejudiceanderror,butisconformabletothecorrectprinciplesofpoliticaleconomy”(Ricardo,1951,vol1,p.387;thirdedition,1821)
“Thisneo-classicalgeneralequilibriumframeworkcanbesaidtocorrespondmostcloselytopresent-daytraditionaleconomic views on technical change and employment.Technological change may indeed result in sometemporaryunemployment, butwithefficientlyoperatinglabour and capital markets there is no basic economicproblemarisingfromtheintroductionofnewtechnology”(Freeman,C.andSoete,L.,WorkforAllorMassUnemployment,London:Pinter,1994,p.25)
ACRITIQUE
EMPIRICALLY:
INNOVATION-INPUT
R&D
ETC
PROD
PROC
PROD&PROC
INNOVATION-OUTPUT
JOBCREATION
JOBDESTRUCTION
PREVIOUSMICROECONOMETRICSTUDIES(1)
The advantage of the firm-level analysis is the possibility to better proxy technological change andinnovationand todealwith largedatasets; thedisadvantage is thatwecannot take intoaccount thecomplex(intersectoral)natureofthecompensationtheory.
CROSS-SECTIONSTUDIESEntorf-Pohlmeier,1990:positiveimpactofproductinnovation,WestGermany.Zimmermann,1991:negativeimpact,WestGermany.Klette-Førre,1998:notclear-cut(negative)impactofR&Dintensity,Norway.Brouweretal.,1993:negativeeffectofR&D,positiveofproductinnovation,theNetherlands.Cross section analyses (mainly based on OLS and or probit) are severely limited by endogeneityproblems, cannot take intoaccount theunobservables andmayover-estimate thepositive impactofinnovationbecauseofthebusinessstealingeffect.Since the second half of the ’90s, attention has been moved to longitudinal datasets and panelmethodologies(GMM-DIF;GMM-SYS;LSDVC).
PREVIOUSMICROECONOMETRICSTUDIES(2)
LONGITUDINALSTUDIES
VanReenen,1997:positiveimpactofinnovation,UK.
Domsetal.,1997:positiveeffectofadvancedmanufacturingtechnologies,US.Smolny,1998:positiveimpactofproductinnovation,WestGermany.
GreenanandGuellec,2000:positiveeffectofinnovationatthefirm-level,butnegativeatthesectorallevel(stillpositiveforproductinnovation),France.
Greenhalghetal.,2001:positiveimpactofR&D,UK,butonlyintheHigh-Tech.
PivaandVivarelli(2005):positiveimpactofinnovation,Italy.Harrisonetal. (2008):positiveeffectofproduct innovationand(slightly)negativeofprocess innovation(strongcompensationinservices),Germany-France-UK-Spain.
Halletal(2008):positiveimpactofproductinnovation,Italy.
Lachenmaier and Rottmann (2011): positive impact of innovation (including process innovation), no sectoraldifferences,Germany.CoadandRao(2011),positiveimpactofinnovation,strongerforfast-growingfirms,US(dataonlyfromhigh-techmanufacturing).
Bogliacinoetal(2012),positiveimpactofR&D,butonlyinservicesandhigh-techmanufacturing,notinthemoretraditionalmanufacturingsectors.
EMPIRICALTESTS
• H1:consistentlywiththepreviousliterature,R&Dexpendituresshouldberelatedtoanincreaseinemploymentatthefirm’slevel;
• H2: in contrast, ETC should be related either to a decrease in firm’semploymentorshoulddisplayanon-significanteffect(mainnovelty);
• H3: consistentlywith theprevious literature, innovationvariablesshouldbemorepositivelyrelatedtoemploymentinthehigh-techsectorsratherthaninthelow-techones;
• H4:innovationvariablesshouldbemorepositivelyrelatedtoemploymentin the large firms rather than in the SMEs, which are more commonlylocated in low-tech sectors and are dominated by ETC and processinnovation.
THEITALIANDATASET• CISdatarepresentativeatbothsectoralandfirmsizeleveloftheentirepopulationofItalian
companieswithmorethan10employees.• Four CIS waves: CIS3 (1998-2000 period), CIS4 (2002-2004), CIS6 (2006-2008) and CIS7
(2008-2010); CIS5 has been excluded since it was mainly conducted through anadministrativeway,resultinginincompleteandmainlyinterpolateddata.
• Inordertogetrelevanteconomic information,CISdatahavebeenmergedwiththe ItalianStatisticalBusinessRegister(ASIA),createdbytheItaliannationalstatisticaloffice(ISTAT).
• Sub-sampleofinnovators:firmsdeclaringthattheyhadintroducedeitherproductorprocessinnovation,orhadstartedinnovativeprojects.
• Final workable sample: 265 medium-large manufacturing firms over 4 periods (947observations).
• Duetoanexcessivecollinearitybetweenvalueaddedandcapitalformation(ρ=0.83)inthisdataset, we decided to test a simplified version of the ideal specification, dropping theinvestmentvariable.
• HypothesesH3andH4willbe investigatedusingtheOECDclassification (Hatzichronoglou,1997)splittingmanufacturingsectorsintohigh-andlow-techsectors,andtheEUthresholdof250employeessplittingfirmsintosmallandmediumenterprises(SMEs)andlargeones.
THESPANISHDATASET• Survey on Business Strategies (Encuesta Sobre Estrategias Empresariales, ESEE)
gathering extensive information on around 2,000 manufacturing companiesoperating in Spain and employing at least ten workers, representative for eachtwo-digitNACE-CLIOsector.
• In this study we use data for the period 2002-2013 and select our workingdatabasefromaninitialsampleof63,648firm-yearcells.
• Firstly,wechecked formissingvalues for thevariables relevant toourempiricalanalysis; secondly, we discarded all firms involved inM&A; thirdly, we retainedonly those firms that have invested inR&D at least once in their life and haveinvestedinETCatleastonceintheirlife;Fourthly-giventhetargettoestimateadynamicequation-weretainedonlyfirmsforwhichtwoconsecutivelagsofthedependentvariablewereavailable.
• Finally,R&DhasbeenlaggedtwoyearswhileETConeyear;thisstructureoflagstakes into account that innovationmay need some time to have an impact onemploymentand that thisdelay is shorter for ETC - that is directly embodied innewmachineries-whilelongerforR&Dexpenditures.
• Weendedupwithanunbalancedpanelof517firmsover12years.
RESULTS• H1:thishypothesisisweaklyconfirmedbytheestimatesbasedontheItaliandataand
not confirmedwhen the Spanish data are used; on thewhole, a generalized labour-friendlynatureofR&Dexpenditures is notdetectable; in case, themagnitudeof thepositiveimpactisverysmall.
• H2:thishypothesisisconfirmedbyourestimates:ETCneverexhibitsalabour-friendlynatureand inone case (within the Spanish SMEs) turnsout to generatea significantlabour-savingimpact.
• H3:thishypothesisisstronglyconfirmedonthebasisofourregressions:intheItaliancase,thepositiveemploymentimpactofthetotal innovationexpendituresandofthesoleR&Dexpendituresistotallyduetohigh-techfirms,whileintheSpanishcase,thejob-creation impactofR&Dexpenditures (afterbeingassessedasnot significantwithregard to theentire sample)becomeshighly significantwhenattention is focusedonthehigh-techfirms.
• H4: thishypothesis isconfirmedbyourestimatesbasedonthe ItalianCISdata (bothtotalinnovationexpendituresandR&Daresignificantwithregardtothelargefirmsandnot significant for the SMEs),but not by the regressions based on the Spanish ESEEdatawhere theR&Dvariable is not significant for both the large companies and theSMEs.
• The job-creation impact is often negligible in magnitude. • However, R&D may fosters labor-friendly product innovation that
leads to job creation. R&D subsidy is a good policy. • The job-creation impact of innovation is limited to product innovation
and to the high-tech sectors. Need for structural policies. • Process innovation may displace labor and create technological
unemployment. Need for labor and education/training policies. • Moreover, safety nets are necessary for the possible job losses due
to process innovation in non-high-tech sectors and SMEs. • Industrial and innovation policies that support R&D and product
innovation, especially in high-tech sectors, can foster job creation.
KEYFINDINGSANDPOLICYIMPLICATIONS
REFERENCES
• Vivarelli, M., 1995. The Economics of Technology and Employment: Theory and EmpiricalEvidence.Aldershot:Elgar.
• Piva,M., Vivarelli,M., 2005. Innovation and employment: Evidence from Italianmicrodata.JournalofEconomics,86,65-83.
• Piva, M., Santarelli, E., Vivarelli, M., 2005. The skill bias effect of technological andorganisationalchange:Evidenceandpolicyimplications.ResearchPolicy,34,141-157.
• Bogliacino,F.,Piva,M.,Vivarelli,M.,2012.R&Dandemployment:AnapplicationoftheLSDVCestimatorusingEuropeandata.EconomicsLetters,116,56-59.
• Vivarelli,M.,2014.Innovation,employmentandskillsinadvancedanddevelopingcountries:Asurveyofeconomicliterature.JournalofEconomicIssues,48,123-154.
• Barbieri, L., Piva, M., Vivarelli, M. (2017), R&D, Embodied Technological Change andEmployment:EvidencefromItalianMicrodata.IndustrialandCorporateChange,forthcoming.
• Pellegrino,G.,Piva,M.,Vivarelli,M.(2017),AreRobotsStealingOurJobs?,IZADPNo.10540,IZA,Bonn.
Recommended