CentrePieceeprints.lse.ac.uk/47044/1/CentrePiece_12_2.pdfCentrePiece Autumn 20074 Another striking...

C e n t r e Pi e c eISSN 1362-3761

The Magazine of The Centre for Economic Performance Volume 12 Issue 2 Autumn 2007

UNIVERSITYECONOMICS■ Incentives for inventionand technology transfer■ The booming labourmarket for graduates■ Reforming highereducation in Europe

ALSO IN THIS ISSUE:Mobile phonesManagement practicesBritain’s trade unionsMonetary policy-makingAustralia’s economy

What contribution do universities make to innovation, productivity and long-runeconomic growth? At the core of their activities are two ‘products’: scientific research and graduates. And as is made clear in the government-commissioned review of the UK's scienceand innovation system, led by Lord Sainsburyand published in early October alongside the comprehensive spending review, both of these outputs are essential forsustaining and improving the nation’seconomic performance.

This issue of CentrePiece explores some ofthe central policy concerns facing institutionsof higher education and their politicalmasters. Nick Butler makes the case foruniversities to have greater independenceand resources. Stephen Machin and SandraMcNally note shortages of graduates in keysubjects, notably science, engineering andtechnology. And Mark Schankerman raisesquestions about the institutional frameworkfor scientific endeavour and thecommercialisation of new inventionsthrough licensing to the private sector – so-called ‘knowledge transfer’.

Knowledge transfer has been central to thework of the Centre for EconomicPerformance (CEP) since its inception –though the long track record of CEPinfluence is more evident in the world ofpublic policy. Stephen Nickell’s account hereof his time on the Bank of England’sMonetary Policy Committee demonstratesthis wider social impact of economicresearch, in terms of the transfer of bothpeople and ideas.

Perhaps less well known is the Centre’sinvolvement in knowledge transfer to theprivate sector. Business-supported researchinitiatives include the Manpower HumanResources Laboratory and a programme onnew technology and productivity sponsoredby EDS. And last year, CEP collaboratedwith McKinsey & Company on a secondsurvey of management practices coveringthousands of firms in Europe, Asia and theUnited States, the first results of which aresummarised here.

CEP’s research on and with the privatesector tends to have a long-run focusthrough the generation of ideas and new

policies, such as changes in competitionpolicy stimulated by Stephen Nickell’s work in the mid-1990s. One recentexample is a study of regulation of Europe’s mobile phone industry described inthis CentrePiece.

Research reported in the previous issueexamines the European Commission’srecently concluded case against Microsoft.The court ruling that the software gianthad abused its monopoly power – and thecareful use of economic arguments insupport of this view – is an example of theeffects that the output of universities canhave on innovation, particularly in the high-tech industries that the Sainsbury review iskeen to promote.

As always, your feedback on CentrePieceis welcome. And do pass the magazineonto colleagues – whether they’re in auniversity or making use of the coreproducts of universities.

Romesh [email protected]

CentrePiece is the magazine of the

Centre for Economic Performance at the

London School of Economics. Articles in this

issue reflect the opinions of the authors, not

of the Centre. Requests for permission to

reproduce the articles should be sent to the Editor

at the address below.

Editorial and Subscriptions Office

Centre for Economic Performance

London School of Economics

Houghton Street

London WC2A 2AE

Annual subscriptions for one year (3 issues):

Individuals £13.00

Students £8.00

Organisations (UK and Europe) £30.00

Rest of world £39.00

Visa and Mastercard accepted

Cheques payable to London School of Economics

CEP director, John Van Reenen

CEP research director, Stephen Machin

Editor, Romesh Vaitilingam

Design, Raphael Whittle

Print, Ghyllprint Ltd

Cover image, Ron Achin

© Centre for Economic Performance 2007

Volume 12 Issue 2

(ISSN 1362-3761) All rights reserved.

Centre Piece

Editorial

CentrePiece Autumn 2007

1

page 2 Harnessing success: incentives for invention andtechnology transfer in universitiesMark Schankerman investigates the motivations for producing new,commercially valuable, scientific knowledge

page 6 Higher education and the labour marketThere are no problems of graduate over-supply or over-qualification,according to Stephen Machin and Sandra McNally

page 12 What drives good management around the world?CEP’s global survey of over 4,000 firms reveals huge variations in thequality of management practices

page 22 Life on the Monetary Policy CommitteeStephen Nickell reflects on his experiences setting the nation’s officialinterest rate

page 25 Regulating the mobile phone industry:beware the ‘waterbed’ effectChristos Genakos and Tommaso Valletti examine the impact of recentprice cap regulation

Contents in brief...

page 10 Europe’s universities – time for reformNick Butler outlines what must be done toimprove the quality of higher education inthe European Union

page 18 OzonomicsAndrew Charlton explores the myth ofAustralia's economic superheroes

page 20 Union bluesAlex Bryson and Paul Willman chart thecontinuing decline of Britain’s trade unions

page 6 Highereducation andthe labourmarket

page 18 Ozonomics


2

Universities are a key source of the newscientific knowledge that drives long-runeconomic growth. But what are the incentivesfor scientists to generate commercially valuableinventions and for university managers tolicense such technologies to the private sector?Research by Mark Schankerman andcolleagues investigates.

Harnessing success:incentives for invention and technology transfer in universities

Perhaps the greatest long-term productivity advancescome throughbreakthroughs in basicknowledge – and a

substantial proportion of the research anddevelopment (R&D) that creates newknowledge and leads to increasedproductivity is done in universities.

University research not only raises theproductivity of private sector R&D (through‘knowledge spillovers’) and encouragesmore of it to be done; it also leads toinventions that can be commercialised,either through licensing to private firms orvia the formation of new start-upcompanies.

Such ‘technology transfer’ byuniversities has grown dramatically in thepast two decades, particularly in the UnitedStates. Between 1991 and 2004, thenumber of US university patentapplications rose from 1,584 to 10,517,and licensing income increased from $218million to $1.4 billion (which is 6% offederal R&D financing for universities).

European and Asian universities are lessinvolved in this form of technology transferbut are rapidly expanding their activities.

The rapid growth of technologytransfer in the United States is in part dueto the Bayh-Dole Act of 1980. This piece oflegislation not only gave universities theright to patent new discoveries but alsomandated them to license inventions madewith federally sponsored research to theprivate sector. Now, nearly all US researchuniversities have a technology licensingoffice and explicit intellectual propertypolicies and royalty-sharing arrangementsfor their scientists.

Analysing technologytransferIn essence, technology transfer involvestwo distinct activities: innovation byuniversity scientists and commercialisationby the university’s technology licensingoffice (see Figure 1). In the first stage,scientists produce both publications andinventions. The mix of these two may beinfluenced by the incentive of money, Im

age:

Ron

Ach

in

either for themselves directly or asenhanced funding for their researchlaboratories. Other incentives – such aspromotion and tenure rules, and intrinsicmotivation to do basic and appliedresearch – are also likely to play a role.

The second stage is thecommercialisation of inventions by thetechnology licensing office, which decideswhether to patent and license inventions,identifies licensees and structurescontracts. The effectiveness of thetechnology licensing office is likely to beinfluenced by the university's objectives,government constraints on licensing andincentives given to its managers.

Our research programme is studying

the role of incentives and otherinstitutional features that can make theprocess of technology transfer moreeffective. Given the importance ofresearch for long-term growth, it is criticalto understand what drives scientificendeavour and technology licensingactivity.

Is research a purely intellectual pursuitdriven by intrinsic motivation, or doeconomic incentives play a role in the waythat scientists structure their work? Whatare the most appropriate incentives formanagers in technology licensing offices?And how is technology transferperformance influenced by whether auniversity is private or public (and henceconstrained by government objectives) andthe degree to which it chooses, or isobliged, to promote local and regionaldevelopment? Our research programmeexplores all of these questions.

Royalty incentives forscientists In a study with Saul Lach, I use US data (assimilar data are not yet available inEurope) to examine how the share oflicensing royalties from universityinventions received by academic inventors(their ‘cash flow rights’) affects thenumber and licensing value of inventionsin universities.

Our central finding is that incentivesare effective: universities that give greaterroyalty incentives do much better in termsof licensing income from technologytransfer. This works both by inducinggreater effort by researchers and through‘sorting’ of the most productive andentrepreneurial scientists into high-royaltyuniversities. We also find that royaltyincentives have a much larger impact inprivate universities than in public ones,and technology licensing activity is morecommercially effective in the former.

In the United States, universitiesusually claim exclusive ownership (‘controlrights’) over inventions made by theirscientists. But the cash flow rights fromlicensing inventions are typically sharedbetween the inventor and various parts ofthe university according to specifiedroyalty-sharing schedules. There issubstantial variation in these arrangementsacross US research universities, whichmakes it possible to estimate their effecton inventive output.

Our study focuses on two outcomes –licensing income and the number ofinventions disclosed by faculty scientists totechnology licensing offices – using datafrom the Association of UniversityTechnology Managers, combined withinformation on the distribution of royaltyshares for 102 US universities during theperiod 1991-99.

The novel aspect of the data is theinformation on the distribution of licensingincome between the university and theinventor(s). The inventor retains a givenpercentage of net licensing income andthe rest is allocated to the inventor'slaboratory, department and college and tothe university. Our criterion for identifyingthe inventor’s share is that the inventormust gain either cash flow rights or musthave direct control rights over the income(for example, lab research money).

In about half the universities, theseroyalty shares vary with the level oflicensing income generated by an invention(‘non-linear royalty schedules’). Theaverage inventor's share is 41% amongthe 58 universities using linear royaltyschedules, but there is substantial variation:the minimum inventor royalty share is 25%and the maximum 65%. The royalty sharesin the 44 universities with non-linearschedules display even larger variability: theaverage royalty share is 51%, but theminimum is 20% and the maximum 97%.


3

University scientists

University technologylicensing

office

‘Knowledge spillovers’

Corporate research funding

‘Market for technology’

Licensed inventions

Start-up companies

Venturecapital

Figure 1:University-private sector science links

Scientists aremotivated byboth love of researchand potentialmonetaryrewards


4

Another striking feature is that inevery university, inventor royalty shares areeither constant or decline with the level oflicensing income per invention. Onaverage, they start at 54% and decline to30% for inventions generating over $1 million. Royalty shares are alsounrelated to various characteristics of theuniversities, such as faculty size, academicquality and the number of technologylicensing office professionals per faculty.

Among the more detailed findings ofour research:

■ Academic research and inventive activityin universities respond to variations ininventors' royalty shares. Controlling fora variety of factors – including universitysize, quality and R&D funding –universities with higher royalty sharesgenerate higher levels of licensingincome. This finding is importantbecause it implies that the design ofintellectual property rights and otherforms of incentives in academicinstitutions can have real effects.

■ Inventors respond both to cash royaltyshare and to royalties used to supporttheir research labs (when the scientistshave direct control over their use). Thus,both high-powered monetary incentivesand intrinsic motivation seem to play arole. This is relevant to the design ofuniversity royalty-sharing arrangements.For example, non-science faculty mayview generous payments to supportresearch labs as less objectionable thandirect cash payments to the scientists.

■ The incentive effects of royalty-sharingwork both by inducing greater effort byscientists and through sorting ofscientists across universities so that the

most productive and entrepreneurialscientists tend to work in higher royaltyuniversities.

■ The response to incentives is muchlarger in private universities than inpublic ones. If universities do not expecta strategic reaction from theircompetitors, the research indicates thatin most private universities, and in abouthalf the public ones, the incentive effectis strong enough to produce a ‘Laffereffect’, where raising the inventor'sroyalty share actually increases thelicense revenue retained by the university(net of payments to inventors).

■ But when universities expect competinguniversities to match changes in theirroyalty share, the benefits to theuniversities of raising inventors’ royaltyshares will be smaller. Thus, high-powered, invention-based incentives areimportant, but so too is the strategicbehaviour among universities in settingthese incentives.

■ Technology licensing offices are moreproductive in private universities,suggesting that private institutions havemore effective, commercially-orientedtechnology transfer activity.

Incentives for technologylicensing officesWhy might university ownership affecttechnology transfer performance? In astudy with Sharon Belenzon, I combineevidence from surveys of US universities’technology licensing offices with paneldata on licensing performance to addressthis question.

Whereas previous research has shownthat technology transfer performance isinfluenced by university characteristics andother factors, including universityownership (public versus private),academic quality, local (high-tech) demandconditions and licensing contract design,our study focuses more on the `black box'of productivity within the technologylicensing office. We examine three keydeterminants of technology licensingproductivity – performance pay; localdevelopment objectives; and governmentconstraints on licensing activity –combining new survey data with paneldata from public sources on 86 USuniversities for the period 1995-99.

The survey data show that universities’two main objectives are generatinglicensing income and promoting local and

regional development, the latter goalbeing more prominent in publicuniversities. Institutions that view localeconomic development as one of theirprimary functions might performdifferently from those that exclusivelypursue income maximisation.

Public universities are also moreaffected by the imposition by stategovernments of a variety of constraints –both statutory restrictions and informalpolitical pressure – on their licensingactivity. Our study quantifies the impact ofincentives and measures the implicit costof these constraints and of concentratingon local development objectives byestimating forgone licensing income.

We find that technology licensingoffices in private universities are muchmore likely to adopt incentive pay fortheir staff than those in publicinstitutions. But ownership does notaffect the licensing performance of thetechnology licensing office once the useof incentive pay is controlled for. From apolicy perspective, this means that itmight be possible to get ‘privateperformance’ from public institutions ifthe right incentives are introduced.

These incentives certainly matter: wefind that technology transfer performanceis strongly influenced by whether atechnology licensing office usesperformance-based pay for their staff.

We also find that technology transferperformance is affected by the extent towhich there is a preference for developinglicensing activity locally rather than morewidely, and by formal and informalconstraints imposed by government.Universities with a stronger localdevelopment focus earn far less licensingincome from a given pool of inventions.This raises important policy questionsabout the right balance between incomemaximisation and local development focusin technology licensing activity.

Among the more detailed findings ofour research:

■ Compared with technology licensingoffices in public universities, those inprivate universities are significantly morelikely to use performance-based pay.Among the private universities surveyed,79% use some form of incentive pay ascompared with only half of the publicuniversities.

■ Performance pay has strong incentive

The incentiveeffect ofroyalty-sharing forscientists ismuch largerin privateuniversities


5

This article draws on research reported in

‘Incentives and Invention in Universities’

by Saul Lach and Mark Schankerman,

CEP Discussion Paper No. 729

(http://cep.lse.ac.uk/pubs/download/

dp0729.pdf) and ‘The Impact of Private

Ownership, Incentives and Local

Development Objectives on University

Technology Transfer Performance’ by Sharon

Belenzon and Mark Schankerman, CEP

Discussion Paper No. 779 (http://cep.lse.ac.uk/

pubs/download/dp0779.pdf).

Mark Schankerman is professor of

economics at LSE and director of CEP’s

research programme on productivity and

innovation. Saul Lach is professor of

economics at the Hebrew University of

Jerusalem. Sharon Belenzon, a CEP research

associate, is a postdoctoral research fellow at

Nuffield College, Oxford.

effects. Universities that use bonus paygenerate, on average, about 30-40%more income per license, aftercontrolling for other factors.

■ While private ownership has a large,positive effect on the adoption ofincentive pay, ownership has noindependent effect on licensingperformance, once we have controlledfor whether the university has adoptedincentive pay.

■ Private universities are much lessconstrained in their freedom ofoperation by state laws and regulations,and are more likely to be focused ongenerating licensing income comparedwith more ‘social’ objectives such aspromoting local and regionaldevelopment.

■ Local and regional developmentobjectives are ‘costly' in terms offorgone license income. Universitieswith strong objectives of this kindgenerate, on average, about 30% lessincome per license, after controlling forother factors. State governmentconstraints also reduce license income.

The finding that local developmentobjectives are costly in terms of theforgone license income raises animportant policy question. There are twoeconomic arguments for having apreference for local licensing. First, pureknowledge spillovers have a tendency tobe geographically localised. Second, the

new economic geography literatureemphasises that growth can be stimulatedby agglomeration effects working throughvarious supply and demand linkages.

But by showing that there is anopportunity cost of promoting localdevelopment in this way, our researchhighlights the importance of comparingthis approach with an alternative policy ofmaximising income from universityinventions (with no preference for localdevelopment) and using the additionallicense income generated to finance localeconomic development in other ways – forexample, through lower business taxes ordirect subsidy programmes.

ConclusionsMany countries, in Europe and beyond,are increasingly concerned about how topromote more effective technologytransfer and other forms of researchcollaboration between universities (andother public research organisations) andthe private sector. Clear ownership rights,incentives and a clear definition of theobjectives of technology transfer are keyelements of that process.

Our research makes a contribution tothat public debate by showing that thebenefits to universities are stronglyaffected by how incentives are set and byidentifying characteristics of technologylicensing offices that influence theeffectiveness of royalty incentives.

One caveat applies to all this work.Our findings contribute to the policydebate about the effectiveness ofuniversity licensing activity, but they arenot a cost-benefit analysis of the'commercialisation' of universities. Manyscholars have expressed concerns aboutthe potential costs of these developments,including the threat to established norms

of ‘open science’ and the potentialredirection of research away fromfundamental science. While there is onlylimited evidence of such costs thus far,continuing vigilance is needed to ensurethat they do not get out of hand.

Important challenges for research andpolicy remain:■ How should the ‘market for technology

licensing’ best be structured, and whatrole, if any, should government have inthat process?

■ Should universities have monopolycontrol over the inventions of theirscientists, as they currently do, orshould the scientists be free to markettheir inventions through otherchannels?

■ Should the use of market-based patentand licensing intermediaries be allowed(while preserving the sharing of cashflow rights between the scientist andthe university)?

■ How much geographic specialisationshould there be (for example, shoulduniversities join into regional technologytransfer offices, as in some countrieslike Germany?) and should such officesspecialise in particular scientificdisciplines?

■ In short, how should the market intechnology transfer be structured toexploit most effectively the economiesof scale and informational advantagesin these activities?

Universities with a strongerlocal development focus earnmuch less licensing income

Royalty incentives forscientists need to becombined withperformance incentiveswithin technologylicensing offices


Higher educationand the labour market

In recent decades, there hasbeen rapid expansion of higher(‘tertiary-level’) educationacross many countries. This hashad important and profoundeffects on labour markets and

the way in which employers use highlyeducated labour.

These expansions have, for the mostpart, been predicated on the assumptionthat more education is good for individualsand for society as a whole, not only interms of economic outcomes like wages oremployment, but also for a wide range ofsocial outcomes like improved health,reduced crime and higher well-being.

But along with expansion of thesystem has come a range of newquestions that have emerged as aconsequence of there being many moregraduates. Is there now ‘over-supply’ ofgraduates? Is there evidence of ‘over-qualification’ and skill mismatch? Are students studying the ‘right type’ ofsubjects? And is there a shortage of science and technology graduates in particular?

In a recent report, we review theevidence on these questions. The reportoffers some conclusions about the way inwhich the expansion of higher educationhas had important effects on economicoutcomes – and draws policy implicationsfor the future.

The increasing supply of graduatesThe labour market consequences ofincreasing supply can be considered withina simple demand and supply framework.Starting from a position where thedemand for and supply of graduates are inbalance, a boost in the supply ofgraduates should, other things beingequal, lead to a reduction in the wagepremium because employers have a widerrange of similarly qualified people to

choose from. But if, for whatever reason,employers demand more graduates, thenthe wage premium may not fall.

The wage premium depends on theinteraction of demand and supply. Inrecent decades, there has been a bigincrease in both the demand for andsupply of graduates. It is the fact thatdemand has outstripped supply that hasgiven rise to an increasing wage premiumfor a university degree. There is much controversy about the reasons for increasing demand for graduates, but the predominant view is that ‘skill-biased technological change’ is a majorcontributory factor.

In most countries, there has beencontinued expansion of higher educationin the last decade. But the wage premiumattached to higher education has increasedin most of them. The exceptions are Spainand New Zealand – two countries withparticularly large expansion of highereducation in the last 10 years – and Korea,where the wage premium declinedmarkedly between 1974 and 1990, aperiod of industrialisation when there wasmassive growth in higher education.

But even in these three countries,there is still a positive return to highereducation. Thus, it makes little sense tospeak of ‘over-supply’ of higher education.The strong, positive and (often) increasing

return to higher education suggests that‘under-supply’ is more of an issue and thatcontinued expansion is justified. In termsof employability, in many countries, therehas been some catch-up of less educatedgroups over the last decade, but graduatescontinue to have a much higherprobability of being in a job.

Mismatches and shortagesNevertheless, it sometimes takes a longtime for some (usually less wellperforming) graduates to find jobs afterleaving higher education and even then,some are not in jobs that appear to bewell matched to their qualifications. At thesame time, there are shortages in certainsectors: this is evident in employer surveysand in some data analysis that shows anegative wage premium associated with‘skill mismatch’.

A body of research has attempted tomeasure these outcomes, and the(sometimes misused) terms of ‘over-education’ and ‘under-education’ haveemerged: the former arises if an individualholds higher qualifications than requiredby his or her job whereas the oppositeapplies for the ‘under-educated’. Butstatistics on over- and under-education aredifficult to interpret as workers arematched to jobs based on a range ofcharacteristics and not just their education

Higher education around the world has expanded rapidly in recent years,yet graduates continue to command a wage premium in the labourmarket. So, as Stephen Machin and Sandra McNally show, there are noproblems of ‘over-supply’ or ‘over-qualification’ – rather there are‘shortages’ in some fields, which further expansion could alleviate.


7

Concerns about the ‘over-supply’ and/or

‘over-qualification’ ofgraduates are misplacedIm

age:

Kev

in L

au


8

level. What’s more, apparent mismatchmay be a temporary phenomenon.

The extent to which such problems areseen as temporary varies across studiesand countries. But one generalisation thatcan be made is that the fact of observing‘over-qualified’ individuals in theworkforce does not mean that there isover-supply of graduates. If there wereover-supply, relative wages andemployment probabilities would fall to thelevel of their closest substitutes – and thishas not happened.

The indications are that skill mismatch(or inadequate levels of skill) is more of aproblem than over-qualification. In somecountries, there is a need to improve thecontent and accreditation of vocationalqualifications so that they provide whatemployers need and are recognised to do so.

This is not to say that higher educationshould be geared to providing highlyspecific skills that are currently needed byemployers. Some studies suggest thatgeneral education and skills are morevaluable because they enable workers torespond to shocks to the economy (forexample, those that require sectoralchange) and advances in technology.

Degree subjectsOne hypothesis put forward to explain skillshortages is that individuals are notchoosing the right type of graduatestudies (whether this education isgeneral/academic or vocational). In otherwords, the choice of higher educationmade by individuals does not correspondto the needs of the labour market in termsof field of study.

As yet, there are relatively few studiesthat estimate returns to higher educationby subject of degree – especially when weare most interested in change over time.One study looks at changes in returns tosubject of degree over time in Britain,Germany, France and the United States,and finds that a return to an arts degreehad the lowest relative return within all countries, for two time periods (theearly 1990s and 2000) and for both menand women.

In contrast, the returns to degrees inscience, engineering and technology aresubstantial (especially for men). Suchfindings are broadly consistent with whatis found for a number of other countries –science/engineering/technology is often

among the category of subjects with arelatively high return (along with somesocial science subjects and professionssuch as law and medicine) whereas arts and humanities are often among the category of subjects with a relativelylow return.

So it may be relevant to talk ofgraduate over-supply in relation to somesubjects of degree. For example, therehave been estimates to suggest that thewage return to an arts and humanitiesdegree is zero in Britain.

This raises the question as to whypeople continue to pursue suchqualifications. There are various possibleexplanations: one is that wages do notcapture important aspects of the ‘value’ of

the degree for individuals – for example,higher education has a ‘consumption’value as well as a value in the labourmarket; and jobs have non-pecuniaryaspects that make them attractive toindividuals. Second, students may not bewell enough informed about the likelyreturns to subject of degree.

The value of science degreesThe existence of the relatively high wagedifferential for science/engineering/technology compared with other subjectsillustrates the high value placed on thefield by employers and indicates highrelative demand for graduates with this field of study. This might beinterpreted as a ‘shortage’ of science andtechnology graduates and would beconsistent with some reports of‘shortages’ that have appeared in severalcountries, including Australia, Belgium,Britain and New Zealand.

There are big differences betweencountries in the proportion of graduateswho qualify with a degree in science andtechnology. Comparing across continents(using data from 2000), Asia has thehighest percentage of graduates withscience and technology degrees (32%),which is just above Europe (28%) andconsiderably above North America (18%),South America (22%) and Oceania (22%).

Within Asia, China has a particularlylarge share of graduates with a degree inscience and technology (53%). Eventhough the EU has a better performancethan the United States in terms ofproducing science and engineeringgraduates, it lags well behind the UnitedStates in terms of the proportion ofscience and technology researchers in thelabour market. Nevertheless, as in othercountries, there are claims of a ‘shortage’in the United States, which economistshave struggled to reconcile with the facts(which belie this concern).

Further analysis suggests that the underlying issue is that the UnitedStates maintains an adequate supply ofscientists and engineers only because ofthe sizeable influx of foreign-bornstudents and employees. This could be arisk to US research if there is anyinterruption of the flow of immigrantscientists and engineers.

The ‘brain drain’ to the United Statesis also a concern for other countries. Forexample, analysis of migration flows in

Just because there areover-qualified individualsin the workforce does not mean that there isover-supply of graduates


9

and out of Europe suggests that Europehas lost out in terms of its own potentialsupply of ‘domestic’ graduates and itsability to attract scientists and engineersfrom other countries. The shortage ofpersonnel in these areas is likely to havecosts in terms of innovation andconsequent productivity growth.

Conclusions and policyimplicationsWhile concerns about over-education arelargely misplaced, there do appear to beproblems with graduates not alwayshaving the skills required by employers.One response to this is to make sure thatvocational courses meet the requirementsof employers and to ensure that theaccreditation system is appropriate.

But it would be unwise to emphasiseacquisition of highly specific skills at theexpense of general education. This is achallenge for whole educational structuresnot just higher education since, in manycountries, students have to make adecision between general and vocationaleducation long before they reach thestage of entering higher education.

There is also a question of the balancebetween employer-provided training andeducation provided by institutions ofhigher education. Employers have a rolein addressing concerns about skillmismatch. And governments have animportant role in improving informationabout training opportunities, settingappropriate legal frameworks andensuring portability of skills.

Potential policy responses to thevariation in returns to higher education bysubject include differential fees (orbursaries) by degree subject so thatgraduates are encouraged to studysubjects for which there is high relativedemand in the labour market. There mayalso be a case for the provision of betterinformation to potential students on jobprospects and earnings by degree subject.

More generally, given the positiverelationship between education andeconomic growth, and the fact thatreturns to higher education are stronglypositive, there is a good argument forcontinuing to expand higher education.

This could be achieved by publicprovision of more places in highereducation. Where capacity constraintsare not the issue, then an importantmatter for investigation is why more

young people do not pursue highereducation. One possibility is the costboth in terms of fees and theopportunity cost (the earnings students –and possibly their families – must forgowhile in higher education). Where suchconstraints exist (most likely for studentsfrom poor social backgrounds), there is agood case for bursaries.

Another possibility is that there isinsufficient information available topotential students about the returns thatmight be gained from pursuing highereducation (or returns in certain subjectareas). In this case again, the appropriatepolicy response would be to provide thisinformation at appropriate stages of anindividual’s education.

The article summarises ‘Tertiary Education

Systems and Labour Markets’ by Stephen

Machin and Sandra McNally, a report

prepared for the OECD.

The full report is available here:

http://www.oecd.org/dataoecd/55/31/

38006954.pdf

Stephen Machin is CEP’s research director

and director of the Centre for the Economics

of Education (CEE). Sandra McNally is

director of CEP’s research programme on

education and skills.

In many countries, thereis relatively higher

demand for graduates in science, engineering

and technology

Skill mismatch – orinadequate levels

of skill – is more ofa problem than

over-qualification

With a few honourable exceptions, the universities ofEurope are failing to provide the intellectual and creativeenergy that is required to improve the continent'srelatively poor economic performance. Too few of themare world-class centres of research and teachingexcellence. Many are desperately short of resources. Someare shamefully poor in every sense and can barely providewhat most objective observers would understand to be aneducation of quality.

The picture is not uniformly bleak. The UK and some ofthe Nordic countries have increased funding in recentyears. Countries such as Austria, Denmark and theNetherlands have greatly improved the way theiruniversities are run. The UK has some of the best researchuniversities in the world, thanks in good measure to therelative autonomy of its institutions and to the way thatresearch funding is allocated – on the basis of peer-reviewed excellence as opposed to the whims of centralgovernment or the need to spread limited resourcesevenly across every region.

But European institutions are not well placed to competein what has become a global competition for talent. Incountries such as France, Germany and Italy, the sector isstruggling to cope with too many students, and deliveringuninspiring teaching in dilapidated buildings. AcrossEurope as a whole, higher education is crying out forreform in six important areas.

The first concerns control and independence. Universitiesneed the autonomy necessary to manage their own affairsin an efficient fashion. Universities that are effectivelyagencies of the state – as is in effect the case in Franceand Italy – have very little control over their resources andare unable to set relevant academic priorities.

Throughout Europe, including the UK, there is a powerful case for universities to be more independent –even if government remains the main purchaser ofservices – including both teaching and research. This isnot about ‘privatisation’ in the sense of institutions being sold to the highest bidder. It is rather aboutuniversities seeking and earning the freedom necessary totransform what one observer has called the last majornationalised industry.

Second, higher education needs to be properly funded.The European Union (EU) countries currently invest about1.2% of their GDP in this area. A figure nearer to 2%would be required to make the EU an effectivecompetitor with the best in the world.

The important difference between Europe and just aboutevery other developed economy is that private financeplays a very modest role in its university funding. Thus,public funding for higher education represents about 1%of GDP for the EU countries, roughly the sameproportion as in the United States. But private fundingfor US universities amounts to a further 1.4% of GDPand the average in OECD countries is 0.8%, comparedwith only 0.1% for Europe.

If the quality is to be maintained and improved, andEurope’s students are to earn degrees worth theparchment on which they are printed, Europeangovernments will sooner or later have to introducetuition fees. These should be backed by strong systemsof maintenance grants to ensure that access is open toall and that students can afford to sustain three or fouryears of study.

The UK has started the process and Germany is movingin the same direction. The political challenge to thestatus quo in France will be enormous. As many Britishand American universities can testify, increasing numbersof the best young French people are already voting withtheir feet, seeking an international education beyond thestultifying constraints of the ancient regime at home.

Third, European countries are going to have to becomemuch more selective in the way they allocate resources.There are nearly 2,000 universities in the EU, most ofwhich aspire to conduct research and offer postgraduatedegrees. By contrast, fewer than 250 US universitiesaward postgraduate degrees and fewer than 100 arerecognised as research-intensive.

Given this concentration of resources, it is no wonderthe United States dominates the league tables of the world's best research universities. Europe needs todevote its available research budget – at national and EU level – on the basis of peer-reviewed


10

Knowledge is an increasingly critical factor in shaping economiclife – but across Europe, the institutions that should be the mainsources of knowledge are failing to meet the challenge. Guestcontributor Nick Butler outlines what must be done to improvethe quality of higher education in the European Union.

Europe’s universities – time for reform

in brief...

excellence. Research funding should not be a cover forregional policy.

Selectivity is also important when it comes to acceptingstudents. World-class universities have to be free to picktheir own talent rather than to take what comes – ashappens now in large parts of Europe. Without some testof merit and potential, universities will remain simply adevice for disguising unemployment numbers, as is thecase in parts of southern Europe.

The fourth area concerns the curriculum reform. This isalready under way in more than 40 countries across thecontinent, through what is known as the Bologna process.The idea is to establish easily recognisable and comparabledegrees based around a two-cycle system of studies,starting with a bachelor degree and moving on to amasters. The UK, with its own traditions, has barelyembraced the process, but change is coming and moreBritish universities need to engage and to experiment inthis area.

Fifth, Europe needs to develop a much more diversesystem of higher education. Rather than attempting tomake them all equal, the aim should be to create a richmix of institutions – some offering world-class teachingand research, others concentrating on regional or localneeds. Germany recognises this challenge with its plans tofund a small group of elite institutions, as do the best ofthe new generation of universities in the UK, which aredeveloping their own specialisms and breaking free of theneed to mimic Oxbridge.

Finally, in pursuing this agenda of excellence throughdiversity, a creative mix of funding is necessary – to rewardtalent, research and teaching; to stimulate entrepreneurialdevelopment; to encourage experimentation; and to

enable universities to reach out to meet the rapidlygrowing global need for education. Much of this fundingwill come from government, but universities also need todevelop their own funding streams, and to see themselvesas self-standing institutions.

Diversity will be one of the principal benefits of a breakfrom the nationalised past. As freestanding organisations,universities will be able to find their own distinctivecapabilities.

Universities may seem slow to change, but historysuggests that over time they do reflect the needs of thesocieties of which they are part. Change in highereducation must be part of a much wider reform of theentire education system – providing ladders ofopportunities; developing talents concealed by poor familybackgrounds and language difficulties; and offeringsecond and third chances for men and women to revisitthe process of education through their lives.

Universities are part of that wider process and vital to itssuccess. Their development and their capacity to respondto the challenges facing Europe will be a crucial leadingindicator of the EU’s success in the fiercely competitiveenvironment of the twenty-first century global economy.


11

Nick Butler is Director of the Cambridge

Centre for Energy Studies. He is co-author with

Richard Lambert (a member of CEP’s policy

committee) of The Future of European

Universities – Renaissance or Decay, published

by the Centre for European Reform in 2006.

Europe needs a muchmore diverse system ofhigher education with acreative mix of funding

Throughout Europe, thereis a powerful case

for universities to be more independent

Imag

e: L

inds

ay B

eyer

stei

n

The exploits of DavidBrent in the televisionseries The Office havemade bad Britishmanagement practicesinfamous around the

globe. But these failings have a far longerhistorical pedigree. The Harvard businesshistorians Alfred Chandler and DavidLandes have both claimed that poormanagement practices held back Britishcompanies. In 1947 as part of theMarshall Aid scheme to revive post-warEurope, American businessmen andengineers concluded that ‘efficientmanagement was the most significantfactor in the American advantage’.

But how do British firms now comparein terms of management practices, notonly with the United States andcontinental Europe but also with the risingindustrial giants of India and China? Untilrecently economists have had little to sayabout the role of management in drivingproductivity and other key performanceindicators. This is largely because there hasbeen an absence of good quality data onmanagement practices. Working inpartnership with McKinsey & Company,CEP has been carrying out a large researchproject that attempts to fill this void.

We have developed an original surveymethod to measure management practicesin a systematic way in more than 4,000

firms in Europe, the United States andAsia. By combining these data with firmaccounts and industrial statistics, we areable to explore in detail the relationshipbetween management practices, theeconomic environment and companyperformance.

Overall, we find compelling evidencethat better management practices aresignificantly associated with higherproductivity and other indicators ofcorporate performance, including returnon capital employed, sales per employee,sales growth and survival. This is true inevery country we look at, suggesting thatour characterisation of good managementpractice is not culturally biased towards‘Anglo-Saxon’ approaches.

We estimate that managementpractices can account for up to a third ofthe differences in productivity betweenfirms and countries. Why are there suchstartling differences in the managementpractices and productivity of competingcompanies? Our research offers somepotential explanations for thesedifferences and suggests areas wherepolicy can encourage the spread of goodmanagement practices.

Measuring managementpracticesMeasuring management requires us tocodify the concept of good and bad


12

What drives good maround the world?It has long been suspected that badmanagement plays a key role in explaining theUK’s productivity gap with the United Statesand some of our European neighbours. CEP’sglobal survey of over 4,000 firms suggests thatthere is indeed a ‘management gap’ – andreveals the forces driving variations in thequality of management practices.

So you thinkmanufacturing is boring...During the summer of 2006, weinterviewed over 4,000 managers. Some ofthese individuals were extremely colourfulcharacters, providing endless entertainmentto the research team with their commentsimmortalised on our team quotes board.Some of our favourites included:

Talent rewards the Indian wayInterviewer:

How do you identify your

star performers?

Indian plant manager:

This is India. Everyone thinks

he is a star performer.

management into a measure applicable todifferent firms. We used an interview-based management practice evaluationtool that defines and scores from 1 (worstpractice) to 5 (best practice) across 18 ofthe key management practices that appearto matter to industrial firms, based onMcKinsey’s expertise in working withthousands of companies across severaldecades. The 18 practices fall into fourbroad areas:

■ Shopfloor operations: have companiesadopted both the letter and the spirit oflean manufacturing?

■ Performance monitoring: how well docompanies track what goes on insidetheir firms?

■ Target setting: do companies set theright targets, track the right outcomesand take appropriate action if the twodon’t tally?

■ Incentive setting: are companies hiring,developing and keeping the rightpeople and providing them withincentives to succeed?

For each company in the study,researchers interviewed by telephone oneor two senior plant-level managers, whoknew only that they were taking part in a‘research’ project. These managers wereselected because they are senior enoughto have a reasonable perspective on whathappens in a company but not so seniorthat they might be out of touch with theshopfloor. The interviews relied on openquestions and the interviewers weretrained to probe for details of practices onthe ground.

The interviews were run by aninternational team of 47 postgraduatestudents (mainly MBAs), who worked fromCEP in a specially created survey centre

management CentrePiece Autumn 2007

13

Figure 1:

US firms are the best managed, followed by the Germans andSwedes, with the Greeks, Indians and Chinese the worst

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

China

India

Greece

Portugal

Poland

France

UK

Italy

Japan

Sweden

Germany

United States

The bars indicate for each country the average score on the 18 management questions (1=worst practice, 5=best practice).

The British chat-upMale production manager:

Your accent is really cute and

I love the way you talk.

Do you fancy meeting up near

the factory for some fun?

Female (Australian) interviewer:

That’s a great offer – how

could I refuse? Unfortunately,

I’m washing my hair every night

for the next three months.

Strong competitionand flexible labourmarkets both lead

directly to improvedmanagementperformance

during the summer of 2006. This was a24-hour operation since the Chinese daystarts at midnight in London, just beforemanagers on the West Coast of the UnitedStates pack up to go home.

Management practicesaround the worldAs Figure 1 shows, there are significantdifferences in management performanceacross countries. The United States is at thetop of the management league table,while Greece, India and China are theworst performers. Germany, Sweden andJapan are (not surprisingly) strongperformers given the manufacturing focusof the survey, while France, Italy and theUK are all solidly mid-table.

But the United States is not entirelydominant. US firms score particularly highlyfor people management, such aspromoting and rewarding talented workersquickly. But as Figure 2 shows, in shopflooroperations management, Sweden, France,Italy, Japan and Germany do relativelybetter.

Overall, cross-country differencesaccount for only 9% of the variation inmanagement practice. Performancedifferences between companies in thesame country are far larger than any cross-country variation. For example, the bestthird of Indian companies outperform theEuropean average. This is worrying for

those who complacently assume that vastlysuperior Western management protectsthem from offshoring.

Managers are very poor atself-assessmentSince good management is strongly linkedwith good performance, why do so manyfirms fail to make a priority of improvingtheir practices? The techniques are prettywell known yet many firms remain poorlymanaged.

To examine possible causes of thisdisconnect, we asked managers as a finalquestion in the interview to assess theoverall management performance of theirfirm on a scale of 1 to10. To avoid falsemodesty, they were asked to exclude theirpersonal performance from the calculation.

As Figure 3 indicates, interviewees’answers to this question are not well


14

Figure 2:

European firms are relatively better at operationsmanagement than people management

-0.350 -0.175 0.000 0.175 0.350

India

Poland

China

United States

UK

Greece

Portugal

Germany

Japan

Italy

France

Sweden

The bars indicate for each country the average score on six questionsfocused on operations management minus the average score on sixquestions focused on people management.

Americans ongeographyInterviewer:

How many production sites do

you have abroad?

Manager in Indiana:

Ummmm… well… we have

one in Texas…

Multinationals tend to achieveexcellent management practices

wherever they are located

correlated with either our managementpractice score or their own businessperformance. At the country level, we findGreek, Portuguese and Indian managers tobe the most over-optimistic about theirmanagement practices, while theJapanese, Swedish and French managersare the most pessimistic.

Government policy plays animportant roleA variety of policy factors have an effect oncompanies’ adoption of good managementpractices. Most significant among these aretheir competitive environment and theflexibility of the local labour market.

When competition (whether measuredby narrow industry profit margins, tradeopenness or number of rivals) is higher,management is better. This could be aresult of two effects: first, good practicespreads quickly in highly competitiveenvironments; and second, poor practice iseliminated by Darwinian natural selection aspoorer performing companies are removedfrom the marketplace.

We also find that flexible labourmarkets matter, since these appear to allowcompanies to adopt better peoplemanagement practices. In countries withrigid employment laws (using the WorldBank’s index), firms find it difficult toimplement effective hiring, promotion,retention and firing practices.

The high position of the United States in the management league table ishelped by its competitive product marketsand flexible labour markets.


15

Figure 3:

Managers are over-optimistic about their own managementpractices across the globe

■ Self-assessed management score

■ Our management score

The difficulties of defining ownership in Europe Production manager:

We’re owned by the Mafia.

Interviewer:

I think that’s the ‘Other’ category… although I guess I

could put you down as an ‘Italian multinational’.

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

China

India

Greece

Portugal

Poland

France

UK

Italy

Japan

Sweden

Germany

United States

The bars indicate for each country the average score on the 18 managementpractice questions and the average score on the self-assessed managementquestion: ‘Excluding yourself, how well managed is your firm on a scale of 1 to 10,where 1 is worst practice, 10 is best practice and 5 is average’. The scores aredivided by 2 to put them on the same scale as our management scores.

management scores are more likely to bebiased downwards by having recentlypurchased badly run firms.

To investigate this we re-plotted themanagement scores for only those firmsthat have had the same ownership for atleast the last three years and found thatdoing this increases the lead of privateequity firms over all other firms, makingthem the best managed in the sample.

Multinationals, familyownership and skillsFirm ownership and the availability ofskilled people, both in management andamong the workforce in general, are alsoassociated with important differencesbetween the better-managed firms andthe rest.

For example, multinational companiesare well managed around the globe,achieving extremely good managementpractices in countries like Greece and Indiadespite the poor management practices oflocal domestic firms.

Family ownership and the traditionalpractice of primogeniture – handingdown the CEO position to the eldest son– are associated with particularly badmanagement practices (see Figure 4). This appears to be an issue for Europe


16

Private equity, public gainOne particular ownership form thatappears to be linked with superiormanagement practices is private equity. AsFigure 4 shows, private equity firms arethe best managed. This superiorperformance of private equity appears tobe quite robust – they come out on topboth with and without controls forcountry and industry.

One possible explanation is thatprivate equity firms only buy well-managed firms so that their highmanagement scores simply reflect theirability to cherry-pick the firms that theybuy. But the usual story of private equitybuy-outs is the reverse: they buy badlymanaged firms with the aim of turningthem around. This suggests that their

Figure 4:

Management scores are highest for private equity owned firms

■ Management score for all firms

■ Management score for firms that did not change ownership in the past three years

-0.25 -0.20 -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 0.20 0.25

Private equity

Dispersed shareholders

Family/founder owned butwith an external CEO

Private individuals

Family/founderowned and CEO

Government

The bars indicate for each type of firm ownership the deviation fromthe country and industry average score on the 18 managementpractice questions. The scores are for domestic firms only, of whichthere are 2,385 in the sample. The red bar for each type of ownershiponly includes the 2,180 firms that have had the same ownership forthe past three years. The number of firms for the blue/red bar are asfollows: government (53/53); family/founder owned and CEO(1,185/1,082); private individuals (364/285); family/founder ownedand external CEO (153/145); dispersed shareholders (624/566); andprivate equity (64/37).

Employee retention the old-fashioned way:Company chairman:

Sex is a great thing! If I can get

my employee a local girlfriend

he’ll never leave.

The best-managed firmsare those withthree or more

years ofprivate equity

ownership

Family ownedfirms that appointthe eldest son asthe CEO areparticularly badlymanaged


17

since in France, Greece, Italy, Portugaland the UK, around 10% of themanufacturing firms are family ownedwith a CEO that has been chosenbecause they are the eldest son. TheUnited States performs much better onthis dimension, with only 2% of its firmsbeing family owned with the CEO chosenbecause he is the eldest son.

The skills of both the managers andnon-managers in the firm also appear toplay an important role. For example,84% of managers in the highest scoringfirms are educated to degree level orhigher, as are a quarter of the non-management work force. Among thelowest scoring firms, by contrast, only54% of managers and 5% of the widerworkforce have degrees.

What can the government do?Our research shows a significantmanagement gap between the UK on theone hand and the United States and someEuropean countries on the other. This is asituation that the government can modifyby encouraging the uptake of goodmanagement practices.

Our research suggests that strongcompetition and flexible labour marketsboth lead directly to improvedmanagement performance. Multinationalcompanies have a strong positive effecttoo, and their influence is felt throughoutthe countries in which they operate. Inthese respects, the British government hasa good track record and it is in otherEuropean countries that these lessonsneed to be taken on board.

The UK performs less well in the areasof skills and family ownership. Britishlevels of basic education are low byinternational standards, and any policiesthat addressed this would have a bigimpact. As regards family ownership, thereis currently a distortion in the inheritancetax system that actually promotes thecontinued ownership of privately heldmanufacturing firms in family hands,keeping these out of private equityownership.

Our research suggests that byappointing managers on the basis of primogeniture rather thancompetitively on the basis of merit, we are possibly promoting more badmanagement and productivity practices in the UK.

More details on this research can be found in

’Management Practice and Productivity: Why

They Matter’ by Nick Bloom, Stephen Dorgan,

John Dowdy, Christos Genakos, Raffaella

Sadun and John Van Reenen, July 2007

(http://cep.lse.ac.uk/management/

Management_Practice_and_Productivity.pdf).

For full details of the survey methodology,

including all the questions, see ‘Measuring

and Explaining Management Practices across

Firms and Nations’ by Nick Bloom and

John Van Reenen, CEP Discussion Paper

No. 716 (http://cep.lse.ac.uk/pubs/download/

dp0716.pdf) and forthcoming in the

Quarterly Journal of Economics.

The research was jointly funded by the

Advanced Institute of Management Research,

the Anglo-German Foundation, the Economic

and Social Research Council and the

Kauffman Foundation.

Nick Bloom is an assistant professor of

economics at Stanford University and a

research associate in CEP’s productivity and

innovation programme. Christos Genakos, a

research associate in CEP’s productivity and

innovation programme, is at Cambridge

University. Raffaella Sadun is a CEP

research economist. John Van Reenen is

director of CEP.

The things we did to get interviewsFrench secretary:

You want to talk to the plant

manager? There are legal

proceedings against him,

so hurry up.


18

Ozonomicsin brief...

The most popular misconception in economics and politicsis that if the economy is humming along, the governmentmust be doing a good job – it must be a capableeconomic manager and its policies must be working.

When the economy is booming, politicians encourage thepublic to believe that they tightly control the economy.The idea that they hold the fortune of the nation in thepalm of their hands appeals to them, and they want thepublic to take the strong economy as evidence of theirskill and omnipotence.

The media too are susceptible to this fiction because we like to have someone to blame for our hardships and praise for our success – heroes and villains make good stories.

The truth, however, is that politicians have much lesscontrol over the economy than they would have usbelieve. Certainly, there can be policy successes and policy failures, but more often than not the condition ofthe economy is determined by factors outside the controlof politicians.

The economy is much like a little boat in a wide sea.Whether the trip is calm or rocky depends much more onthe weather and the wash from the bigger boats thananything that might be done internally. If the weather isbad, you cannot blame the skipper for the bumpy ride. Infact, unless you know a lot about sailing, it’s hard to knowwhether the skipper is doing a good job in badcircumstances or whether he’s making it worse.

The same is true of skippering the economy. If you askpeople about the state of their economy in the last fiveyears of the global boom, they will remember the low

unemployment rates, the low inflation, the low interestrates and the general mood of affluence that suffused theworld. Incumbent leaders have generally turned thisprosperity into stunning electoral success.

But most people are much sketchier on the followingquestion: ‘Which of our leader’s policies, if any, canactually be shown to have produced the boom for whichhe has received so much credit?’ This question is critical tothe issue of whether a government’s leadership isresponsible for the prosperity it presided over – orwhether that prosperity was produced by the policies ofprevious governments or the influence of benigninternational economic forces.

In Australia, the incumbent conservative government hasbenefited from an extraordinary period of prosperity, onethat makes even the recent success of the British economylook modest by comparison. Australia has had 16 years ofcontinuous expansion during which its wealth hasdoubled, labour productivity increased by a half and jobsincreased by a quarter so that unemployment is at a 33-year low of 4.3%.

Australians have generally chosen not to think too hardabout where this prosperity has come from. They haverichly rewarded Prime Minister John Howard at the ballotbox without looking too closely at causes and effects.

But a more accurate analysis of different governments’contribution to Australia’s ‘miracle’ economy involvesrecognising that economic policies act with a lag, so the prosperity reaped today may have been sown manyyears before.

In Australia the battle for the title of ‘better economic

Australia’s extraordinary period of prosperity has allowedthe incumbent government to position themselves aseconomic superheroes. But as Andrew Charlton warns ina new book, we should always adopt a sceptical attitude toany politician’s claims about their contribution to nationaleconomic success.

Letting politicians take credit for abooming economy distorts the publicdebate about economic policy

manager’ has become one of competing narratives.Howard and his finance minister Peter Costello, backlit bythe conspicuous boom, have an obvious advantage overthe Labor Party’s story, which is unflatteringly silhouettedagainst the recession they presided over.

Labor says its macroeconomic framework in the 1980sand microeconomic reform in the 1990s created thebedrock for later success and that the recession was theprice to be paid for progress.

In contrast, the conservatives say the Howardgovernment’s decisions to pay down the national debtand deliver budget surpluses have directly produced thelong boom. Howard has taken credit for the prosperityand used it to position himself as an economic superherodeserving of complete trust from the electorate.

Every government’s claims about the economy deservethorough scrutiny because what we believe about the

foundation of our wealth makes a great deal ofdifference to how we should pursue our future. If, as

our leaders would have us believe, the prosperity ofBritain and Australia has been due to the superior

performance of our leaders, then our job ascitizens is merely to continue passively to re-elect

those leaders.

But if our success is a complex combinationof long-term policies and globalcircumstances, then our challenge is toensure we continue constant public debate

with a view to finding the right policies tosustain our position in a changing world.

Letting politicians take credit for a boomingeconomy that is not of their making distorts the

public debate about economic policy.


19

Andrew Charlton is a research economist

in CEP’s globalisation programme. His new

book Ozonomics: Inside the Myth of

Australia’s Economic Superheroes has just

been published by Random House.

Economic success is acomplex combination of

long-term policies andglobal circumstances


20

Union bluesin brief...

Trade unions are in the doldrums. Although the rate ofdecline in union membership has slowed since Labourcame to power in 1997, the latest figures show that theycontinued to lose members in 2006. The reasons are clear.Unions are less able to organise new workplaces and newworkers than they used to be. As a consequence, anincreasing proportion of all workers have never beenunion members, and new workplaces rarely recogniseunions for pay bargaining.

Less well known is the effect that the loss of membershipis having on unions as organisations. These effects areidentified for the first time in our research, whichmeasures the resources available to unions both on theirown balance sheets and within establishments.

As Figure 1 suggests, the finances of British unions are ina parlous state. This is not particularly surprising since theyrely very heavily on members’ subscriptions as theirprimary source of income. Thus, their income flows fall asmembership falls unless they can increase membershipfees substantially or generate income from other sources.Moreover, expenditure has exceeded subscription incomefor some time. Of course, this is not a sustainable strategyin the long run and it has implications for their ability tosustain assets.

Few unions like to raise membership fees. It’s generallyviewed as impractical and unacceptable. Most unionsallow their annual conferences to debate a higher fee but,not surprisingly, there is rarely an appetite for it so itdoesn’t happen. Yet unions elsewhere in the world haveadopted this approach. For example, the Health ServicesUnion of Australia recently decided to increase theirsubscriptions dramatically, arguing that the quality of theirservices, including collective bargaining, requires them tobe on a stable and viable financial footing.

In countries like the United States, members are used topaying a much higher percentage of their wages in unionfees: in return, they get the largest union wage premiumin the world. Some unions in Britain are studying this issuecarefully and are considering running experiments to testthe sensitivity of demand for membership to the price ofjoining – whether and by how much numbers might fall iffees were higher.

Meanwhile, the chief response of unions to thediminishing pool of unionised labour has been to engagein what we call ‘market share unionism’, in which theyhave sought to grab a greater share of the remainingunion pie. It is this strategy that lies behind the wave ofunion mergers we’ve witnessed in Britain in recent years.

The membership of Britain’s trade unions continues todecline and, despite a series of mergers, most of them facesevere financial difficulties. Alex Bryson and PaulWillman examine their organisational failings and find aglimmer of hope in the handful of success stories – unionsrepresenting professional workers.

Figure 1:

Changes in union resources 1984-2004

75

80

85

90

95

100

105

110

115

2004199619901984

■ Solvency index■ Reserves index■ Membership index■ Offbal index

Each of the four indices is set at100 in 1990. The Offbal (‘off-balance sheet resources’) indexis constructed using theWorkplace EmploymentRelations Surveys, from which aworkplace can be rated between0 and 3, scoring 1 each time ithas one of the following: check-off; management recommenda-tion of union membership or aclosed shop; and an on-siteunion representative. The otherthree indices are derived fromCertification Officer returns.Solvency is the margin of totalincome over total expenditure.Reserves are total funds dividedby expenditure.

The most successful unionsorganise occupations rather than

industries or workplaces

Between 1990 and 2005, the number of unions fell byover 40%. In 2007, there have been further reductionsand greater membership concentration following themerger between AMICUS and the Transport and GeneralWorkers Union.

The rationale for such mergers is analogous to therationale for corporate mergers and acquisitions: unions tryto consolidate their resources, creating a more substantialorganisation that, at least in theory, is capable of grabbinga bigger share of existing members and, if they are lucky,attracting new members through greater ‘reach’.

At the same time, the logic goes, unions can reduce theircost base by stripping out duplicate union services(offices, officials and the like). In practice, it hasn’t reallyworked out like this. By definition, the big unions eachhave a greater share of union membership than they didwhen they operated separately. But there is no evidencethat they have succeeded in expanding membershipbeyond their traditional base.

What’s more, they have not been cutting costs so as totake advantage of the economies of scale that mergersoffer. This is partly because, despite mergers, theseunions rarely operate as general unions. Instead, theycontinue to operate along sectoral, industrial oroccupational lines, often because unions recognise that different types of members require different types of service.

This organisational knowledge is often locked away inparticular sections of unions used to servicing parts of theunions’ membership. If unions are to reap returns toscale, they’ll have to work out how to share thisknowledge around the organisation and create efficientstructures that permit them to service their members withfewer staff and offices.

What does all of this mean for unions in the future? First, it is doubtful whether unions will be able to servicetheir existing members without greater reliance on thevoluntary endeavours of lay union activists and theirbroader membership. This is what we call their ‘off-balance sheet resources’. Alas, our research shows thatthese resources are also in decline.

Second, unions’ organising capability is severely damaged.When finances are tight, unions are less likely to riskspending money organising in new workplaces unless they can be fairly sure of success. More generally, theysimply do not have the organisational capacity on theground to reach out to new workers and bring them intothe union movement.

But it would be wrong to conclude that all is doom andgloom. Although the general picture looks bleak, there arehuge differences in the finances of different unions andthere are some real success stories.

What’s more, success depends, at least in part, on thebusiness models that unions are deploying. The mostsuccessful unions, both in membership terms andfinancially, are those representing professional workers, inparticular, those representing doctors (the British MedicalAssociation), nurses (the Royal College of Nurses) andteachers (the National Union of Teachers).

These unions continue to organise occupations, ratherthan industries or workplaces, because union membershipremains strongly linked to occupational identity. As well asrepresenting their members in pay and conditionsnegotiations, grievance procedures and the like, theseunions also protect their members against clients andstate interference. Other unions would do well to takenote – before it’s too late.


21

This article summarises ‘Accounting for Collective Action:

Resource Acquisition and Mobilisation in British Unions’

by Alex Bryson and Paul Willman, CEP Discussion Paper

No. 768 (http://cep.lse.ac.uk/pubs/download/dp0768.pdf)

and forthcoming in Advances in Industrial and

Labor Relations.

Alex Bryson is research director at the Policy Studies

Institute and the Manpower Research Fellow at CEP.

Paul Willman is professor in employment relations and

organisational behaviour at LSE.

The latest figures on union membership are

in a 2006 DTI/ONS report by Heidi Grainger and Martin

Crowther (http://www.berr.gov.uk/files/file39006.pdf).

Unions no longer have theorganisational resources to

reach out to new workersand new workplaces


22

Life on theMonetary PolicyCommittee

Stephen Nickell was the first externalmember of the Bank of England’s MonetaryPolicy Committee to serve two three-yearterms. He reflects on his experiences settingthe nation’s official interest rate.

One Tuesday eveningin the spring of2000, I received aphone call from GusO’Donnell, the thenpermanent secretary

at HM Treasury, suggesting that I might liketo attend a small meeting at the Treasuryon the following day. After somediscussion, he revealed that it concerned ajob but refused to be more specific.

On the Wednesday morning, I wentinto the Treasury where he and Ed Ballsasked if I wished to become a member ofthe Bank of England’s Monetary PolicyCommittee (MPC). On the Wednesday andThursday, I organised leave from the LSEfor four days per week and had a chatwith Eddie George, then the Bank’sgovernor. Gordon Brown announced myappointment on the Friday morning.


23

Such speedy appointments to publicbodies are relatively unusual but membersof the MPC have always been, and stillare, appointed with great rapidity andsecrecy, much to the irritation of theHouse of Commons Treasury Committee.

Before starting on the MPC andhaving purchased a new suit, the firstexciting event was to appear before theTreasury Committee for my confirmationhearing. This involved first, completing alengthy ‘exam’ paper on various aspects ofmonetary policy-making (which was set, asit happens, by Charlie Bean, then anadviser to the committee and a fellow LSEprofessor, and from later that year theBank’s chief economist); and second, agrilling by the committee in the presenceof the financial press.

When I made my appearance, I wasthe second person on, the first havingbeen Chris Allsopp who had beenappointed at the same time. Chris wasgiven such a hostile reception by thecommittee (with Brian Sedgemore leadingthe charge) that by the time they got tome, the MPs were merely mildlyaggressive, having run out of steam. This proved to be the last time thatprospective MPC members went beforethe Treasury Committee for theirconfirmation hearings without a longpractice session at the Bank.

The MPC operates on a monthly cycle,essentially because, by law, it must meetto set interest rates in every calendarmonth. The interest rate decision is madeon the first or second Thursday of eachmonth. On the previous Friday morning,the committee meets to listen to the Bankstaff going through the economic news.The following week, the committeespends the Wednesday afternoon

discussing the current and futureprospective state of the economy and theThursday morning making the interest ratedecision, which is announced at noon.

These discussions form the basis of theminutes that are put together by asecretariat of four Bank staff who arepresent throughout. These minutes arediscussed at length and amended by the entire committee on the Monday,eleven days after decision day and arepublished, along with details of the vote,two days later.

One of the key features of this processis the fact that the decision on interestrates is taken by strict majority vote. Onthe Thursday morning, the governorinvites each member of the committee topresent their vote on rates along withtheir reasons, each member beingexpected to talk for about ten minutes.The deputy governor in charge ofmonetary analysis (currently Rachel Lomax)is always invited to go first, the next sevenmembers are then asked in apparentlyrandom order and the governor alwaysgoes last.

Of the nine members, five areinternals who are permanent Bankemployees: the governor, two deputies,the chief economist and the head ofmarkets. The other four are externalmembers appointed in the same fashionas myself. On one occasion, I was presentat an exceptionally unusual event. Therandom order of speaking by the sevenmembers between Rachel Lomax and thegovernor just happened to be in anti-clockwise order around the table. If theorder were truly random at each meeting,this event would only be observed, onaverage, once every 438 years.

Extensive study of the voting records

of the committee has failed to elucidateany patterns. There is no block voting, andwithin both internal and external groupsthere are often divisions with deputygovernors voting against the governor, forexample. Indeed the committee is notconcerned with consensus, split votes arecommon and the governor has twice beenon the losing side. Note that sincegovernors always vote last, they canchoose whether to join the winning orlosing side except in the unusualcircumstance when they have the castingvote, with the previous eight members’votes tied at 4-4.

The style of decision-making, usingstrict majority voting based on theindividual views of nine independent MPCmembers, is reinforced by the fact thatmembers are held to account for theirindividual votes. Formally, they arequestioned in public by the TreasuryCommittee, where they must appear threeor four times a year. Furthermore,individual members frequently usespeeches, papers and interviews to justifytheir particular positions.

In my view, this method of decision-making – that is, nine individuals comingto their own decisions and then usingmajority voting to aggregate them –generates outcomes that are superior tomethods based on a search for aconsensus under the auspices of adominant leader favoured by some centralbanks.

The desirability of having independentvoices on the committee suggests thatwhile professional economists should havea strong representation, it is a good ideato have a number of members with adifferent point of view. This helps toensure that decision-making is not


24

dominated by a rigid consensusperspective. Indeed to guard furtheragainst this, during my time thecommittee often used to invite membersto prepare short papers focusing onfactors that might lead to decisions beingseriously mistaken.

Of course, aside from the actualprocedures involved with the monthlyround of decision-making, there is a greatdeal of background activity. Within theBank, all the external members engage inresearch and analysis, each assisted by twoBank researchers. This results in numerousspeeches, press interviews and papersprepared for consumption by academics,financial journalists and the general public.

An important part of the job is toensure that individual views andcommittee decisions are explained to theworld at large as clearly as possible. Tohelp with this, the MPC publishes aforecast every three months explaining itsview of where the UK economy is goingover the next three years. The productionof these forecasts involves numerousmeetings of the committee, interactingextensively with large numbers of theBank staff. The monetary analysis divisionof the Bank contains around 120economists, which reveals the sheer scaleof the whole monetary policy operation.

On top of this, it is part of the remit ofMPC members to go on regional visits. Soabout ten times a year, I would set off forsome distant part of the UK both to listenand to explain. These visits usually lastedtwo days and involved a great deal ofeating. The basic format was to meetlarge groups of business people, trade unionists, academics and so on,over, successively, lunch, dinner, breakfast, lunch.

The idea was to give a brief talk aboutthe economic situation and then engagein discussions, sometimes heated, abouthow things looked from the individualperspectives of the people present. Thesewere organised by the Bank’s regionalagents, who had numerous business andother contacts, and often involved thelocal CBI (Confederation of BritishIndustry), Chamber of Commerce andsimilar organisations.

The two most popular topics underdiscussion concerned either the direconsequences of official regulation forbusiness or what was going to happenwhen manufacturing industry disappeared

entirely. Monetary policy rarely seemed tobe a cause for concern.

Between meals, I would visit localworkplaces, talk to the local press andgive interviews on local radio. The purposeof all this was partly to get a feel for thestate of the economy on the ground andpartly to fly the flag and explain what thecommittee was up to.

By and large, these visits were greatfun. I got to visit places as far afield as theIsle of Lewis and the Isles of Scilly, as wellas less exotic places like Aberdeen,Enniskillen, Pwllheli, Truro and Wakefield,plus all the major cities of the UK. I got tosee the making of steel, aluminium, glass,brake linings, sandwiches, Smarties,Formula One cars and stair-lifts, wentdown the deepest mine in the UK (apotash mine in Cleveland), wanderedaround call centres, docks, farms, batteryhen sheds, garden centres and shoppingcentres, and only failed to get to an oil rigbecause the health and safety procedureswould have taken too long.

Overall, the hospitality was splendid,especially in Northern Ireland. The OmaghChamber of Commerce annual dinner wasparticularly memorable: arriving at 7pmfor pre-dinner drinks, sitting down todinner at 9.30pm, standing up to speakon monetary policy at around midnight,and finally taking my leave of an event stillin full cry well after 3am.

During my entire time on thecommittee, any time I appeared at apublic event or conference, whatever thesubject and wherever the place, journalistsfrom the wire services were always there.Their job was to obtain a quote. Thesecharacters became a part of my life –indeed one of them followed me around,at the Bank’s invitation, for an entire two-day regional visit to North Wales.

So I became famous in a ratherlimited, ex officio sense. Then on 1 June2006, I became an ex-member of the MPCand I have never seen any of them again.It was great fun while it lasted.

The MPC decision-making method generates superior

outcomes to methods based on asearch for a consensus under the

auspices of a dominant leader

UK monetarypolicy is decided by strict majority

voting based on committee

members’individual views

Stephen Nickell is Warden of Nuffield

College, Oxford and a longstanding

participant in CEP’s research programme

on labour markets.

Regulators have capped the charges thatmobile operators can levy on other networksfor connecting calls to their subscribers. But asnew research by Christos Genakos andTommaso Valletti shows, this leads to a‘waterbed’ effect, where prices rise elsewhere.Their analysis has implications for recent EU caps on ‘roaming charges’.

The prices that mobileoperators charge othernetwork operators(fixed or mobile) forconnecting calls totheir subscribers – so-

called termination charges – have becomea hotly debated issue among regulatorsand academics worldwide. The level ofthese charges is perceived to be high both in absolute terms, but also in relation to similar prices charged by fixednetwork operators.

Industry analysts argue that suchcharges may inhibit the future growth oftelecoms services. What’s more, especiallyfor fixed-to-mobile termination rates, alarge body of theoretical research ineconomics has demonstrated thatindependently of the intensity ofcompetition for mobile customers, mobileoperators have an incentive to set chargesthat will extract the largest possiblesurplus from fixed users.

This problem has provided thejustification for regulatory intervention tocut termination rates. But reducing thelevel of charges can potentially increase thelevel of prices for mobile subscribers. Thisis what is known as the ‘waterbed’ effect,where pressing down prices in one part offirms’ operations causes another set ofprices to rise. Understanding andquantifying the effect – as our researchaims to do – is critical for assessingconsumer benefits from mobile terminationregulation.

Regulating terminationchargesMobile termination charges have been animportant issue ever since 1997 when thefirst regulatory debate started in the UK.Price controls on the two largest operatorswere put into effect from 1998 to2002, requiring termination

charges to be reduced by 9% per year inreal terms.

After a lengthy consultation andinvestigation, the UK telecoms regulator(then called Oftel) concluded at the end of2001 that mobile termination charges werestill substantially in excess of cost. Itproposed additional price controls for thenext four years on the four major mobilecompanies, Vodafone, O2, Orange and T-Mobile. The companies objected and thematter was referred to the CompetitionCommission.

The Commission broadly endorsedOftel’s proposals. It concluded thatcompetition in the mobile industry did notconstrain fixed-to-mobile terminationcharges and that a pricecap was the only remedy

Regulating the mobile phone industry:

beware the‘waterbed’ effect


25


26

likely to address these detrimentseffectively. The Commission consideredthat this would yield significant benefitswithout an increase in average retail pricesor a significant loss of retail subscribers. Infact, it was during these investigationsthat the term waterbed was first coined bythe late Paul Geroski, chairman of theCompetition Commission.

Other countries have followed theUK’s lead. The European Commissionintroduced a new regulatory frameworkfor electronic communications in 2002.Every member state of what was then theEU-15 was obliged to conduct a marketanalysis of the mobile termination marketand, to the extent that market failureswere found, remedies would have to beintroduced. Indeed, all the countries thatcompleted the analysis did find problemsand imposed (differential) cuts totermination rates.

In 2005, the New Zealand CommerceCommission introduced similar regulationand while it was convinced that thewaterbed effect is a theoretically generalphenomenon, it doubted its empiricalimportance. Similarly, the most recenttermination rate proposals by Ofcom(Oftel’s successor organisation)acknowledge the importance of thewaterbed effect, but question whether theeffect is ‘complete’, arguing that this canonly be the case if the retail market issufficiently competitive.

Analysing the waterbed effectMobile operators compete in themarketplace to win subscribers, who thenprovide a stream of revenues. Theycompete by offering attractive prices forsubscriptions and outbound mobile callsand also, in the case of monthlysubscribers, subsidised handsets. In doingso, they consider all the revenues that will

accrue from acquiring a customerand all the costs of servicing

that customer. Part of their revenues are the charges that they receive fromother networks for connecting calls totheir subscribers.

When considering its pricing policy, amobile operator will take thesetermination revenues into account. Thehigher these revenues, the lower the totalprice an operator would charge itscustomers. It makes sense for any mobileoperator to pass-through some of itsrevenue to consumers, because bylowering prices, it increases the number ofmobile subscribers, which in turn increasesthe termination revenues earned.

Of course, the reverse is also true: ifregulation reduces termination chargesand hence revenues, operators will have toraise their prices to subscribers. To assesshow widespread this waterbed effectmight be, we need to analyse how it canemerge under different market scenarios.

If the mobile market werecharacterised by very strong (perfect)competition, then operators would expectto make zero excess economic profits.Consider now what would happen as aresult of an increase in terminationcharges. This would increase the revenuesassociated with each new customer, henceincreasing their value to mobile operators.Mobile operators would start competingby lowering their prices to acquire thesenew customers.

The result of competition would be acomplete waterbed effect: any additionalprofit would be simply passed on toconsumers via lower prices, so thateconomic profits remain zero. Under thismarket structure, the introduction ofregulation to cut fixed-to-mobiletermination charges would affect thestructure of prices but not the overallprofitability of operators.

While competition appears to bestrong in the mobile industry, bothregulators and market analysts agree thatoperators possess a significant amount ofmarket power. Hence, there is also a need

to consider the nature of the waterbedeffect when competition is not perfect.

Consider the extreme case of amonopolist. Economic theory suggests thata firm with full market power maximisesprofits by setting the price at the pointwhere marginal revenue equals marginalcost. In other words, the price is such thatthe extra revenue the monopolist earnsfrom selling his last unit is equal to hisextra cost of producing this unit.

Consider again the effect of anincrease in termination charges. Assumingfor simplicity that each subscribergenerates the same amount of terminationrevenue, this would increase the revenueearned on each consumer. As a result, themonopolist would now charge a lowerprice to attract more customers, againcausing the waterbed effect.

Notice that the waterbed effect wouldnot be complete now, as the firm wouldpass-through as lower prices only part ofthe extra revenue. Hence, even in theextreme scenario of a firm possessingcomplete market power, a reduction intermination charges is expected to causelower marginal revenues that wouldincrease optimal subscription prices.

In practice, mobile markets worldwideare dominated by a small number of firms.Competition among them is expected tobe somewhere between the two extremescenarios of perfect competition andmonopoly. Under these more general(oligopolistic) market conditions, the sameeconomic logic applies.

The size of the waterbed

effect is key tounderstanding thecosts and benefits

of terminationcharges regulation

The magnitude of the waterbed effectwill depend on the intensity ofcompetition as well as the shapes of thedemand and cost functions underlying themobile industry. The previous two extrememarket structures provide us with thebounds within which we expect thewaterbed effect to lie.

Measuring the waterbedeffect in the mobile phone industryDespite the importance of the waterbedeffect and the wide range of marketconditions under which it arises, untilnow, there has been no systematicevidence to back up the theory. The mainpurpose of our research is to examine theexistence and magnitude of the waterbedeffect in the mobile phone industry. Usinga new dataset of mobile operators acrossmore than 20 countries during the lastdecade, we analyse the impact of fixed-to-mobile termination rate regulation onprices and profit margins.

Both the timing of the introduction ofregulated termination rates and theseverity with which they were imposedacross mobile operators in practice variedwidely, driven by legal and institutionalcharacteristics of each country. Thisvariability allows us to identify andquantify the waterbed effect for the firsttime by looking at the impact ofregulation on prices (and profits) inreforming countries compared with thegeneral evolution of prices (and profits) innon-reforming countries.

Figure 1 plots the average retail pricefor mobile phone usage in countries thathave experienced a change in regulation,six quarters before and after theintroduction of regulation. Notice first thatcompared with prices in the rest of theworld, average prices in countries thatexperienced a change in regulation wereactually lower before the introductionof regulation. This is important

because it refutes the argument thatregulation was introduced as a result ofhigh retail prices for making mobile calls.

Most importantly, in line with ourwaterbed prediction, the introduction ofregulation has a clear positive impact onprices that becomes stronger as regulationbecomes progressively more binding.

The full empirical analysis allows us tocontrol both for common global trendsand for any country and operatorcharacteristics that remain constant overtime and may influence both regulationand prices. Our estimates suggest thatalthough regulation reduced terminationrates by about 10%, this also led to amore than 10% increase in mobileoutgoing prices on average.

But although the waterbed effect is


27

Figure 1:

Average mobile phone usage price around the introduction ofregulation of fixed-to-mobile termination charges

-0.10

-0.05

0.00

0.05

0.10

0.15

T+6T+5T+4T+3T+2T+1TT-1T-2T-3T-4T-5T-6

Quarters around the introduction of regulation (T), so T-1 means one quarter before the introduction of regulation

Ave

rage

pri

ce p

aid

(P

PP

-ad

just

ed e

uro

s/ye

ar)

per

usa

ge p

rofi

le

Price caps ontermination

charges have ledto significant

price increasesfor mobile phone

subscribers


large, our analysis also provides evidencethat it is not complete: accountingmeasures of profits are positively related totermination rates, thus mobile firms sufferfrom cuts in those rates as they possesssome degree of market power.

ConclusionsThe existence and magnitude of thewaterbed effect following the regulation oftermination rates is key to understandingthe social costs and benefits of theregulation. Regulators have generallyaccepted that the waterbed effect is likelyunder perfect competition, but they haddoubts about its validity and empiricalsignificance under conditions of imperfectcompetition. Our research shows that thewaterbed effect exists under a wider rangeof possible conditions and is bothsignificant and strong in practice.

These results have importantimplications for the mobile industry, butalso for other ‘two-sided’ markets’ –industries like shopping malls, credit cardsand dating agencies, where firms need toattract two (or more) groups of customersif they are to succeed.

■ First, it implies that any analysis of thecosts and benefits of regulation oftermination rates cannot ignore thepresence of the waterbed effect. Theimpact of regulation on (unregulated)prices for mobile subscribers should betaken into account when assessing theoverall costs and benefits of regulation.

■ Second, the mobile phone industryexhibits features typical of two-sidedmarkets. The market for subscriptionand outgoing services is closelyinterlinked to the market for terminationof incoming calls. As in any other two-sided market the structure of prices(who pays for what) is fundamentallyimportant for the development of the

market. Therefore, any regulatoryanalysis must take these linkages intoaccount either at the stage of marketdefinition or market analysis.

■ Finally, our analysis has implications forthe current debate about regulation ofinternational ‘roaming charges’ – theprices customers pay when using theirphones outside their home country.

The European Commission has votedto cap roaming charges for making andreceiving phone calls within the EU. Thesecharges account for 5-10% of operators’revenues globally and a larger proportionof their profits. The aim is to reduce thecost of making mobile phone calls whileabroad and hence encourage moreoverseas (but within EU) phone use.

A reduction in roaming charges maycause a similar waterbed phenomenon,whereby prices of domestic calls mayincrease as operators seek to compensatefor their lost revenue elsewhere. Whereaswith fixed-to-mobile terminationregulation, it was fixed users thatessentially subsidised mobile users, in thiscase, it would be the mobile subscribersthemselves that would bear the burden, ascalling abroad would be cheaper butcalling at home could become moreexpensive.

While the likely magnitude of thewaterbed effect caused by this newlegislation is debatable, our resultsdemonstrate that regulators have toacknowledge its existence and carefullyaccount for it in their calculations ofconsumer benefits.

This article summarises ‘Testing the

“Waterbed” Effect in Mobile Telephony’

by Christos Genakos and Tommaso Valletti,

CEP Discussion Paper No. 827

(http://cep.lse.ac.uk/pubs/download/

dp0827.pdf).

Christos Genakos is a lecturer in economics

at Cambridge University and a research

associate in CEP’s productivity and

innovation programme. Tommaso Valletti is

a professor of economics at Tanaka Business

School, Imperial College London.

Further reading

Mark Armstrong (2002), ‘The Theory of

Access Pricing and Interconnection’, in

Martin Cave, Sumit Majumdar and Ingo

Vogelsang (eds), Handbook of

Telecommunications Economics, North-

Holland.

Mark Armstrong and Julian Wright (2007),

‘Mobile Call Termination’, mimeo, University

College London.

Jerry Hausman and Julian Wright (2006),

‘Two-sided Markets with Substitution: Mobile

Termination Revisited’, mimeo, MIT.

Jean-Charles Rochet and Jean Tirole (2006),

‘Two-sided Markets: A Progress Report’,

RAND Journal of Economics 35: 645-67.

Tommaso Valletti and George Houpis (2005),

‘Mobile Termination: What is the “Right”

Charge?’, Journal of Regulatory Economics

28: 235-58.

Julian Wright (2002), ‘Access Pricing under

Competition: An Application to Cellular

Networks’, Journal of Industrial Economics

50: 289-316.

Regulation to caproaming charges

within the EU may makedomestic calls

more expensive

28

Publications

For further information on CEP publications please contact:The Publications Unit


Houghton Street, London WC2A 2AE

Telephone +44 (0) 207 955 7284

Email [email protected]

Recent Manpower Human Resources Lab DiscussionPapers available free from:http://cep.lse.ac.uk/pubs/opsp.asp?prog=MHRLDP

ARE THERE DAY OF THE WEEK PRODUCTIVITY EFFECTS?Alex Bryson and John ForthMHRL Discussion Paper No. 4July 2007

TEMPORARY AGENCY WORKERS AND WORKPLACE PERFORMANCE IN THE PRIVATE SECTORAlex BrysonMHRL Discussion Paper No. 3May 2007

THE SHIMER PUZZLE AND THE CORRECTIDENTIFICATION OF PRODUCTIVITYSHOCKSRégis BarnichonCEP Discussion Paper No. 823August 2007

SEARCH FRICTIONS, REAL RIGIDITIES ANDINFLATION DYNAMICSCarlos Thomas CEP Discussion Paper No. 822August 2007

THE EVOLUTION OF INEQUALITY INPRODUCTIVITY AND WAGES: PANEL DATAEVIDENCEGiulia Faggio, Kjell Salvanes and John Van ReenenCEP Discussion Paper No. 821August 2007

TRUST-BASED TRADELuis Araujo and Emanuel Ornelas CEP Discussion Paper No. 820August 2007

PRODUCTIVITY, AGGREGATE DEMANDAND UNEMPLOYMENT FLUCTUATIONSRégis BarnichonCEP Discussion Paper No. 819August 2007

IS DISTANCE DYING AT LAST? FALLINGHOME BIAS IN FIXED EFFECTS MODELS OFPATENT CITATIONSRachel Griffith, Sokbae Lee and John Van ReenenCEP Discussion Paper No. 818August 2007

MEASURING ORGANIZATION CAPITAL INJAPAN: AN EMPIRICAL ASSESSMENTUSING FIRM-LEVEL DATAYoungGak Kim and Tsutomu MiyagawaCEP Discussion Paper No. 817August 2007

SPEND IT LIKE BECKHAM? INEQUALITYAND REDISTRIBUTION IN THE UK, 1983-2004Andreas Georgiadis and Alan ManningCEP Discussion Paper No. 816August 2007

FREEDOM FRIESGuy Michaels and Xiaojia ZhiCEP Discussion Paper No. 815July 2007

TRADE LIBERALIZATION, OUTSOURCINGAND FIRM PRODUCTIVITYRalph OssaCEP Discussion Paper No. 814July 2007

WAGE DISTRIBUTIONS BY BARGAINING REGIME: LINKED EMPLOYER-EMPLOYEE DATA EVIDENCE FROM GERMANYKarsten Kohn and Alexander LembckeCEP Discussion Paper No. 813July 2007

THE EFFECT OF INFORMATION ANDCOMMUNICATION TECHNOLOGIES ONURBAN STRUCTUREYannis Ioannides, Henry Overman,Esteban Rossi-Hansberg and KurtSchmidheinyCEP Discussion Paper No. 812July 2007

THE DIVISION OF LABOUR,COORDINATION AND THE DEMAND FOR INFORMATION PROCESSINGGuy MichaelsCEP Discussion Paper No. 811July 2007

INTERGENERATIONAL MOBILITY AND THEINFORMATIVE CONTENT OF SURNAMESMaia Güell, Jose Rodriguez Mora andChris TelmerCEP Discussion Paper No. 810July 2007

HISTORY AND INDUSTRY LOCATION:EVIDENCE FROM GERMAN AIRPORTSStephen Redding, Daniel Sturm and Nikolaus WolfCEP Discussion Paper No. 809July 2007

THE MAASTRICHT CONVERGENCECRITERIA AND OPTIMAL MONETARY POLICY FOR THE EMUACCESSION COUNTRIESAnna LipinskaCEP Discussion Paper No. 808July 2007

Recent CEP Discussion Papers available free from:http://cep.lse.ac.uk/pubs/dp.asp

CentrePiece is the magazine of the Centre for

Economic Performance at the London School of

Economics. Articles in this issue reflect the

opinions of the authors, not of the Centre.

Requests for permission to reproduce the

articles should be sent to the Editor at the

address below.

Editorial and Subscriptions Office


London School of Economics

Houghton Street

London WC2A 2AE

© Centre for Economic Performance 2007

Volume 12 Issue 2

(ISSN 1362-3761) All rights reserved.

Centre Piece

The back issues

Date post:	06-Oct-2020
Category:	Documents
Upload:	others
View:	2 times
Download:	0 times

CentrePieceeprints.lse.ac.uk/47044/1/CentrePiece_12_2.pdfCentrePiece Autumn 20074 Another striking...

Documents