the magazine of the centre for economic...

ISSN 1362-3761

C e n t r e Pi e c eThe Magazine of The Centre for Economic Performance Volume 17 Issue 3 Winter 2012

America’s fiscal cliff Gender gapsImmigrant workers Peace dividendsHooray for GDP! Working from home

The UK’s housing crises

Getting back to growth remains an elusivequest across Europe. But beyond theshort-term need to reduce high levels ofunemployment, is increasing GDP –national output of goods and services –still the best measure of human progress?In a submission to the LSE GrowthCommission summarised in thisCentrePiece, Nicholas Oulton provides acombative response to those whoquestion the desirability of GDP growth.He argues that GDP remains a valuableindicator of the overall wellbeing ofsociety and that rising GDP remains apolicy objective that the majority ofpeople support.

A key element of growth, especially inthe UK, is the housing market. In ourcover story, Henry Overman evaluates thewide variety of proposals to address thebig problem of ‘not enough houses of thekind people want in the places wherethey want to live’. The solution, heconcludes, is simply to build more houses;

the challenge is to develop a planningsystem that allows the constructionindustry to respond to demand.

Another topic of constant publicconcern is immigration. Here, DavidMetcalf, chair of the Migration AdvisoryCommittee and a stalwart of the Centrefor Economic Performance (CEP), outlinesthe current policy framework. He notesthat there is a new focus on the gainsfrom immigration to UK residents ratherthan simply to GDP.

One high-profile immigrant in 2013will be the Canadian Mark Carney, thenext governor of the Bank of England.CEP’s director John Van Reenen haswelcomed the appointment: ‘It signalsthat top jobs in Britain are open to talentfrom all over the world – the opennessthat makes our universities and footballclubs successful. Let's hope that thegovernment's crazy net immigrationtarget doesn't mean that he’s denied awork visa!’

Elsewhere in this issue, we reportresearch findings on whether home-working is a good idea; on gender gapsin nineteenth century factories andtwenty-first century universities; and onthe value of ending conflict as reflectedin house prices. The ‘peace dividend’study is co-authored by Tim Besley, John’sco-chair of the Growth Commission,which will publish its recommendationsat the end of January.

The start of the new year may also be when the US economy falls over theso-called ‘fiscal cliff’. Our final articleexplains what’s going on and why it threatens a return to recession. That outcome would be yet anotherconstraint on Europe’s ability to return tosustainable growth.

Romesh Vaitilingam, [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, DesignRaphael Ltd

Print, Hastings Printing Co. Ltd

© Centre for Economic Performance 2012

Volume 17 Issue 3

(ISSN 1362-3761) All rights reserved.

Centre Piece

Editorial

CentrePiece Winter 2012

1

in brief...

page 2The UK’s housing crisesNot enough houses of the kind people want in the places where they want to live – that’s the big problem, according to Henry Overman

page 6Hooray for GDP!Nicholas Oulton responds to arguments that promoting thegrowth of GDP is undesirable or even irresponsible

page 12Immigration and the UK labour marketDavid Metcalf, chair of the Migration Advisory Committee,outlines recent changes in the regulatory framework forimmigrant workers

page 26America’s fiscal cliffEthan Ilzetzki and Jonathan Pinder explain the US policydebate on taxes, spending and public debt

Contentspage 18Working or shirking?Nicholas Bloom and colleagues analysethe first randomised experiment in havingemployees work from home

page 20Sexism at workTim Leunig and colleagues uncover paydiscrimination in Sweden’s late nineteenthcentury tobacco industry

page 22Science: why the gender gap?Thomas Breda and Son Thierry Ly ask ifprofessors are to blame for the lownumbers of young women in science

page 24‘Stop fighting and wealth will follow’Tim Besley and Hannes Mueller usehouse prices to measure the peacedividend in Northern Ireland

page 12Immigrationand the UKlabour market

page 18Working or

shirking?

page 26America’s fiscal cliff


2

The UK is facing not one but two housing crises,according to Henry Overman. The first is a short-term fall in prices and construction – which is both driven by and driving the recession.The second and more profound is the overallshortage of housing and the problems ofaffordability that this generates.

The UK’s housing crises

Reduced demand duringthe recession has beenreflected in falling houseprices outside London andthe South East and a

slump in construction everywhere. Datafrom the Department for Communitiesand Local Government show that betweentheir peak in 2007 and 2011, house prices fell by 12% in the North East andby 10% in the North West. BetweenSeptember 2011 and September 2012,land registry data show further price fallsof 3.2% in the North East and 2.2% inthe North West.

There has also been a severe slump inconstruction in these parts of the country.In the North East, construction (measuredas permanent dwellings completed) hasfallen by 36% from its 2007/08 peak. In the North West, construction has fallenby 47%.

Prices have held up much better inLondon and the South East. A large initialfall in London (9.3%) was quickly offset bysubsequent rises, leading to an overallincrease of 3.8% between 2007 and2011. The South East saw a similar initialfall and a slightly weaker recovery, butprices still increased by 1.4% between2007 and 2011. Between September 2011and September 2012, land registry datashow a continued recovery with pricesrising by 5.5% in London and by 2.3% inthe South East.

Despite these different pricetrajectories, construction has still slumped,even if the impact is somewhat lesspronounced: 12% down from the peakfor London; and 19% down from thepeak for the South East.

While we should not downplay short-term concerns about the impact of fallinghouse prices on consumer demand, it isthe construction figures that are mostworrying. The UK’s more profoundhousing crisis is the overall shortage ofhousing and the problems of affordabilitythat this generates (see, for example,Hilber and Vermeulen, 2012).

These problems, already acute in theSouth East, can only be exacerbated byrecent trends of slow income growth,falling construction and rising prices. Even with the general price falls inNorthern regions, affordability remains aproblem for the more successful cities,such as Manchester.

Focusing on the short-run slump alsodistracts from the longer-run problem thatthe UK was building very few houses evenduring the good times. At the height ofthe boom, England was building 170,000new homes a year. The annual averagebetween 1998 and 2007 was 150,000new homes.

To put these numbers in perspective,note that between 2001 and 2011, about1.4 million new homes were built inEngland, while the population rose by justunder four million. With an averagehousehold size of a little over two peopleper household, it is not surprising thatprices rose so sharply. The now defunctNational Housing and Planning AdviceUnit suggested that we would need tobuild around 270,000 homes a year just tostabilise (not reduce) prices.

Managing the demand for housingSo what can be done? Some have called

for the problem to be tackled ‘head-on’ byrent control measures to restrict risingprices. That would be great for peoplewho get rent-controlled houses, awful foreveryone else. In the long run, it wouldalso massively distort the rental sector(because it reduces the returns to rentingand increases the cost of moving). Quitesimply, this is a very bad option.

Instead of intervening directly tocontrol prices, an alternative set ofmeasures aims to intervene on thedemand side of the housing market. Thesemeasures are of two broad kinds. The firstseeks to dampen demand, for example,through changes in the way in whichhousing is taxed. Subjecting housing tocapital gains taxes would be a good if veryunpopular example.

To the extent that these arguments areabout reducing the pro-cyclicality of thehousing market (the tendency for demandto rise very rapidly as the economybooms), there is something to recommendthem. It is less clear to what extent theywould solve the fundamental problembecause most suggestions would havelittle effect on existing homeowners, butinstead reduce demand from youngergenerations (or renters). With the overallsupply of housing fixed, measures toimprove affordability need to haveprecisely the opposite incidence.

This leads to the second kind ofdemand-side intervention: policies toincrease demand. As with policies todampen demand, whether this helps interms of affordability depends on theextent to which it redistributes the abilityto demand housing from the house ‘rich’to the house ‘poor’.


3

The UK wasbuilding very

few houseseven when the

economy wasbooming


4

Once again, many recommendationsare likely to work in the wrong direction.For example, the Smith Institute hassuggested that the government shouldconsider tax concessions along the lines ofmortgage interest tax relief (Miras, whichwas abolished for principal residences in2000) to encourage access to homeownership.

Unfortunately, US evidence suggeststhat there is only a very weak linkbetween mortgage interest relief andhome ownership (Hilber and Turner,2010). In fact, in tightly regulated housingmarkets, relief has a negative effect onhome ownership because the price effect(through increased demand) more thanoffsets the income effect (from the taxdeduction). In less regulated markets, reliefdoes have a positive effect on homeownership rates, but only for higherincome groups.

As the UK market is very highlyregulated, these findings suggestconsiderable caution in using Miras as ameans to increase home ownership.Reintroducing it could prove to be a costlyand ineffective intervention, which has theopposite impact of that intended.

Unfortunately, many other demand-side proposals do not stand up to even themost basic scrutiny. The most recentexample is Nick Clegg’s suggestion thatparents should be allowed to use theirpension to help younger people buyproperty. A two-step assessment of this

policy (which can, of course, be applied tomany other housing policy initiatives)would run as follows.

First, how many people are likely to beaffected? This can be tricky to work outprecisely, but for Clegg’s announcement –as with a number of recent schemes – theconclusion from a rough estimate seemsto be ‘not many’. Second, if the policydoes affect relatively large numbers ofpeople, what will be the likely impact onthe market?

This second step involves applyingsome basic insights from supply anddemand. There are essentially two ways tohelp people struggling to get a toehold inthe housing market. One is to increase thesupply of (suitable) housing. The other isto redistribute some of the existinghousing stock from older people toyounger people. Nick Clegg’s proposaldoes neither of these things – so even if itwere to ‘work’, it would not ‘help’.

If it is hard to formulate gooddemand-side measures when the market isleft to its own devices, then anotherpossibility is to intervene directly toredistribute housing. Two popularexamples here relate to ‘empty bedrooms’and ‘empty houses’.

The Intergenerational Foundation hascalled for the government to adoptmeasures to stop older people from‘hoarding’ housing. Based on anassessment of housing ‘need’, policies likethis argue that we should take bedroomsfrom people who currently ‘under-occupy’their houses and give them to those wholive in overcrowded conditions. In practice,this tends to mean getting old people tomove out of large houses.

Ironically, the fact that planningdecisions are made on the basis of ‘need’but housing is allocated through themarket is one of the reasons why thehousing market is in such a mess. Marketsseek to balance supply and demand(rather than need) and it turns out that associeties get richer, they unsurprisinglytend to demand more space not less.

One response would be to switch to a‘needs’-based mechanism for allocatinghousing. My colleague Paul Cheshire haslight-heartedly suggested one option: ‘Ifwe are intent on allocating land for eachuse without regard to price, then logicallywe need to introduce space rationing. Ifprice does not determine the supply ofland, then price must not determine its

consumption. Each adult could, forexample, have a ration of, say, 40 squaremetres with dependent children having, say,another 20 square metres each. We could, if we wanted, even introduce atrading system so young adults or thosewilling to live in more cramped conditionscould sell some of their space ration,perhaps buying back space in later life’(Cheshire, 2009).

The Intergenerational Foundationsuggests something that seems lessextreme: measures to encouragehomeowners to consume less space. Thesewould be of two kinds. The first wouldstrongly penalise people who ‘over-consume’ space. Such penalties build upfrom a logic of housing need and areproblematic for all the reasons that spacerationing would be: who decides howmuch space is enough?

The second approach would be toremove barriers and distortions thatencourage people to over-consumehousing. I would have no problems withsuch measures apart from the fact that theywould be highly costly and remarkablyineffective.

Take, for example, the idea of removingstamp duty on people downsizing. At themoment, the huge wealth gain that theywould get by moving into a smaller houseis insufficient to offset the benefits ofstaying put. Removing stamp duty wouldchange this balance for a small number ofpeople at the margin, but it would be at ahuge cost to the public finances.

Removing single person allowances oncouncil tax or removing universal benefitsfor those in valuable houses would have asimilarly small impact on the number ofpeople willing to downsize. But it wouldimpose high costs on a small number ofpeople who are income-poor and who donot want to move for whatever reason. For more wealthy people, this wouldessentially be an irrelevance.

Changing the treatment of capital gainstax would provide a disincentive forownership (which may or may not be agood thing) but it would dampen theincentives to downsize. An annual capitalgains tax would constitute a punishmentbased on arbitrary decisions on how muchspace is enough.

What about empty homes? In England,there are around 280,000 homes that areempty for more than six months. But for2011, the data show that only 29,500 of

The system hasfailed to deliverenough housesof the kindpeople want inthe places wherethey want to live

these long-term empty homes were inLondon, with another 32,500 in the SouthEast. In other words, in high demand areas,very few houses are empty.

Using empty homes will (sometimes)make sense, but it will not do much tosolve the UK’s housing problem. Just aswith empty bedrooms, the reallocation ofempty homes does not represent a long-term solution to the housing crisis. So what can be done? The short answer isthat we must do things to increase thesupply of housing.

Increasing the supply of housingThe under-supply of housing in the UK hasbeen a long-term problem, which theprevious government was unable to tackleeffectively. Labour was slow to recognisethat something needed to be done aboutthe planning system. Once the problembecame clear, top-down regional planswere introduced, which tried to force localauthorities to build more housing.

These plans were very unpopular withlocal authorities in parts of the countrythat needed more housing and werequickly abolished by the coalitiongovernment. The new ‘national planningframework’ intends to replace the top-down system with more ‘localism’ and apackage of financial incentives toencourage development – with a target of240,000 new homes to be built each year(see Nathan and Overman, 2011).

These reforms should be welcomed fora number of reasons, but the governmentmay yet regret the immediate abolition ofregional plans. Uncertainty createsproblems for developers who tend torespond by postponing investment untilthat uncertainty is resolved. Add to this theeffects of the recession and you have twounderlying reasons for concerns about thesupply of housing.

In addition to these short-term issues,there is the longer-term issue of what thegovernment will do if its package offinancial incentives is insufficient toencourage more development. With thenew system yet to bed in, it could be a fewyears before the government is able toassess whether the system is working. Theassessment is likely to be close to anelection, when a change of governmentcould see a change of policy and yet moreuncertainty for developers.

Another area with which the previousgovernment struggled was its insistenceon high brownfield targets. There aresome problems with these targets, butthey remain very popular. This means thatthere is a danger that the coalitiongovernment will not be able to resist callsto strengthen constraints on building ongreenfield land. The government hasalready committed to maintaining greenbelts, but there are many other categoriesof ‘protected land’ where long-term policyremains uncertain.

But truly dealing with the problem ofaffordability requires a market-ledresponse in the areas of highest demand.This in turn requires the planning systemto allow a proper supply response.Addressing long-term affordability is not amatter of short-term stimulus. Instead, itrequires a private sector response whenthe market finally picks up. Developing aplanning system that allows that tohappen is the real challenge.

In all the debate around thegovernment’s planning reforms, we are indanger of losing sight of the fundamentalproblem: the current system has failed todeliver enough houses of the kind peoplewant in the places where they want tolive. Supposedly ‘radical’ solutions areeither insufficiently important to makemuch difference (empty homes) or soradical that it is hard to believe theyrepresent a good solution (emptybedrooms). The real solution isstraightforward: build more housing.


5

Henry Overman is director of the Spatial

Economics Research Centre (SERC),

professor of economic geography at LSE and

a research associate in CEP’s globalisation

programme.

For more commentary on housing

and other urban and regional policy

issues, see the SERC blog:

http://spatial-economics.blogspot.co.uk/

Further reading

Paul Cheshire (2009) ‘Urban Containment,

Housing Affordability and Price Stability –

Irreconcilable Goals’, SERC Policy Paper No. 4

(http://www.spatialeconomics.ac.uk/

textonly/SERC/publications/download/

sercpp004.pdf).

Christian Hilber and Tracy Turner (2010)

‘The Mortgage Interest Deduction and its

Impact on Homeownership Decisions’,

SERC Discussion Paper No. 55

(http://www.spatialeconomics.ac.uk/textonly/

SERC/publications/download/sercdp0055.pdf).

Christian Hilber and Wouter Vermeulen

(2012) ‘The Impact of Supply Constraints

on House Prices in England’,

SERC Discussion Paper No. 119


SERC/publications/download/sercdp0119.pdf).

Max Nathan and Henry Overman (2011)

‘Assessing the Government’s Proposals

to Reform the UK Planning System’,

SERC Policy Paper No. 11


SERC/publications/download/sercpp011.pdf).

Policies on‘empty homes’and ‘emptybedrooms’ do notoffer a long-termsolution to thehousing crisis

The real solution to the

UK’s housingcrisis is

straightforward:build more

houses


6

The idea of having GDP growth as the maintarget of economic policy has been under attackin recent years. Nicholas Oulton answers someof the criticisms and argues that continued GDPgrowth would be good for the UK – and not justin the short term to reduce high levels ofunemployment.

The much-loved poetJohn Betjeman isreported to have saidon his deathbed thatthe one thing heregretted in his life was

not having had more sex. This reminds usthat there is more to life than just buyingand consuming stuff. And that is whatGDP measures: the output of goods andservices on which we collectively spendour income.

Many people today would say thatpromoting the growth of GDP isundesirable or even irresponsible. Here Iconsider three common criticisms of GDPas a target of policy and explain why I think they are wrong:

� The first criticism is that GDP ishopelessly flawed as a measure of humanwelfare. For example, the argument goes,

it takes no account of pollution. � The second criticism is that GDPignores distribution. In a rich country likethe United States, some say, the typicalperson or family has seen little or nobenefit from growth since the 1970s. Atthe same time, inequality has risensharply. � The third criticism is that above acertain level, a higher material standard ofliving does not make people happier. Thisview concludes that we should stop tryingto raise GDP and look instead for policiesthat promote happiness.

‘GDP is a flawed measure of human welfare’GDP has always been a measure ofoutput, not of welfare. Using currentprices, it measures the value of goods andservices produced for final consumption,private and public, present and future.

(Future consumption is covered since GDPincludes output of investment goods.)Converting to constant prices makes itpossible to calculate the growth of GDPover time or the differences betweencountries across space.

But although GDP is not a measure ofhuman welfare, it can be considered acomponent of welfare. The volume ofgoods and services available to theaverage person clearly contributes towelfare in the wider sense, though ofcourse it is far from being the onlycomponent. So it is possible to imagine asocial welfare function that has GDP asone of its components alongside health,equality, human rights, etc.

GDP is also an indicator of humanwelfare. In cross-country data, GDP percapita is highly correlated with otherfactors that are important for welfare. Inparticular, it is positively correlated with

Hooray for GDP!


7

life expectancy and negatively correlatedwith infant mortality and inequality. Sinceparents naturally feel grief for childrenthey have lost, infant mortality might bethought of as an indicator of happiness.

Figures 1-3 illustrate these facts forlarge samples of countries, plottinghousehold consumption per capita (whichclosely tracks GDP per capita) againstthree measures of human welfare. Theyshow that richer countries tend to havegreater life expectancy, lower infantmortality and lower inequality. Of course,correlation is not necessarily causation,although there is a strong case for theview that higher GDP per capita leads toimproved health (Fogel, 2004).

According to the Commission on theMeasurement of Economic Performance,policy should be concerned withwellbeing, which encompasses manydimensions, including material living

standards, health, education, politicalvoice, social relationships and theenvironment (Stiglitz et al, 2009). Inresponse to the Commission’s report, boththe OECD and the UK’s Office forNational Statistics are now developingmeasures of these aspects of life.

Few will disagree that thesedimensions of life are important forhuman welfare and no one can object toimproved measurement. But for the UK, Iquestion whether the ONS is capable oftaking on a potentially vast newprogramme when even the basiceconomic statistics on which GDP restsare not fully in accordance with theOECD’s best practices for measuringproductivity and capital (Oulton, 2004a).

What’s more, we can go a long waytowards measuring welfare just using theapparatus of the national accounts.Martin Weitzman’s concept of Net

National Product (NNP) is key here. It isdefined in real terms as consumption plusnet investment (gross investment lessdepreciation), all deflated by the priceindex for consumption. Weitzman (1976)showed that his NNP could be thought ofas the yield on society’s wealth and wastherefore equal to the maximumsustainable level of consumption.

It is fairly simple to calculateWeitzman’s NNP from published nationalaccounts (Oulton, 2004b). In principle, wewould want to include the net change toall assets that are relevant to humanwelfare, including environmental stocks. Inpractice, we are a long way fromachieving this: for example, the UK’snational accounts include mineral oilexploration as part of gross investment butdepletion of oil and gas stocks byextraction is not included in depreciationand so official NNP is overstated.


8

Yet the statistical infrastructure builtto estimate GDP can be used to estimatea welfare measure such as Weitzman’sNNP. So GDP retains its usefulness as ameasure of output and as a welfareindicator.

‘Most people don’t benefitfrom GDP growth’Many people assert that the typical UShousehold’s living standards havestagnated since the 1970s, despite therelatively rapid growth of labourproductivity and GDP per capita. Butwhile it is uncontroversial that US incomeinequality has been rising for decades,does this mean that the typical householdhas received no benefit from growth? Theresults of a comprehensive recentexamination of these issues reveal quite adifferent picture (Wolff et al, 2012).

The measure of household welfarethat this study uses is what is known asthe Levy Institute Measure of EconomicWellbeing (LIMEW). Table 1 shows thegrowth rates of median and mean LIMEW(and GDP per capita) over the period1959-2007 and sub-periods within thatnear half-century. The difference betweenthe growth rates of mean and medianLIMEW is an indicator of changes ininequality: if the mean rises faster thanthe median, then inequality is increasing.

Over the whole period, equivalentmedian LIMEW grew at an annual rate of1.01%. The period 1959-72, supposedlythe golden age of economic growth, wasactually a comparatively poor one forhouseholds. And it was followed by a fall

in living standards over 1972-82.Far and away the best time for

households was the period 1982-89,which coincides roughly with the Reaganadministration. True, this excludes theVolcker deflation and recession of 1980-81: GDP per capita was 2.8% below its1979 level in 1982, which helps to explainsome of the subsequent rapid growth.But GDP per capita still grew at 2.43% ayear during the period 1980-88, which isfaster than in any sub-period except1959-72. Since 1989, the growth rateof living standards has beendeclining, but it has still beenpositive, even in 2004-07.

Table 1 also shows that equivalentmedian LIMEW grew less than half asfast as GDP per capita, which grewat 2.18% a year. What accountsfor this huge gap? It is partlydue to rising inequality sincefor nearly the wholeperiod, the mean was growingfaster than the median, theexception again being duringthe Reagan administration.

But this only accountsfor a small proportion ofthe gap. Most of it is accounted for bythree factors. First, the study deflateshousehold incomes by the consumerprices index: arguably, a better choicewould have been the price index forconsumption from the national accounts,which rises about 0.5% a year moreslowly. Second, GDP includes investmentas well as consumption – and investmenttends to rise more rapidly. Third, LIMEW

Table 1:

Real income measures, per capita and per household, in theUnited States: annual percentage rates of growth, 1959-2007

Sources: Wolff et al (2012) and US National Income and Products Accounts

Notes: LIMEW (Levy Institute Measure of Economic Wellbeing) is defined as income

minus taxes plus cash and non-cash benefits plus individual public consumption plus

household production, with property income valued on an annuity basis, per household.

‘Equivalent’ means that household income is measured after adjusting for household

size and composition.

1959-72 1972-82 1982-89 1989-2000 2000-04 2004-07 1959-20071. Equivalent median LIMEW 0.94% -0.13% 3.22% 0.97% 0.84% 0.42% 1.01%2. Equivalent mean LIMEW 1.11% 0.14% 3.27% 1.94% 0.10% 0.93% 1.31%3. GDP per capita 2.73% 1.34% 3.37% 2.03% 1.26% 1.58% 2.18%

GDP retains itsusefulness as a

measure ofoutput and an

indicator ofhuman welfare


9

includes a slow-growing component,household production, which GDPexcludes.

The typical US household hastherefore gained significantly from growthsince 1959 and also since 1980. Thisremains the case even though the medianhousehold would have gained more (tothe extent of 0.30% a year) if inequalityhad not widened. But most of the gapbetween the growth rates of GDP percapita and median LIMEW is not due torising inequality but to other factors.What’s more and contrary to the commonview, there were large gains in the 1980s,which continued, albeit at a slower rate,in the 1990s and even into the 2000s.

‘GDP growth doesn’t makepeople happier’Surveys of wellbeing or happinessrepeatedly find that in any given countryat any point in time, richer people reportthemselves to be happier than poorerpeople do. But when the same survey isrepeated in the same country over time,there is no rise in the average level ofhappiness, despite the fact that per capitaincome has gone up. Most of this timeseries evidence (which is disputed byStevenson and Wolfers, 2008) is for theUnited States and the result is known asthe ‘Easterlin paradox’ (Easterlin, 1974).

The most common explanation for theparadox, suggested by Richard Easterlinhimself, is that at least above a certainlevel of income, people care more abouttheir relative position in the income scalethan they do about their absolute

position. They are motivated less by puredesire for stuff and more by envy, by thepressure of ‘keeping up with the Joneses’and by the satisfaction of looking downon the less successful.

This explanation reconciles the cross-section and time series evidence. But itleaves the implication that stoppinggrowth would have no effect onhappiness. It also suggests that moreredistribution from rich to poor wouldraise overall happiness (provided it did notreduce GDP too much through adverseeffects on people’s incentives).

I find the results of these happinesssurveys puzzling because they areinconsistent with other facts aboutpeople’s behaviour. First, if people caremainly about their relative position, whyhas there been so much fuss about thefinancial crisis? After all, for most peoplein the UK, the drop in income has been(on this view) trivially small, no more than8% – and at least initially, it felldisproportionately on the rich.

Second, if people care about theirrelative position, why does this have to beexpressed in terms of annual income?After all, most workers today can workpart-time if they want. So why can’t Aboast that his daily rate of pay is higherthan B’s even if B’s annual earnings arehigher and this is because smart A worksonly three days a week while poor dumbB, a slave to the rat race,works five?

Yet surveys of part-time workers regularlyshow that many would

like to work longer hours if only theycould. And while it is true that someleisure activities like skiing require a lot ofcomplementary expenditure on stuff,many other activities – watching TV,surfing the internet, chatting with friendsin pubs or cafés or avoiding Betjeman’sregret – do not.

In fact, people’s leisure choicesprovide powerful evidence against theview that only relative position matters.

The classical economists argued thatthe amount of time people were

prepared to work depended on therange of goods and services

People’s choicesbetween work

and leisure show that they

value higherconsumption inabsolute terms

not just relativeto others

Despite wideninginequality, the typical US household has gained significantlyfrom GDP growthsince 1980


10

Figure 1:

The relationship between a country’s householdconsumption per head and its rate of infant mortality

Source: Oulton (2012); Notes: 146 countries.



Figure 2:

The relationship between a country’s householdconsumption per head and its life expectancy

Figure 3:

The relationship between a country’s householdconsumption per head and its inequality

1 10 100 200

2

10

50

100

1 10 100 200

40

80

90

70

60

50

1 10 100 200

65%

75%

55%

45%

35%

Infa

nt m

orta

lity

(dea

ths

per

1,00

0 liv

e bi

rths

)

Real household consumption per head (DRC=1)



Life

exp

ecta

ncy

(yea

rs)

Gin

i coe

ffic

ient

available for consumption. John StuartMill, for example, wrote this in hisPrinciples of Political Economy, firstpublished in 1871:

‘A people may be in a quiescent,indolent, uncultivated state, with all theirtastes either fully satisfied or entirelyundeveloped and they may fail to putforth the whole of their productiveenergies for want of any sufficient objectof desire. The opening of a foreign trade,by making them acquainted with newobjects or tempting them by the easieracquisition of things which they had notpreviously thought attainable, sometimesworks a sort of industrial revolution in acountry whose resources were previouslyundeveloped for want of energy andambition in the people: inducing thosewho were satisfied with scanty comfortsand little work, to work harder for thegratification of their new tastes and evento save and accumulate capital, for thestill more complete satisfaction of thosetastes at a future time.’

Let’s perform a simple thoughtexperiment. Imagine that over the 220 orso years since the Industrial Revolutionbegan, process innovation has taken placeat the historically observed rate but thatthere has been no product innovation inconsumer goods (though I allow productinnovation in capital goods).

UK GDP per capita has risen by afactor of about 12 since 1800 (Maddison,2003). So people today would havepotentially vastly higher incomes thanthey did then. But they can only spendtheir incomes on the consumer goods andservices that were available in 1800.

In those days, most consumerexpenditure was on food (at least 60% ofthe typical family budget), heat (wood orcoal), lighting (candles) and clothing(mostly made from wool or leather).Luxuries like horse-drawn carriages wereavailable to the rich and in my imaginaryworld, they would now be available tomany more people. But there would beno cars, refrigerators, washing machinesor dishwashers, no radio, cinema, TV orinternet, no rail or air travel and nomodern healthcare such as antibiotics and antiseptics.

How many hours a week, how manyweeks a year and how many years out ofan expected lifetime would the averageperson be willing to work? My guess isthat in this imaginary world, people


11

would work a lot less and take a lot moreleisure than real people do today.

After all, most consumer expenditurenowadays goes on products that were notavailable in 1800 and a lot goes onproducts not invented even by 1950.Today, only about 10% of the familybudget goes on food – and even withinthe food basket, many items (such asmicrowave-ready chicken tikka masala,the UK’s national dish) were not availablein 1800.

In summary, people’s choices betweenlabour and leisure demonstrate that theyvalue higher consumption in an absolutesense, not just a relative sense. So risingGDP per capita would be in accordancewith people’s desires and preferences.Philosophers and social critics may objectthat the average person’s desires andpreferences are trivial, ill informed andmisguided (an attitude which can betraced back at least as far as Plato’sRepublic). But in a democracy, people’spreferences should be respected.

Nicholas Oulton (2012) ‘The Wealth and

Poverty of Nations: True PPPs for 141

Countries’, CEP Discussion Paper No. 1080

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

dp1080.pdf).

Betsey Stevenson and Justin Wolfers (2008)

‘Economic Growth and Subjective Wellbeing:

Reassessing the Easterlin Paradox’,

Brookings Papers on Economic Activity.

Joseph Stiglitz, Amartya Sen and

Jean-Paul Fitoussi (2009) Report by the

Commission on the Measurement of

Economic Performance and Social Progress

(www.stiglitz-sen-fitoussi.fr).

Martin Weitzman (1976) ‘On the Welfare

Significance of National Product in a

Dynamic Economy’, Quarterly Journal of

Economics 90: 156-62.

Edward Wolff, Ajit Zacharias and Thomas

Masterson (2012) ‘Trends in American Living

Standards and Inequality’, Review of Income

and Wealth 58(2): 197-232.

Further reading

Richard Easterlin (1974) ‘Does Economic

Growth Improve the Human Lot? Some

Empirical Evidence’, in Nations and

Households in Economic Growth: Essays in

Honor of Moses Abramovitz edited by Paul

David and Melvin Reder, Academic Press.

Robert Fogel (2004) The Escape from

Premature Hunger and Death, 1700-2100:

Europe, America and the Third World,

Cambridge University Press.

Angus Maddison (2003) The World Economy:

Historical Statistics, OECD.

Nicholas Oulton (2004a) ‘A Statistical

Framework for the Analysis of Productivity

and Sustainable Development’, CEP

Discussion Paper No. 0629

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

dp0629.pdf).

Nicholas Oulton (2004b). ‘Productivity

versus Welfare: or, GDP versus Weitzman’s

NDP’, Review of Income and Wealth 50(3):

329-55.

This article summarises ‘Hooray for GDP!’

by Nicholas Oulton, CEP Occasional

Paper No. 30 (http://cep.lse.ac.uk/pubs/

download/occasional/op030.pdf) and a

submission to the LSE Growth Commission.

Nicholas Oulton is a senior visiting research

fellow at CEP.

Rising GDP percapita is in

accordance withpeople’s desiresand preferences


12


13

Net migration (immigrationminus emigration) intothe UK over the lastseven years totals aroundone million. In the period

since 1997, the annual immigrationinflow has doubled from around a quarterof a million to over half a million. Relativeto the size of the total population, suchnumbers are only matched by the inflowsof the seventeenth century, when theProtestant Huguenots were expelled from France.

Figure 1 shows inward and outwardmigration over the last half-century. Thestatistics are based on the United Nationsdefinition of a migrant: an individualintending to alter his or her country of

residence for over one year. In the UK,these data are collected by the Office forNational Statistics (ONS) through itsInternational Passenger Survey (IPS).

For the period 1964-82, net migrationwas negative except for one year. In thelast three decades, net migration hasbeen positive in all bar three years. Butthe story is much richer than this: in thefirst half (1983-97), net inward migrationwas never greater than 80,000; bycontrast, in the second half (1998-2011),it was always above 80,000. In four ofthose most recent years, the number wasover 200,000.

Initially, the higher inflow since 1998consisted of more workers from outsidethe European Economic Area (EEA)

coming to jobs in restaurants, health andcare work and information technology.Then in 2004, when eight central andeast European countries (the so-called A8)joined the European Union (EU), workrestrictions on their nationals were lifted,again boosting the inflow. Finally, in theperiod since 2005, the inflow of non-EEAstudents has doubled.

InstitutionsNet migration is the outcome of threedistinct reasons for migration – work,study and family (including asylum) – fromthree different geographical sources – theUK, the EU and outside the EU. Thisproduces a three-by-three matrix andeach of the nine cells has both an inflow

Changes in the UK labour market brought about byimmigration over the past 15 years always rankhigh among the public’s concerns. David Metcalf,chair of the Migration Advisory Committee and anactive CEP researcher for three decades, sets outthe numbers on net inward migration and outlinesrecent changes in the regulatory framework andother major policy initiatives.

Immigration and the UK labour market

Figure 1:

International migration to and from the UK, 1964-2011

Notes: Estimates for 1964-90 are from the

International Passenger Survey (IPS).

Estimates for 1991-2011 are for ‘long-term

international migration’, which include the

IPS estimates as well as estimates of flows to

and from the Republic of Ireland, asylum-

seekers and migrant-visitor switchers.

Source: Office for National Statistics-100

1964

0

100

200

300

400

500

600

700

566

350

216

80

1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012

■ Outflow

■ Inflow

■ Net■


14

and an outflow. The government can onlydirectly control the flows in the three non-EU cells. In broad terms, the net inflowfrom the EU equals the net outflow of UKcitizens – so the non-EU net figure issimilar to aggregate net migration.

In the mid-2000s, after almost adecade of rising inflows and substantiallyhigher net migration, the then-government decided to alter both theregulations around non-EU migration and the institutions central to the often heated and sometimes toxic debateabout migration.

First, the ‘points-based system’ (PBS) was introduced for economicmigration incrementally from 2008. The PBS replaced the previous 80-plusimmigration routes and it initiallyconsisted of five tiers.

Tier 1 covered highly skilled workerswho could come to the UK without a joboffer to search for skilled work. This was asupply-side initiative designed to boostthe stock of human capital in the UKworkforce.

Tier 2 was a demand-side scheme,covering skilled workers with a job offerwho filled a vacancy the employerotherwise could not fill from the UK and

EU labour markets. Tier 2 now has threemain routes:

� The ‘shortage occupation list’ route asdefined by the MAC (for example, MAC,2011). This covers skilled jobs that are inshortage and where it is sensible to fillthe vacancy with a worker from outsidethe EU. At present, the list covers jobswhere the number of employees (notimmigrants) totals less than 1% of the UKworkforce.� The ‘resident labour market test’ route.This is where an employer advertises thevacancy in the UK and/or the EU and if nosuitable worker responds, the firm canthen fill the vacancy from outside the EU.� The ‘intra-company transfer’ route.Traditionally this involved, for example, aJapanese auto engineer from Toyotacoming to work at the Toyota plant inDerby for a few years. But in recent years,the intra-company transfer route has beendominated by information technologyworkers coming mainly from India towork as third-party contractors.

Tier 3, which has never been activated,was for low-skilled workers. Tier 4 coversthe study route into further and higher

The governmentcan only directly

control migrationflows from

outside the EU

Lon

g-te

rm i

nte

rnat

ion

al m

igra

tion

to

the

UK

in

th

ousa

nd

s


15

education. And Tier 5 covers temporaryand youth mobility.

Each tier had points attached tocharacteristics. For example, for Tier 2,points initially varied according to age,qualifications and pay. The potentialmigrant had to pass a particular pointsthreshold and have a ‘certificate ofsponsorship’ (similar to the old workpermit) from a sponsor licensed by the UKBorder Agency.

The second change in the regulatoryframework was the establishment of theMigration Advisory Committee (MAC) toprovide independent, evidence-based andtransparent advice on migration issues.The MAC is economics-oriented with fouracademic economists plus the chair. TheUK Commission for Skills andEmployment is also represented on theMAC.

The key to the MAC’s modus operandiis that the government decides whichquestions and issues it wishes to beinvestigated. For example, successivegovernments have wanted greaterselectivity in non-EU migration. The MACwas asked to implement this, not todebate whether such selectivity isdesirable. Nevertheless, the MAC wascentrally involved in each of the threemajor policy initiatives discussed next.

Major policy initiativesUnder the coalition government, therehave been three major policy initiativesaround non-EU work migration. Theseinvolve greater selectivity, limits onmigrant numbers and a re-examination ofthe impact of immigration. The imperativeis strong for the coalition governmentbecause it has set a target for netmigration of below 100,000 by the endof this parliament in 2015. This impliesnearly halving the most recent figure of183,000 for the year ending March 2012.

Skill and pay thresholdsWhen the PBS was implemented in 2008,the minimum skill level for migrant jobswas set at NQF3 (which is roughlyequivalent to two A-levels). Subsequently,the threshold has been ratcheted up, firstin 2011 to foundation degree (NQF4) andin 2012 to degree level (NQF6). Theminimum pay threshold for Tier 2 jobs hasbeen raised in tandem with skills. All Tier2 migrants now have a default minimumof £20,000 a year, but the codes of

practice specify substantially higherminima for many occupations and jobs.

Occupations skilled to graduate levelare defined by the MAC according tothree measurable criteria: pay; theproportion of the occupation qualified toat least NQF6; and the skill level of theoccupation as determined by the ONS(MAC, 2012c). On this basis, 97 of the369 four-digit occupations in the 2010Standard Occupation Classification countas skilled to graduate level. These 97occupations employ roughly six millionfull-time workers.

There are around 28,000 job titlesdefined within the 369 occupations.Under the shortage route, stakeholderscan argue that a particular job title isskilled to graduate level even though theoccupation is not. Such arguments turnon innate ability and the training andexperience required for a job. Examples include chefs at the toprestaurants and ballet dancers in leadingcompanies such as Ballet Rambert or theRoyal Opera House.

The focus on skilled workers underTier 2 is sensible. Skilled workers are more

likely than low-skilled workers to becomplementary with capital and otherlabour. In addition, they may well havedynamic effects, for example, raising theproductivity of colleagues and innovating.And on average they make a muchstronger contribution to the publicfinances than the unskilled.

The emphasis on skills also suggeststhat the EU will continue to be the sourcefor any non-UK, low-skilled workers.Some sectors that previously relied on lessskilled non-EU migrants must now trainUK workers or look to the EU. Previously,a high proportion of employees in, forexample, Asian restaurants, care homesand work riders in racehorse training werenon-EU migrants.

Limits on migrant numbersAs part of the initiative to reduce netmigration, the MAC was asked in 2010 torecommend a limit on numbers enteringthe UK for work under Tiers 1 and 2. Wesuggested that the work route take itspro rata share of any reductions inimmigration (with the family and studyroutes taking their pro rata shares) andthat any such reductions were spreadequally over the four-year period 2011-15(MAC, 2010).

We recommended that the number ofentry clearance visas issued in 2011/12 forTiers 1 and 2 combined should bereduced by between 6,000 and 13,000.(Any cut in entry visa numbers is largerthan the reduction in the IPS numbersbecause some visas are not used andsome migrants come for under one year.)We recommended that between 37,000and 44,000 entry clearance visas beavailable. Drawing on stakeholderevidence, we suggested that Tier 1 should have a larger proportionate cutthan Tier 2.

The government essentially adoptedthe MAC’s recommended reduction inaggregate, but the cake was cut a littledifferently.

First, the government largely shutdown Tier 1. This route was designed forhighly skilled people without a job offer,but it also permitted any studentgraduating with a bachelor degree to stayin the UK for two years to search for askilled job – the ‘post-study work’ (PSW)route. There was some tentative evidencethat many Tier 1 entrants and PSWstudents were working in low-skilled jobs.

Greaterselectivity has

boosted the skilllevel of work

immigrants fromoutside the EU

and cut theirnumbers


16

Closing Tier 1 provided extra headroomfor Tier 2.

Second, a limit of 20,700 was put onTier 2 General (shortage and residentlabour market test routes), but the intra-company transfer route was not limitedby quantity but by price. The minimumannual pay thresholds were raised from£20,000 to £24,000 for intra-companytransfer workers coming for under a yearand to £40,000 for those coming formore than a year.

In the event, only half of the 20,700certificates of sponsorship available weretaken up in 2011/12 (roughly 1,500 forthe shortage route and 8,500 for theresident labour market test route).Therefore, the MAC recommended thatthe limit be unchanged for 2012/13(MAC, 2012b). The government adoptedthis recommendation and alsoannounced that this limit would remainfor 2013/14.

Figure 2 shows the impact of thepolicy initiatives on selectivity and limits.Between 2004 and 2011, non-EU workimmigration more than halved – from114,000 to 47,000. By contrast, inflowsfor study almost doubled – from 95,000to 180,000.

The impact of immigrationMost government policy initiatives requirean ‘impact assessment’. The MAC wasasked to analyse the approach taken inrecent immigration impact assessments(MAC, 2012a). Our recommendations,which have been adopted, will notinfluence immigration levels in theimmediate future, but they may well do soover the longer term.

Impact assessments concerningalterations to immigration had previouslyfocused on GDP and therefore suggestedgains from greater immigration and lossesfrom lower levels. This approach neglectedtwo key features of immigration: it involvesa change in population as well as output;and most of the gains go to the migrant.

The MAC therefore suggested a morenuanced approach that focuses on thegains or losses to UK residents (howeverdefined). This will normally involveanalysing five questions:

� What are the dynamic benefits? Theseinclude innovation and raising the humancapital and productivity of co-workers.Such benefits are elusive to measure, butthey may be of major importance.� What is the contribution of the migrant

Figure 2:

International Passenger Survey estimates of non-EUinflows by main reason for entry, 1991-2011

Notes: International Passenger Survey

estimates do not include flows to and from

the Republic of Ireland, asylum-seekers

and migrant-visitor switchers. The ‘other’

category includes those responses where

no reason was given.

Source: Office for National Statistics

Sectors that haverelied on less

skilled non-EUmigrants mustnow train UK

workers or lookto the EU

Skills policyremains centralto ensuring thatbusiness, health

and educationcan meet

their labourrequirements

1991

40

20

0

60

80

100

120

140

160

180

200

180

Points-based system introduced,

February 2008

For 1995 only, looking for work figures were included in Other.

17

5247

1993 1995 1997 1999 2001 2003 2005 2007 2009 2011

Formal study ■

Work ■

Other ■

Accom/join ■

Non

-EU

in

flow

s in

th

ousa

nd

s


17

to the public finances? Skilled workersgenerally contribute more than low-skilledworkers and those coming for familyreasons. The duration of the stay matters too.� Do the migrants displace UK workers?This is a controversial area. The MACreport suggested evidence of displacementin recession periods.� Do migrants impose ‘congestion costs’?This involves examining any extra burdenson the UK’s health, education, housing andtransport systems.� What are the distributional effects onUK residents of particular migrants? This isa rather neglected topic, but we need abetter understanding of whether it is high-or low-income groups that gain or lose, orwhether distribution is unaffected.Similarly, the distribution of gains betweencapital and labour are important.

Other policy changesThe MAC has produced a number of other reports whose recommendations –adopted by the government – will influence immigration and the labour market.

Regulations around settlement(‘indefinite leave to remain’) in the UK havegradually been tightened. On the issue ofwork routes to settlement, the MAC wasasked what level of pay a migrant shouldbe earning if and when he or she appliesfor settlement after five years or soworking here. We suggested a rangebetween £31,000 and £49,000, reflectingdifferent points on the pay distribution ofskilled workers. The government opted for£35,000, a higher pay threshold that mayraise the outflow of migrants.

Immigration for family reasons canhave an indirect impact on the laboursupply. The MAC was asked to recommenda minimum income level required tosponsor a spouse, so that the new familyunit would not be a burden on the state.We analysed the pay at which income-related benefits terminate (£18,600) andthe pay required for a neutral impact onthe public finances (£25,700). Thegovernment chose the lower figure, buteven this pay requirement excludes nearlyhalf the current sponsors.

When the A8 countries joined the EUin 2004, the UK put no restrictions on theirnationals’ right to work. Because mostother EU countries did restrict work rights,the A8 inflow was much greater than

expected. Therefore, restrictions were puton Bulgarians and Romanians when thosecountries joined the EU in 2007. In tworeviews (required by EU law), the MACrecommended retaining these restrictionson the grounds that lifting them wouldpotentially exacerbate the already seriouslydisturbed labour market.

But all such restrictions must end in December 2013. This raises aconundrum. The main source of work forBulgarians and Romanians is the ‘seasonalagricultural workers scheme’ (SAWS),largely picking fruit and vegetables. The National Farmers Union is concernedthat when the restrictions are lifted, theSAWS workers will instead choose to work in other sectors, such as constructionand hospitality.

The NFU is therefore suggesting areplacement scheme involving non-EUworkers, perhaps from Ukraine. This wouldimply a de facto opening of a less skilledlabour route – albeit only for temporarymigration – at a time when 25 millionpeople are unemployed in the EU. TheMAC will report on this matter in thespring of 2013.

Conclusions Non-EU work immigration is under control.Greater selectivity has boosted the skilllevel and cut the numbers. This greaterselectivity, coupled with the recession,largely explains why only half thecertificates of sponsorship available for Tier 2 are currently being used. Intra-company transfers in the informationtechnology sector continue to requirecareful monitoring, but limiting them byprice (minimum pay thresholds) isoperating well.

The MAC has been commissioned toanalyse whether there should be a sunsetclause for jobs on the shortageoccupations list for more than a certainperiod of years. Thus, skills policy remainscentral to ensuring that business, healthand education can meet their labourrequirements.

Finally, the new focus on gains fromimmigration to UK residents, rather thansimply GDP, puts the onus on firms andanalysts to demonstrate that dynamic andfiscal benefits from skilled immigrationoutweigh any displacement andcongestion costs.

There is a newpolicy focus onthe gains fromimmigration to

UK residentsrather than

simply to GDP

David Metcalf is an emeritus professor at

LSE, a research associate in CEP’s labour

markets programme and chair of the

government’s Migration Advisory Committee.

This article summarises a longer report, Work

Migration from Outside the European Union:

Fifteen Years of Turnabout, which will be

available in 2013.

Further reading

Migration Advisory Committee (2010) Limits

for Tier 1 and Tier 2 for 2011/12 and

Supporting Policies, November.

Migration Advisory Committee (2011) Fourth

Review of the Recommended Shortage

Occupation Lists for the UK and Scotland,

September.

Migration Advisory Committee (2012a)

Analysis of the Impacts of Migration,

January.

Migration Advisory Committee (2012b) Limits

on Migration: Limit on Tier 2 (General) for

2012/13 and Associated Policies, February.

Migration Advisory Committee (2012c)

Analysis of the Points Based System: List of

Occupations Skilled at NQF Level 6 and

Above and Review of the Tier 2 Codes of

Practice, October.

All these documents are available here:

www.ukba.homeoffice.gov.uk/mac

More than half of employees in the UK say that theywould like the chance to work from home, according to aYouGov survey conducted during the London Olympics.But while large numbers of office staff believe thatworking from home would improve their productivity,most admit that their firm does not allow it.

Firms elsewhere in the world have also been slow toadopt this management practice. Indeed, there are widelydifferent practices even among firms operating in thesame industry in the same country. In the US airlinebusiness, for example, Jet Blue allows all regular call-centre employees to work from home; Delta andSouthwestern have a policy of no home-working; whileUnited has experimented with a mix of practices.

In our research, we have had the opportunity to evaluateone such experiment. The firm, CTrip, is China’s largesttravel agent with 13,000 employees and a nearly $3billion valuation on NASDAQ. The firm wanted toexperiment with home-working before deciding whetherto roll it out across its whole operation.

The motivation was both to reduce office costs, whichwere becoming an increasingly high share of total costsdue to rising rental rates at the firm’s Shanghai base, andto reduce the firm’s high annual rate of staff turnover(50%). On the downside, the management wasconcerned that allowing employees to work at homeaway from the supervision of their managers could have anegative impact on their performance.

The experiment is unusual because one of our researchteam is also the co-founder and chairman of CTrip. Thishas naturally provided us with excellent access to both theexperimental data and also to the management’s views onworking from home. As such, the experiment provides aninsight into the adoption of a modern managementpractice by a large publicly listed firm, helping to addresssome of the questions about why so many firms fail toadopt potentially beneficial management practices.

CTrip decided to run a nine-month experiment with the Airfare and Hotel divisions in the firm’s Shanghaiheadquarters. All employees with at least six months’experience with the firm and their own room at homewere offered the option to work there for four days eachweek. Of the 508 eligible employees, 255 asked to workfrom home and after a lottery draw, those with even-numbered birthdays were selected for home-workingwhile those with odd-numbered birthdays stayed in theoffice to act as a ‘control group’.

Both home- and office-based employees worked the same shift period in the same teams under the samemanager as before, logged on to the same computersystem with the same equipment and the same workorder flow. The only difference between the two groupswas the location at which they worked.

So what were the results of CTrip’s experiment? First, theperformance of the home-workers went up dramatically,increasing by 13% over the nine-month experiment. Thisimprovement came mainly from a rise in the number ofminutes they worked during each shift, which was dueto a reduction in the number of breaks and sick daysthat they took. The home-workers were also moreproductive per minute worked, answering more calls,presumably because of the quieter working conditions at home.

Second, there were no negative effects on theemployees left working in the office: there was nochange in the performance of the control group. Third,the rates of staff turnover fell sharply for the home-workers, dropping by almost 50% compared with thecontrol group. The home-workers also reportedsubstantially higher work satisfaction and less ‘workexhaustion’ in a psychological attitudes survey.

At the end of the experiment, CTrip’s management teamwas so impressed by the success of home-working thatthey decided to roll it out to the entire firm. They also


18

Working or shirking?Do employees who start working from home improve their productivityor does it simply lead to ‘shirking from home’? The first randomisedexperiment on home-working – run by a big Chinese travel agent andanalysed by Nicholas Bloom and colleagues – has produced encouragingresults for both staff performance and staff satisfaction.

in brief...

Employees who started working from homeimproved their productivity dramatically


19

offered both the original home-workers and the controlgroup a fresh choice of working arrangements.

To their surprise, a half of the home-workers changedtheir minds and returned to the office, citing theloneliness of working at home. At the same time, threequarters of the control group – who had initially allrequested to work from home – decided to stay in theoffice. This outcome suggests that before theimplementation of these types of management practices,their likely effects are as unclear to employees as they areto managers. It may also help to explain the typically slowadoption of such practices.

How do our findings compare with previous research?There is an extensive body of case studies of individualfirms that have adopted home-working programmes, andthey tend to show large positive impacts. But therobustness of these results are hard to evaluate because ofthe non-randomised nature of the programmes, both interms of the selection of firms into programmes and theselection of employees to work at home.

This self-selection effect is evident even in the case ofCTrip: when the firm allowed a general roll-out of home-working, employees performing well at home typicallychose to stay home while employees performing badlyreturned to the office. We are continuing to collect dataon both current and former employees to evaluate longer-run impacts on recruitment, promotion and other workand non-work outcomes.

This article summarises ‘Does Working from

Home Work? Evidence from a Corporate

Experiment’ by Nicholas Bloom, James Liang,

John Roberts and Zhichun Jenny Ying

(www.stanford.edu/~nbloom/WFH.pdf).

Nicholas Bloom, a professor of economics at

Stanford University, is a research associate in

CEP’s productivity and innovation programme.

James Liang is the co-founder and chairman

of CTrip. John Roberts and Zhichun Jenny

Ying are at Stanford University.

The success of CTrip’s experimentled to home-working being

rolled out across the firm


20

Employers have paid men more than women since thebeginning of time: according to the Bible, the Lord toldMoses to value a female servant at three-fifths the value ofa male servant. But economists don’t see such headlinewage differentials as proof that there is sexism in the labourmarket. After all, men and women may have different levelsof productivity.

One reason for productivity differences is that men aretypically stronger than women. This allows them to earn apremium, particularly in the past when strength was oftenimportant. In many jobs, experience helps and men – whoare less likely to take time out to bring up children or tocare for the elderly – have generally worked longer. Thisaffects the jobs that men can get and the wages theyreceive for any given job.

It can be hard therefore to work out whether a particularwage differential represents sexism or not. Many jobs areundertaken almost exclusively by one gender or the other.Even when both genders work alongside each other,productivity is often hard to measure.

But in recent research, we have been fortunate enough toget access to just such measurements. In 1898, the Swedishgovernment commissioned a remarkable survey, in whichstatistician Henning Elmquist and his team interviewed everyfirm and every employee in the country’s tobacco industry.

We know the job each worker did, the hours they workedand the earnings they received. In addition, we know ahost of other details – the workers’ ages, their experienceon the job, whether they had children, whether they wereunion members and so on. All of the individual datasurvive, giving us a high quality employer-employeematched data set.

Cigar-making is a good industry to investigate. Strengthwasn’t important, but experience did matter. Rolling cigarswas skilled work, even if you were using a wooden mouldto help you. A few factories were exclusively male and afew were exclusively female, but in most cases men andwomen worked alongside each other doing exactly thesame job.

Our study uses a modern analytical technique to calculatewhether men and women were paid fairly, given their levels

of experience and so on. In essence, this means we workout what women would have been paid had they been paidthe same as men of the same age, experience and so on.We can assess the extent of discrimination by comparingwhat they ‘should’ have been paid with what they wereactually paid.

Some workers were paid by the hour – those preparing theraw tobacco, for example. Here we find clear evidence ofsexism. Men were paid ‘men’s wages’ and women werepaid ‘women’s wages’ – and there was a big differencebetween the two.

Other workers were paid a piece rate, notably cigar-rollers,who were paid by the cigar. Here we find a completelydifferent result: all of the difference in the earnings of menand women can be explained by differences between thecharacteristics of men and women rolling cigars. Men were older, more experienced and more productive. Asa result, they produced more cigars per week and earnedmore per week in direct proportion to their additionaloutput.

What’s more, we find that men and women were equallylikely to get promoted to cigar-rolling, given theircharacteristics. Experienced workers got offered thesepositions, irrespective of gender. There was no gender baror discrimination against women at this level. There was a‘glass ceiling’, but it was higher up: all of the supervisorsand managers were male.

These results show that the Swedish tobacco firms wereinternally inconsistent. They were sexist when it came topaying workers by the hour. But they were gender-blindwhen it came to putting the best workers into their piecerate sections and paying them equally well. In all cases, themagnitudes are sufficiently large that firms must have beenaware of the inconsistencies in their behaviour.

Pay rates and pay systems don’t operate in a vacuumbecause firms don’t operate in a vacuum. For men paid bythe hour, firms had to pay the economy-wide ‘going rate’for male labour. If they didn’t, the men would have left forother jobs. So even though cigar-making didn’t requirestrength, most men would be able to get a job that did andtherefore commanded the ‘strength premium’ even whentheir strength wasn’t used.

Sexism at workAre differences in men and women’s pay an indication of discriminationin the labour market? In a study of Sweden’s cigar-making industry inthe late nineteenth century, Tim Leunig and colleagues find clearevidence of sexism at work.

in brief...


21

In contrast, women’s primary alternative employment wasas domestic servants, where wages were low. Firms wereable to attract women even when they offered pay ratesfar lower than those paid to men.

Men paid by the hour represented bad value for money tothe firms that employed them because they were paid adisproportionately large premium relative to theirproductivity advantage. Firms seem to have recognised this:the industry was almost all male in 1860, but it feminisedsteadily thereafter.

The presence of more than 50 Swedish tobacco firms wassufficient to make the industry reasonably competitive,even though imports were low. After all, cigars are lightand cheap to transport.

Economists have always instinctively believed thatcompetition disciplines firms. A firm that employsunsuitable workers or pays a wage disproportionate to their productivity will increase its costs and lower its profits. This means that a firm takes a hit every time it discriminates. The implication is that firms thatemployed a disproportionately large number of men wouldhave higher costs, lower profits and thus be less likely to survive.

Using data on firm survival rates, we are able to test theeffect on firm survival of a more feminised workforce. The results are clear: the more men that a firm employed,the more likely it was to go out of business between 1863and 1915. Employing men made bad commercial sense.

In this way, product market competition imposed adiscipline on firms. At the going male pay rate, theycouldn’t employ men without risking their firm’s verysurvival. But we shouldn’t get too excited about the effecton women’s lives. Product market competition may havecreated jobs for women, but this only happened becausewomen’s wages in the wider labour market were lowerthan those for men. The market delivered more jobs forfemale cigar-makers, but it didn’t deliver equal pay.

This article summarises ‘Gender, Productivity and the Nature

of Work and Pay: Evidence from the Late Nineteenth-century

Tobacco Industry’ by Tim Leunig, Maria Stanfors, Björn

Eriksson and Tobias Karlsson, CEP Discussion Paper No. 1053

(http://cep.lse.ac.uk/pubs/download/dp1053.pdf).

Tim Leunig is reader in economic history at LSE, an associate

in CEP’s research programme on globalisation and currently a

policy adviser in the UK government’s Department for

Education. Maria Stanfors, Björn Eriksson and Tobias

Karlsson are at Lund University.

The market deliveredmore jobs for female

cigar-makers, but itdidn’t deliver equal pay

For the Swedish tobaccoindustry, employing men made

bad commercial sense

ww

w.t

obak

soch

tand

stic

ksm

useu

m.s

e

Gender differences have disappeared in many educationalsettings, yet male and female students remain stronglysegregated across what Americans call their ‘majors’, themain subject they study at university. Of the US workforcewith a background in science, technology, engineeringand maths (STEM), only a quarter are women. And in theUK, the Institute of Physics recently reported that physicsis the fourth favourite A-level subject for boys but only thenineteenth most popular among girls.

Understanding the origin of these discrepancies isimportant from an economic perspective. Genderdifferences in entry into science as opposed to non-science careers account for a significant part of the genderpay differential among graduates. They may also reduceaggregate productivity because of misallocation of talent.

The reasons for the under-representation of women inscience have been debated in research papers andgovernment reports. We know that gender differences inmaths and science test scores at 15 years old have fallenin recent decades and, according to the OECD’sProgramme for International Student Assessment (PISA),they are now very small in most developed countries.(PISA also shows that the UK’s gender differences in mathsare among the highest, probably because of earlyspecialisation.)

These small gender differences in ability cannot explainthe gender gap that emerges later on in young people’sscience careers. For example, even looking at studentswith identical abilities, women are still between 50% and70% less likely than men to complete a degree in theSTEM subjects.

So what explains the gender gap in science? A potentialexplanation is that women are discouraged or evendiscriminated against by professors when they choose to study science at university. Evidence certainly indicates that professors influence students’ educationalchoices by serving as role models: having a femaleteacher in traditionally ‘masculine’ subjects stronglyincreases female students’ attainment and their likelihoodof majoring in science.

Other studies also suggest that gender stereotypes –

such as ‘boys excel at maths and science while girls dobetter in other subjects’ – may serve as the basis forsteering young women towards more ‘female’ occupationsand be partly responsible for gender gaps at university andin the labour market.

But to date, there is almost no evidence of a direct linkbetween stereotypes and discrimination. We don’t reallyknow how professors in different subjects evaluate their students. Do science professors want femalestudents? This is a key concern if we are to ensure thatyoung men and women are given equal opportunitiesand are equally treated when they make their educationaland career choices.

In our research, we use a unique dataset on the entranceexam for a leading higher education institution in France –the École Normale Supérieure (ENS) – to investigate thepotential links between gender stereotypes anddiscrimination. To get into the ENS, each student is testedon subjects where boys are usually thought to be betterthan girls – maths and physics, for example – as well as onsubjects that are assumed to be better suited for girls – forexample, biology and foreign languages.

This specific context enables us to identify precisely howboth the direction and degree of gender discriminationvary with gender stereotypes. We use the fact thatcandidates to get into the ENS have to take both a blindwritten test (their gender is not known by the professorwho marks the test) and a non-blind oral test.

The ‘difference-in-differences’ between the young menand women in the blind and the non-blind test scoresgives a measure of professor-driven genderdiscrimination in a given subject. Moreover,since students are tested in more than onesubject, it is possible to investigate howprofessors’ gender bias varies across subjectsfor the same candidate.

We find that discrimination goessystematically against gender stereotypes: themore masculine a subject is thought to be,the more favoured are the female candidates.In maths and physics tests, for example,


22

Science: why the gender gap?Are professors to blame for the fact that there are so few youngwomen in science? Analysing data on the entrance exam for atop French university, Thomas Breda and Son Thierry Lyargue that it is not the selection process in higher education thatis perpetuating this gender gap.

in brief...


23

young women overtake about 10% of the young mendue to discrimination while the exact opposite happens inbiology and foreign languages tests.

This implies that the demand for students in differentmajors is biased in favour of the minority gender: forexample, the share of female students who are admittedto major in maths and physics jumps from 8% to 12%.These results show that professors’ evaluations are notdirectly driven by simplistic stereotypes such as ‘girls areno good at science’.

Having seen that professors react to gender stereotypes ‘inopposition to them’, we may wonder how candidatesthemselves react to these gender stereotypes. After all,our study focuses on a very competitive contest: it may bethat the female candidates at the ENS feel especially self-confident in maths, which explains their goodperformance in the oral tests.

But this is not what we find. The performance of thefemale candidates we analyse is consistent with what isusually found in other contexts: although the differencesare small, female candidates tend to perform slightlyworse in written tests in more male-dominated subjects(such as maths) and slightly better in more female subjects(such as foreign languages). What’s more, when they haveto choose an additional test, females are a lot less likely tochoose the most masculine one. This is true evencomparing candidates with the same ability.

These results imply first, that in opting for non-sciencesubjects, young women behave exactly as the stereotypeswould predict; and second, that this choice is irrationalgiven professors’ actual evaluations of their performancein masculine subjects. To maximise their chances ofsuccess, young women should choose masculine subjectsmore often and benefit from professors’ seeming biasagainst gender stereotypes.

Different mechanisms could explain the fact thatprofessors tend to favour young women in typically male-dominated subjects. One is that we may simply observe‘affirmative action’ to produce more equal sex ratios in thedifferent majors. But unlike in the United States, there isno legal base for affirmative action in France. The ENS isalso one of the most prestigious higher education

institutions in the country and it has a strong reputationfor rewarding pure talent only. Thus, there are probablyno coordinated decisions among the professors towardsfavouring female candidates for science majors.

This leaves us with two other possible explanations. Thefirst is pure preference-based discrimination: mathsprofessors are just happier when they have the unusualoccasion to interview a female candidate; and the sameis true for literature professors with respect to malecandidates.

The second and more plausible mechanism is directlylinked to students’ abilities. Paradoxically, professors mayrationally favour young women in science even if theyhave negative stereotypes about their abilities. For agiven test performance, the professor may think that thefemale candidates signal higher effort, self-investment orperseverance and they therefore reward these non-cognitive attributes.

These mechanisms need to be investigated further. Butwe already know that stereotypes do not always harmyoung women, which can be seen as good news aboutthe capacity of our societies to move quickly fromawareness to action against longstanding imbalances. Itwould be valuable to know if such behaviours arealready widespread and to what extent they may help toreduce the very large gender gap in science that stillexists in most countries.

This article summarises ‘Do Professors Really

Perpetuate the Gender Gap in Science?

Evidence from a Natural Experiment in a

French Higher Education Institution’

by Thomas Breda and Son Thierry Ly,

Centre for the Economics of Education

Discussion Paper No. 138

(http://cee.lse.ac.uk/ceedps/ceedp138.pdf).

Thomas Breda is a research fellow in CEP’s

labour markets programme. Son Thierry Ly is

a PhD student at the ENS and the Paris School

of Economics.


24

When Nobel peace prize-winner Archbishop DesmondTutu delivered a sermon for South Sudan’s first birthday inJuly 2012, he told the nation’s leaders: ‘stop fighting andwealth will follow’. But what is the evidence that endingconflict can deliver economic regeneration? And might the possibility of a ‘peace dividend’ itself contribute tolasting peace?

The peace process in Northern Ireland provides anopportunity to estimate the value of peace. Violentconflict flared up there in the late 1960s and ultimatelyclaimed around 3,500 lives before the Downing StreetDeclaration was signed nearly 20 years ago in December1993. The conflict caused major economic dislocation, butas peace took hold, the economy began to be repaired.

In a recent study, we use house prices to measure thepeace dividend. Our basic idea is that willingness to payfor a house will reflect investors and homeowners’perceptions of the value of peace. Between 1992 and2009, the average house price in Northern Ireland morethan quadrupled. But before assuming this represents thevalue of peace, we need to recognise that the period sawprices rising similarly in England and indeed across theBritish Isles.

We know that while some regions in Northern Irelandwere heavily affected by the violence, others werepeaceful throughout the conflict. We can therefore drilldown into the housing data to compare prices in violentand non-violent regions before and after the cessation of violence.

It turns out that after the peace process began, houseprices started to converge (as Figure 1 illustrates). Thepeace dividend can then be measured as the difference inhouse prices between the violent and peaceful regionsduring the conflict.

One of the biggest concerns during any peace process isthe chance of a relapse. Despite the Omagh bombing in1998, which killed 29 people, Northern Ireland saw afairly permanent shift away from violence after 1993. Thismatters for estimating the peace dividend, especially whenpeople are making long-term investment decisions such asbuying houses.

At what stage did expectations about the future changeand by how much? Research methods developed foranalysing business cycles can be used to deal with theproblem that we don’t know exactly when peace arrived

‘Stop fighting and wealth will follow’Tim Besley and Hannes Mueller use house prices to measurethe peace dividend in Northern Ireland – and draw lessons forother parts of the world that have endured long conflicts.

in brief...

As the peaceprocess in NorthernIreland began,house prices inviolent andpeacefulregionsconverged

in Northern Ireland and whether its timing varied acrossregions. Data on killings are potentially informativeabout what people care about – peace or conflict – sothey can measure expectations about peace.

This approach can be used to explore the size of thepeace dividend region by region. We find that Belfastbenefited most with an increase in house prices ofbetween 6% and 17%. Mid- and Southwest Down andLondonderry/Strabane experienced house price increasesof between 2% and 8%. Other regions – North Down,for example – were largely unaffected by the conflict andtherefore did not benefit as much.

Did the peace dividend help to build peace? Anecdotessuggest that even Republican sympathisers weregambling on houses on Belfast’s sectarian dividing linesin the 1990s. We have heard of one who bought sixhouses nearby, paying around £7,000 per house in thebelief that if he spent £3,000 on each and the peaceprocess held, then they would be worth about £35,000apiece. If the peace process didn't hold, the investmentwould become worthless. In this way, the possibledividends helped to create a vested interest in peace andmight have helped to stabilise it.

Every conflict is different, but what might our findingsimply elsewhere? Baghdad, for example, is a city 10-15times the size of Belfast and where the current level ofviolence per capita is around 2.8 times higher than inBelfast during the 1980s. But the mixture of insurgencyand sectarianism has parallels with Northern Ireland.

We can use our estimates from Northern Ireland to askwhat would happen if Baghdad entered a phase ofpeace comparable to the peace in Belfast. These suggestthat the peace dividend in Baghdad would be between16.4% and 46.4%.

Stretching our argument even further, we can get an ideaof the economic effects of current events in Syria. The cityof Homs experiences violence per capita more than 48times as intense as Belfast in the 1980s. If there were afunctioning housing market, our estimates suggest acollapse in house prices to only an eighth to a third ofwhat they were before the conflict started. Reversing thiscollapse suggests a sizeable stake for citizens in endingthe conflict if only this could be engineered.

So the example of Northern Ireland is not only useful forcalibrating the peace dividend: it can also give hope toother places where long conflicts have reduced the qualityof life. The potential gains and losses in the value of fixedassets vividly illustrate Archbishop Tutu’s maxim: stopfighting and wealth will follow.


25

This article summarises ‘Estimating the Peace Dividend:

The Impact of Violence on House Prices in Northern Ireland’

by Tim Besley and Hannes Mueller, American Economic

Review 102(2): 810-33.

Tim Besley is School Professor at the LSE and co-chair

with CEP’s director John Van Reenen of the LSE Growth

Commission. Hannes Mueller is at the Barcelona Graduate

School of Economics.

1990q1

10.4

10.6

10.8

11

11.2 Downing Street Declaration

Peace dividend ■

1992q1 1994q1 1996q1 1998q1

Six least violent regions ■

■ Five most violent regions

■

The possibility of a peace dividend

can itself help to build peace

Figure 1:

Average house price movements in the most andleast violent regions in Northern Ireland around the time of the peace process

In (

mea

n o

f h

ouse

pri

ces)


26

There is cross-party consensusin the United States thatpublic debt levels are a seriousproblem, at least in the

medium term. Debt levels rose significantlyin the Great Recession (see Figure 1), butmuch of the increase can be explained bythe automatic responses of publicspending and taxes to the state of thebusiness cycle.

The more worrying fact is that thelong-run debt trend is clearly upwards.Public spending has risen secularly overthe past decades, due to increases inpublic healthcare (Medicare for the elderlyand Medicaid for the poor) and socialsecurity (pensions), up from 3% of GDP inthe 1950s to almost 12% today andprojected to rise to 16% by 2037. Since

tax revenues have not kept up with thesespending trends, public debt has beenmarching upwards. In contrast, publicconsumption as a share of income has notincreased since the 1950s and publicinvestment has actually declined (seeFigure 2).

In December 2010, the bi-partisanNational Commission on FiscalResponsibility and Reform (known as‘Simpson-Bowles’ after the two co-chairs)released a majority report proposing amixture of cuts in entitlement spendingand tax reform, backloaded to avoidexacerbating the recession (see Table 1).

The commission recommendedeliminating most tax deductions,increasing revenue by about $1.1 trillion.Part of this higher revenue would be used

to reduce tax rates and overall taxrevenues would be targeted as less than20% of GDP in the long run. Theremainder would be allocated to debtreduction.

Social security would be brought intobalance through broadening payroll taxbases and increasing the retirement age.Simpson-Bowles recommended settingtargets to contain Medicare’s growthbeyond 2020 (unfortunately without muchdetail) and containing discretionaryspending growth to half the rate ofinflation.

Although the Commission’s proposalsare broadly seen as the starting point forany serious reform, President Obama hasnot fully embraced them and Paul Ryan, acommission member and the defeated

After the presidential election, US politicalattention turns to addressing the challenge of thecountry’s long-term public debt. Ethan Ilzetzkiand Jonathan Pinder explain the most prominentproposals for reform – and the economics of theso-called ‘fiscal cliff’.

America’s fiscal cliff

include tax increases. With no ‘grandbargain’ on resolving longer-term debtproblems, agreement was reached to raisethe debt ceiling temporarily and to find acompromise on longer-term challengesover the coming year.

To give incentives to both sides toarrive at a long-term compromise, the


27

vice-presidential candidate in November’selection, explicitly voted against themajority report. Simpson-Bowles did notobtain the super-majority required to bringforward legislation.

Political conflict peaked in the summerof 2011 over the ‘debt ceiling’. Due to(outdated) rules from the early twentiethcentury, Congress is required not only toapprove tax and expenditure laws in itsbudget but also, in separate legislation, toset a limit on total government debt. As USfederal debt approached this limit in July2011, Congress was unable to reach anagreement on a change to the debt ceiling.

Republicans demanded that increasesin the debt ceiling be linked to legislationon spending restraint. Democrats insistedthat the debt ceiling be increasedunconditionally or for the agreement to

Figure 1:

Debt of US federal government (percentage of GDP)

100%

90%

80%

70%

60%

50%

40%

30%

20%

1970 1973 1976 1979 1988 1991 19941982 1985 1997 2000 20062003 2009 2012

■ Held by the public

■ Total

Figure 2:

Expenditure breakdown of US federal government (percentage of GDP)

19%

17%

15%

13%

11%

9%

7%

5%

3%

1%

1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012

■ Social transfers

■ Government investment

■ Government consumption

■ Interest payments

Source: Bureau of Economic Analysis

Source: Bureau of Economic Analysis

The tax rises andspending cuts of

the fiscal cliffwould almost

certainly plungethe US economy

into recession


28

The fiscal cliffdoes not address

the root causesof the US debt

problem –healthcare and

pensions

This article is an extract from a briefing

published as part of CEP’s series of US

Election Analyses: ‘Recession and Recovery:

The US Policy Debate on Taxes, Spending and

Public Debt’ by Ethan Ilzetzki and Jonathan

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

download/cepusa001.pdf).

Ethan Ilzetzki is an assistant professor in

LSE’s economics department and a research

associate in CEP’s macroeconomics

programme. Jonathan Pinder is a PhD

student at LSE.

TAX POLICYIncome taxes

Eliminate most ‘income tax

expenditures’, that is, all

deductions from income taxes.

Current income tax expenditures

are estimated at $1.1 trillion

annually. Use part of the savings

to lower tax rates, limiting the

top income tax rate to 29% and

maintaining or increasing the

progressivity of the tax code.

Payroll taxes

Broaden the base of social

security taxes to apply to 90%

of personal income by 2050.

Corporate taxes

Lower the corporate tax rate to

no higher than 29%. Eliminate

all ‘tax expenditures’ for

businesses. Move to a territorial

tax system.

Revenues

Revenues to increase gradually,

stabilising at just under 20% in

the long run.

EXPENDITURE POLICYDiscretionaryHold spending in 2012 equal to

or lower than spending in 2011

and return spending to 2008

levels in real terms in 2013.

Limit future spending growth

to half the projected inflation

rate through 2020. Require

equal cuts from both security

and non-security spending.

Medicare and social security

The commission only proposes

small fixes to Medicare in the

short run, while setting targets

to contain the programme’s

rate of growth after 2020.

Increase the social security

retirement age to 67 by 2027

and index the retirement age to

average life expectancy

thereafter. Index social security

benefits to chain-indexed CPI.

The plan is projected to close

the social security shortfall over

a 75-year horizon.

DEFICIT Reduce the deficit gradually to

2.3% by 2015, with most

deficit reductions scheduled to

coincide with economic

recovery. Put in place a credible

plan to stabilise the debt over

time, with debt (held by the

public) stabilising at around

65% of GDP in 2020, after

peaking at 72% in 2013.

Table 1:

The Simpson-Bowles proposals

TAX POLICYRevenues

Total revenue as a share of GDP

projected to rise from 15.7% of

GDP (2012) to 18.4% (2013)

and 20.3% (2015). Personal

income tax take to increase by

1.8% of GDP, social security

taxes by 0.5% of GDP and

corporate income taxes by

0.4% of GDP.

Income taxes

Scheduled to rise automatically

from 2013,reversing the 2001

tax cuts. Tax rates to rise from

10-15%, 25%, 28%, 33% and

35% to 15%, 28%, 31%, 36%

and 39.6% respectively

Payroll taxes

Temporary payroll tax cut of

2 percentage points is set

to lapse

Capital gains taxes

Scheduled to rise from 15% to

a maximum rate of 20% for

most taxpayers from 2013.

EXPENDITURE POLICYTotal

Total outlays projected to fall

from 22.9% of GDP (2012) to

22.4% (2013) and 21.5%

(2015). The Budget Control Act

has defence and non-defence

budgets falling by $55 billion

each year from 2013 to 2022

(0.7% of GDP in 2013).

Defence

$55 billion of cuts, almost

entirely discretionary spending,

amounting to 10% of

discretionary defence spending

in 2013. These cuts are not

restored in future years but, as

the economy and the size of

the defence budget grow, they

fall to 8.5% of the planned

discretionary defence budget

in 2022.

Medicare, Medicaid and

social security

Medicare is shielded from cuts:

90% of Medicare spending can

only be cut by a maximum of

2%; a further 9% is exempt

entirely. Medicare and social

security are exempt from cuts.

In January 2013, doctors’

payments under Medicare are

due to fall by 27%. These cuts

have been reversed by Congress

each year since 2003 (the ‘doc

fix’). Under current law, these

cuts to payments would reduce

expenditures by $10 billion.

Unemployment benefits

Extensions in emergency

unemployment benefit are set

to lapse. Total expenditure on

unemployment benefit is set to

fall by over a third from $94

billion to $60 billion in 2013,

despite a baseline CBO scenario

that has unemployment rising

over the course of the next year.

Table 2:

The fiscal cliff: changes in tax and expenditurespolicy scheduled in current law

legislation required automatic publicspending cuts – the Budget Control Act –to be triggered in 2013 if no agreement isreached by then. The spending cuts weredesigned to target the essential priorities ofboth parties to force the two sides to anagreement.

With the presidential electionapproaching, no attempt was made to findan alternative to the ‘sequester’ scheduledin the Budget Control Act. With no changein legislation now that the election haspassed, the sequester will be triggered inJanuary 2013. In addition to the spendingcuts in the Budget Control Act, tax cutsintroduced by the Bush administration in2001 will expire at the end of 2012.

The combination of these two factorsis known as the ‘fiscal cliff’ (see Table 2),which will mean that taxes rise by about2.7% of GDP and spending falls by almost1%. This fiscal contraction of close to3.7% in 2013 relative to current planswould almost certainly plunge the US intorecession, even on the most optimisticestimates.

To make things worse, the law doesnot address the root causes of the US debt problem – healthcare and pensions. If the economy falls over the fiscal cliff, itwill cut only discretionary spending, which is not the main cause of the long-run debt problem.

PUBLICATIONSCEP Discussion Papers are available as electronic copies free to download from the Centre’s website:http://cep.lse.ac.uk/_new/publications

MINIMUM WAGES AND WAGEINEQUALITY: SOME THEORY AND AN APPLICATION TO THE UKTim Butcher, Richard Dickens and Alan ManningCEP Discussion Paper No. 1177November 2012

ESTIMATING THE INFLUENCE OF LIFESATISFACTION AND POSITIVE AFFECT ON LATER INCOME USING SIBLING FIXED-EFFECTSJan-Emmanuel De Neve and Andrew Oswald CEP Discussion Paper No. 1176November 2012

THE ENDURING IMPACT OF CHILDHOODEXPERIENCE ON MENTAL HEALTH:EVIDENCE USING INSTRUMENTED CO-TWIN DATARachel Berner Shalem, FrancescaCornaglia and Jan-Emmanuel De Neve CEP Discussion Paper No. 1175November 2012

SELECTION EFFECTS WITHHETEROGENEOUS FIRMSMonika Mrázová and Peter NearyCEP Discussion Paper No. 1174October 2012

RESILIENCE TO ECONOMIC SHOCKS AND THE LONG REACH OF CHILDHOOD BULLYINGNattavudh PowdthaveeCEP Discussion Paper No. 1173October 2012

OPTIMAL POLICY FOR MACRO-FINANCIAL STABILITYGianluca Benigno, Huigang Chen,Chris Otrok, Alessandro Rebucci and Eric YoungCEP Discussion Paper No. 1172October 2012

THE CAREERS OF IMMIGRANTSAna Damas de MatosCEP Discussion Paper No. 1171October 2012

POLICY DESIGN IN A MODEL WITH SWINGS IN RISK APPETITEBianca De Paoli and Pawel ZabczykCEP Discussion Paper No. 1170October 2012

EXPORTERS, IMPORTERS AND CREDIT CONSTRAINTSMirabelle MuûlsCEP Discussion Paper No. 1169October 2012

CEP US ELECTION ANALYSESDuring the 2012 US presidential election, CEP published aseries of briefings to provide non-technical, evidence-basedand politically neutral introductions to the main economicissues facing the American people:� Taxes, spending and public debt� Economic recovery and policy uncertainty� Healthcare reform� Inequality and opportunity

The relevance of the analyses extends far beyond the presidential election: as the economically largest and most powerful nation on the planet, what happens in the United States has a profound influence on the rest of the world.

CEP’s US Election Analyses are available here:http://cep.lse.ac.uk/_new_publications/series.asp?prog=CEPUSA

For further information on CEP publications and events, please contact:

The Publications Unit, Centre for Economic Performance, Houghton Street, London WC2A 2AE

Telephone: +44 (0) 207 955 7673

Fax: +44 (0) 207 955 7595

Email: [email protected]

We accept Visa, Access and Mastercard for payment by post, fax and phone.

Overseas rates on requestFollow CEP on Twitter @CEP_LSE

CEP EVENTSCEP PUBLIC LECTURE

Eurozone Deadlock – Finding a Path Out of the Crisis

Speaker: Professor Luis GaricanoChair: Professor Francesco CaselliDate: Wednesday 23 January 2013Time: 6.30-8.30pm Venue: Sheikh Zayed Lecture Theatre, NewAcademic Building, LSE

It is still possible, both economically andpolitically, to find a way out of the eurozonecrisis if policy-makers separately address twoproblems: dealing with the legacy costs ofthe initially flawed design of the eurozone;and fixing the design itself. In this lecture,Luis Garicano will discuss how thesesolutions can be implemented.

Luis Garicano is professor of economics andstrategy in LSE’s management department, aresearch associate in CEP’s productivity andinnovation programme and co-editor ofNadaesGratis.es, the most widely readeconomics blog in Spanish.

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

© Centre for Economic Performance 2012

Volume 17 Issue 3

(ISSN 1362-3761) All rights reserved.

Centre Piece

The back issues

Follow CEP on Twitter @CEP_LSEAnd CEP’s director on @johnvanreenen

the magazine of the centre for economic...

Documents