Corporatism and job satisfaction

Journal of Comparative Economics 41 (2013) 35–47

Contents lists available at SciVerse ScienceDirect

Journal of Comparative Economics

journal homepage: www.elsevier .com/ locate/ jce

Corporatism and job satisfaction

0147-5967/$ - see front matter � 2013 Association for Comparative Economic Studies Published by Elsevier Inc. All rights reserved.http://dx.doi.org/10.1016/j.jce.2013.01.005

⇑ Corresponding author. Fax: + 1 212 854 3735.E-mail address: [email protected] (E. Phelps).

Edmund Phelps a,⇑, Gylfi Zoega b,c

a The Center on Capitalism and Society, 1126 International Affairs Building, Columbia University 420 W. 118th st., MC 3334 New York, NY 10027, USAb Department of Economics, University of Iceland, Saemundargata 2, 101 Reykjavik, Icelandc Department of Economics, Mathematics and Statistics, Birkbeck College, University of London, Malet St., London WC1E 7HX, UK

a r t i c l e i n f o a b s t r a c t

Article history:Available online 18 January 2013

JEL classification:P1P47P51

Keywords:CorporatismEconomic performanceJob satisfaction

Phelps, Edmund, and Zoega, Gylfi—Corporatism and job satisfaction

We introduce reported job satisfaction as a measure of economic performance and find itpositively correlated with GDP per capita and the labor force participation rate in a sampleof OECD countries and negatively correlated with unemployment. Moreover, we find thatmany measures of corporatism, which we define in the wider sense as institutions thathamper with the allocation of the factors of production and the distribution of income ina capitalist economy, are negatively correlated across countries with job satisfaction. Thusjob satisfaction is positively correlated across countries with measures of the protection ofproperty rights and negatively correlated with the volume of regulations of credit markets,labor markets and businesses, in addition to barriers to entrepreneurship, corruption andlack of access to capital. In contrast, measures of capitalism, such as the number of listedcompanies and market capitalization, are positively correlated with job satisfaction. Journalof Comparative Economics 41 (1) (2013) 35–47. The Center on Capitalism and Society, 1126International Affairs Building, Columbia University 420 W. 118th st., MC 3334 New York,NY 10027, USA; Department of Economics, University of Iceland, Saemundargata 2, 101Reykjavik, Iceland; Department of Economics, Mathematics and Statistics, Birkbeck Col-lege, University of London, Malet St., London WC1E 7HX, UK.� 2013 Association for Comparative Economic Studies Published by Elsevier Inc. All rights

reserved.

1. Introduction

In this paper we propose job satisfaction as a measure of economic performance and relate it to a variety of social insti-tutions that fall under corporatism and capitalism. We are interested in testing to what extent capitalist institutions mayenhance job satisfaction and economic performance while corporatist institutions that hamper with the allocation of re-sources and income may act to lower performance. We will test to see whether some signature features of corporatistand modern-capitalist economies – features such as the high employment protection in a country and the collective bargain-ing favored by corporatism; the large public sector; the bureaucratic red tape; and the individual freedoms relied on by cap-italism – are conducive or inimical to job satisfaction. Our thesis is that capitalist systems may provide a better opportunityfor individuals and firms with new ideas to replace existing ones, for new firms to leapfrog existing ones and for individualswith ideas to prosper while the corporatist system has a stifling effect by protecting vested interests in the form of jobs, busi-nesses and industries.

Our thesis that the institutions of corporatism may have a detrimental effect on economic performance is closely relatedto the contributions of Thrainn Eggertsson (such as Eggertsson, 2005), whose work we celebrate in this symposium. In his

http://dx.doi.org/10.1016/j.jce.2013.01.005

mailto:[email protected]


http://www.sciencedirect.com/science/journal/01475967

http://www.elsevier.com/locate/jce

36 E. Phelps, G. Zoega / Journal of Comparative Economics 41 (2013) 35–47

work, he describes how productivity depends not only on physical technology but also on social institutions that providepeople with incentives. The set of possible institutions in society is then determined by technology in the narrow senseand also social knowledge and social technologies, as well as factors such as geography and natural resources. In Eggertsson’swork, societies do not always end up choosing the best institutions from the set of feasible institutions. There are forces thatintervene and prevent societies from optimizing; forces such as the failure of collective action, the narrow self-interest ofruling groups, beliefs, values and cognitive limitations. In our case the belief in the beneficial effects of corporatist institu-tions and the vested interests of many social groups in maintaining this system may have affected overall job satisfaction insociety and economic performance.

The structure of this paper is as follows. We first discuss corporatism as an economic system, then review the literature onjob satisfaction and take a preliminary look at the statistical relationship between job satisfaction and the conventional mea-sures of economic performance; output per capita, labor force participation and unemployment, in a sample of 27 countries.We then introduce and discuss different types of corporatist institutions that are used in our analysis of the data. The dataanalysis starts with a series of scatter plots and tables that report statistical relationships between job satisfaction and var-ious institutional variables. We also estimate principal components in order to explore the relationship between job satis-faction and institutions further. In order to test for robustness we also use data on 47 thousand individuals where we cancontrol for individual attributes that may affect job satisfaction in addition to our institutional variables. The final sectionconcludes.

2. Corporatism as an economic system

Corporatism has a long history on the continent of Europe. In essence, the difference between corporatism and the liberalcapitalism that developed in the United States is the notion of individuals forming groups, such as labor, business and thescientific community, which represent their interests and together determine how society develops, in contrast to theuncontrolled development of the capitalist economy through the trials and errors of individual entrepreneurs. As in Phelps(2013) we define corporatism as the set of institutions and interventions in the functioning of a capitalist system that isintended to prevent capitalism from harming the objects of traditional values – state, family, community and religion. Ina corporatist system, competition in the market place is deemphasized and in its place comes a sense of shared objectivesfor society that are meant to ensure that each individual serves the interests of society. Individuals and businesses shouldthus have social responsibilities that transcend their private interests.

The recent labor economics literature has made much of corporatism. The exact definition of the term is often lose andvaries from one author to another. However, the term is generally used in a much narrower sense taken to mean that wagesare formed in labor markets where encompassing unions and employers’ associations represent the interests of their mem-bers. Corporatism then takes on a meaning close to ‘‘centralization of bargaining,’’ where the main implication is that themacroeconomic effects of the wage bargaining process are taken into account by the two parties to the bargain. Brunoand Sachs (1985) add to this definition the active involvement of the government in wage agreements taking the form ofincomes policy. Measures of corporatism are then used to explain differences across countries in labor market performance,such as real wage flexibility and the inflation–unemployment trade-off. While most studies, such as Crouch (1985) andBruno and Sachs, find a beneficial effect of measured corporatism, Calmfors and Driffill (1988) detect a hump-shapedrelationship between unemployment and centralization so that both countries with centralized wage bargaining and thosewith the most decentralized labor markets perform better than those in the middle. Tarantelli (1986) extends the definitionof corporatism to encompass the degree to which there is a high ideological and political consensus and also a high level ofintegration and cooperation of trade unions and employer’s representatives with the government. His extended index showsa high level of correlation with the ‘‘misery index,’’ defined as the sum of the rate of inflation and the unemployment rate, forthe OECD countries.

In this paper we widen the definition of corporatist institutions. Clearly, the corporatism found on the European continentreaches its influence beyond the labor market. It involves the existence of tripartite relationships between strong labor un-ions, employers’ associations and the government as social partners not only in reaching wage agreements but also in man-aging the direction of the national economy. Through this tripartite relationship the social partners affect both the allocationof the factors of production as well as the income distribution. Moreover, corporatism predates the modern use of the termby labor economists. In fact, it traces its roots to the late 19th century and the beginning of the 20th century. In the 1920s itstenets became popular among European nationalists who felt the need for a return to unity and purpose in society, the intel-lectuals who felt the need for economic order, the peasants, artisans and other interest groups losing ground to moderniza-tion who wanted protection; and the scientists who wanted state support for research and artists who wanted it for the arts.All these factions hoped for some way of curbing or over-riding various tendencies and impulses of the modern capitalisteconomy. This had to mean putting the private-enterprise economy under political control.

A corporatist model was built by Benito Mussolini in Italy in the1920s and 1930. Germany’s development of corporatismstarted sooner than Italy’s with one of the early corporatist critics of capitalism, Ferdinand Tönnies, who proposed the thesisin 1887 that communities and guilds were being destroyed by capitalism, and Emile Durkeim, who maintained that

E. Phelps, G. Zoega / Journal of Comparative Economics 41 (2013) 35–47 37

capitalism raised conflicts without rules1. In the 1920s German politics gradually gave voice to the elements of corporatistthought found in Italy, in particular a dislike of individualism and a rejection of laissez-faire economic policy. Adolf Hitler playeda pivotal role in implementing the corporatist ideas, much as Mussolini did in Italy. These ideas were subsequently introducedin countries such as Spain, Portugal and France and in South America in Brazil and Argentina.

In spite of the attempts by post-war Germany and other continental economies to shed some of the corporatist inheri-tance from the inter-war era in favor of a more competitive and liberal economy, as exemplified by the Treaty of Rome thatestablished competition at the forefront of the EU, important elements of the old regime remain. The organization of the la-bor market remains corporatist, dominated by large unions and employer associations. The tripartite system continues. Pub-lic employment has continued to grow, in particular in France, Italy and Germany; the share of government consumption inGDP has risen; and the volume of government regulations and red tape remain much greater in the corporatist economiesthan in the UK and the US. In France, Italy and Germany, as well as in the United Kingdom, labor unions rose to positions ofpolitical power during the post-war years. Unions and employers came to exert influence through non-market channels anda large increase in the activity of the public sector and the volume of regulations has resulted. Only in the UK was this devel-opment pushed back under the government of Margaret Thatcher, a development that was not reversed under hersuccessors.

Our previous research has implicated a few corporatist institutions in the failure of most European economies to grasp theopportunities of the internet revolution of the late 1990s – bureaucratic ‘‘red tape’’ and employment protection legislationwere among these – and the findings credited some institutions with helping some of the other OECD economies with seiz-ing the new opportunities – a relatively high proportion of the labor force with a university degree, for example (Phelps andZoega, 2001). More recently (Phelps and Zoega, 2004) we have been looking at specific institutions in the corporatist land-scape to see whether some of them are statistically speaking, harmful for economic performance. Among the institutionalvariables considered were the degree of employer- and union-coordination in the industrial bargaining process, weightedby the extent to which wages are ‘‘covered’’ by union scales; the penalty for employee dismissal provided in employmentprotection legislation; also, the volume of required licenses hindering or deterring the establishment of new firms andnew projects, as measured by the OECD index of ‘‘red tape’’. Our findings suggest that the effects of these institutions areharmful for market capitalization, which is a strong sign that they are harmful for economic performance.

3. Job satisfaction and economic performance

Job satisfaction has been found to be closely related to life satisfaction (see Judge and Watanabe (1993)). The economicsliterature on job satisfaction was started by Freeman (1978) and Borjas (1979). Freeman found job satisfaction to be a majordeterminant of labor market mobility and found it to be dependent on union membership. These results were consistentwith those of Borjas who also found that while union members were less satisfied, the magnitude of the effect was strongerat higher tenure levels. Phelps (1968) described the relationship between job satisfaction and quit behavior. Ingelhart (1990)uses the Eurobarometer surveys and reports data on well-being for a large number of countries. Andrew Clark andco-authors have studied job satisfaction in Britain, the characteristics of jobs most likely to make workers satisfied andthe relationship between age and job satisfaction.2 They find, using British data that men, workers in their thirties, those work-ing longer hours and workers in larger establishments have lower levels of job satisfaction. Sousa-Poza and Sousa-Poza (2000)analyzes the levels and determinants of job satisfaction in a cross-national setting for a group of 21 countries. Job satisfaction isquite high in most countries; highest in Denmark and very low in Japan and Russia; job satisfaction has fallen in Germany andthe United States; and job satisfaction tends to be positively correlated with having an interesting job, having good relationswith management and negatively correlated with having an exhausting job. Hamermesh (2001) detected increased dispersionof job satisfaction in the US between higher and lower income groups of young workers. He found that job satisfaction isespecially responsive to surprises in the returns to observable skills. Huppert and So (2009) use the European Social surveyand explore the determinants of what they label ‘‘flourishing,’’ and find that higher flourishing3 is associated with highereducation and income and that married people are more likely to flourish than those no longer married.

We will use survey responses from Ingelhart (1990) for 27 countries4 and relate these to observed institutional differencesusing a series of cross-section regressions. Table 1 below shows the numbers for mean job satisfaction in two waves of theWorld Values Survey (1990–1991 and 1999–2000); the average of the two that we use in the empirical work that follows;in additionto the unemployment and labor force participation rates in 1995 and the GDP per capita relative to the US levelin 2001. Note the high value for job satisfaction in Switzerland in the 1990 survey. Other countries that rank highly areDenmark, Austria, Canada, Norway, Sweden, the US and Iceland. The lowest ranking countries are Turkey, Greece, Spain, Franceand Slovakia. We also note that the job satisfaction numbers do not change much between waves although they do risesomewhat in Germany and the Czech Republic and fall in Sweden and Hungary.

1 See Durkheim (1997) and Tönnies (2001).2 See Clark (1996), Clark (1998), Clark et al. (1995) and Clark (1997).3 By flourishing they mean having positive emotions; engagement and interest, meaning and purpose, in addition to having self-esteem, optimism, resilience,

vitality, self-determination and positive relationships.4 Countries included were: Austria, Belgium, Canada, the Czech Republic, Denmark, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Japan,

Luxembourg, the Netherlands, Norway, Poland, Portugal, Slovakia, Slovenia, Spain, Sweden, Switzerland, Turkey, the UK and the US.

Table 1Measures of economic performance. Source: OECD.

Mean job sat. 1990–1991

Mean job sat. 1999–2000

Average jobsatisfaction

Unemployment1995

Participation1995

GDP per capita2001

Austria 8.0 7.7 7.9 3.7 48.5 80.9Belgium 7.7 7.6 7.7 13.0 42.6 79.5Canada 7.9 7.9 9.5 50.4 81.8Czeck Rep. 6.7 7.3 7 4.1 50.1 46.9Denmark 8.2 8.1 8.2 7.1 53.5 82.1Finland 7.6 7.7 7.7 15.5 49.1 74.0France 6.8 7.1 7 10.3 44.8 74.2Germany 7.2 7.7 7.5 8.2 48.2 74.5Greece 6.9 6.9 10.0 39.9 55.1Hungary 7.3 6.8 7.1 10.5 39.7 37.3Iceland 7.9 7.9 7.9 4.9 55.7 84.9Ireland 7.8 7.8 7.8 12.2 40.5 85.8Italy 7.3 7.3 7.3 11.7 40.3 76.1Japan 7.7 7.7 3.2 53.1 72.9Luxemburg 7.6 7.6 2.32 2.4 53.4 150.3Netherlands 7.5 7.5 7.5 7.2 47.9 85.8Norway 7.9 7.9 5.0 50.2 103.4Poland 7.8 6.6 7.2 13.3 45.0 30.5Portugal 7.6 7.6 7.2 47.4 51.6Slovakia 6.9 6.7 6.8 13.1 46.1 33.7Slovenia 7.2 7.2 7.2 51.4Spain 7.1 7.1 23.0 41.6 63.0Sweden 7.9 7.3 7.6 9.2 49.8 78.7Switzerland 8.4 8.4 3.2 58.6 89.5Turkey 6.1 6.1 7.6 38.1 24.0UK 7.4 7.3 7.4 8.6 48.3 76.9US 7.9 7.9 5.6 50.2 100.0


In Fig. 1, we show the relationship between job satisfaction, GDP per capita, labor force participation and unemploymentusing the numbers from Table 1. The four measures of economic performance are correlated in our sample so that job sat-isfaction is positively correlated with output and participation and negatively correlated with unemployment.

4. Corporatism and job satisfaction

We will classify corporatist institutions into three groups. These may influence the economic culture or, alternatively,work through institutions, such as expenditures and subsidies, entitlements and regulation. First, we have institutions thatintervene in resource allocation. Second, there are institutions that intervene in the distribution of income. Third, there are theinstitutions of neo-corporatism which are driven not by the state but by the attempts of the social partners – unions andemployers’ associations – and other powerful interest group to harness the state’s power to further their interests. Beforediscussing the three groups of corporatist institutions we briefly cover the corporatist institutions emphasized by laboreconomists.

4.1. Corporatism in labor economics

We start by examining briefly corporatism in the narrow definition used by labor economists. It should be noted thoughthat our interests lie primarily in the wider definition of corporatism that encapsulates the government’s attempts to redis-tribute both income and the factors of production in order to influence economic development.

In labor economics corporatism is used interchangeably with the centralization of wage bargaining. Many authors – suchas Crouch (1985), Bruno and Sachs (1985) and Layard et al. (2005) – have found centralized labor unions to be conducive tolow unemployment. Several reasons have been proposed for this finding. First, centralized unions have different objectivesfrom decentralized unions since there is no alternative employment available for the unemployment union members while adecentralized union can always count on unemployed members being able to find employment elsewhere. Second, whenhigher wages translate into higher prices, a centralized union realizes that a wage increase implies higher prices for all mem-bers while a decentralized union knows that a member’s higher wages are mainly another man’s higher prices. Third,employers may be more resistant to wage demands by a centralized labor union since they cannot pass on the wage increasein the form of higher relative prices.

We will discuss some of the empirical work on corporatism in the labor market below and relate it to measures of jobsatisfaction but before doing that we need to explain how the institutions of corporatism in the wider sense manifest them-selves in the economy.

20

40

60

80

100

120

140

160

6.0 6.5 7.0 7.5 8.0 8.5

Job satisfaction

GD

P pe

r ca

pita

rel

ativ

e to

US

leve

l

lux

tur

slk polhun

czrgre

slv

spa

fra it uknet

gerswebel

finjap

ireicecanaus

norusa

denswi

por

36

40

44

48

52

56

60

6.0 6.5 7.0 7.5 8.0 8.5

Job satisfaction

Part

icip

atio

n (%

)

tur

gre hun

polslk

fra

czr

ita

bel

ire

hunjap

ice

den

swi

uk gernet

aus

swefin aus

canusanor

spa

0

4

8

12

16

20

24

6.0 6.5 7.0 7.5 8.0 8.5

Job satisfaction

Une

mpl

oym

ent (

%)

tur

spa

slk

grefrahun

pol

ita

uk

fin

belire

can

den

swiczr

luxjap aus

icenorusa

sweger

netpor

Fig. 1. Job satisfaction and economic performance (1995). Abbreviations: Austria (aus), Belgium (bel), Canada (can), Check Republic (czr), Denmark (den),Finland (fin), France (fra), Germany (ger), Greece (gre), Hungary (hun), Iceland (ice), Ireland (ire), Italy (ita), Luxemburg (lux), Japan (jap), Netherlands (net),Norway (nor), Poland (pol), Portugal (por), Slovakia (slk), Slovenia (slv), Spain (spa), Sweden (swe), Switzerland (swi), United Kingdom (uk), United States(usa) and Turkey (tur).


4.2. Institutions affecting resource allocation

At the heart of corporatism is an intervention in what the economy produces. This may come about through the level andsystem of taxation, an extensive body of regulation of businesses, barriers to entry into selected industries and industrialpolicy.


One measure of the effect of government on the allocation of resources is the size of tax revenues. Since corporatists careabout what is produced they may use the tax system to affect the direction of the private sector. Consumption not deemeddesirable by the state is heavily taxed as well as most income in order to finance subsidies of employment in the privatesector, transfers or public employment. Other examples include regional subsidies and subsidies to selected industries suchas high-technology industries.

Another measure is the volume of recorded regulations since corporatists may also intervene by regulating private indus-try. Corporatist economies go beyond the body of good regulation to a large body of bad regulation, beyond what is needed togrease the wheels of commerce and innovation by providing protection against fraud to investors, workers and savers fromunscrupulous companies, employers and banks. Corporatist economies use barriers to entry to protect important industries,professions and communities in various ways. New industries may be prevented from growing in order to protect existingones. Extensive red tape is a manifestation of attempts to establish controls over the formation and operation of businessesin general. There may be permits for the entry and operation of businesses, permits for different transactions and licenses forthe entry of people into different professions. Informal regulation may take the form of the length of time it takes to enforcecontracts and tolerance of judicial delays. An aim of red tape may be to prevent one interest group from harming another butthe net effect may be to stifle efficiency and innovation.

4.3. Institutions affecting income distribution

Corporatist doctrine also puts emphasis on who should benefit in society rather than just what should be produced.Retirement pension, unemployment benefits and subsidized health care are just few examples of the state using its powersto redistribute income.

The size of labor unions and their centralization matters for the income distribution. The labor market pits large aggre-gations of workers represented by unions against large aggregations of companies, represented by an employer association.What results is what economists tend to call the battle of the markups: unions want to raise money wages for a given level ofprices and employers want to raise prices for a given level of money wages. In equilibrium we need unemployment to besufficiently high to bring union’s demands in terms of real wages down to what employers are willing to offer. Governmentsfrequently intervene in this process through incomes policies and tri-partite agreements.

The valuation of businesses in stock markets is depressed by the corporatist state. The state taxes businesses and empow-ers workers to acquire a larger share of company profits. Labor unions threatening strikes are one clear example of this.Moreover, attempts to redistribute output between industries trough regulations and red tape that is intended to protectone group from another will lower the valuation of businesses.

4.4. Neo-corporatist institutions

Neo-corporatism distinguishes itself from classic corporatism in not having the state taking the initiative in setting thedirection of the economy but instead in having the initiative taken by powerful business interests. This type of corporatismhas several features.

A weak corporatist state has needs of powerful friends or cronies. The state may therefore protect particular industriesand businesses, which may smell of cronyism. Industrial policy is aimed at the favored industries, not the ones that may turnout to be most profitable or important for the national economy. Government contracts may for the same reason be given tofriendly companies to buy their support for the government in power. In some countries the cronies are primarily relatives orlong-time friends of the rulers.

Backdoor deals between politicians and their trusted cronies take the form of lobbying of legislators, regulators and agen-cies by businesses for legislation or rule-setting. This may over time generate volumes of laws, regulations and interpretativerulings aimed at giving various preferences to all or most groups and individuals. Businesses may also resort to political con-tributions, in the extreme bribing and other activity to reduce the burden brought about by taxation and regulation dictatedby government officials and excessive wage demands by labor unions. Paying politicians and political parties for favors isanother symptom of the same phenomenon. The proportion of the labor force engaged in the practice of the law may bea good proxy for the proportion of income diversion from those who earned it to those receiving various sorts ofcompensation.

5. Exploring the data

In order to provide a glance at the relationship between corporatist economic systems and job satisfaction we use data onjob satisfaction taken from the World Values Survey for the 27 countries listed above and the measures of institutions takenfrom the Fraser Institute5 (observations for 1995) and other sources.

Before moving onto our main results we first consider the measures of corporatism used in labor economics. In theirempirical work, Layard et al. (2005) find in a cross section of 20 OECD countries using data from the 1980s that the average

5 See http://www.fraserinstitute.org/.

http://www.fraserinstitute.org/

Table 2Institutions and job satisfaction.

Dependent variable: job satisfaction Constant term Independent variable

Independent variable Estimate t-Ratio Estimate t-Ratio R2 Obs.

Resource allocationSize of government a,b 7.0 28.5 0.13 2.4 0.44 27

Government consumption a 7.9 23.5 �0.09 1.2 0.09 27Government investment a 7.3 53.2 0.04 1.6 0.08 25Top marginal tax a 7.5 56.6 �0.01 0.1 0.00 25Transfers and subsidies a 7.5 30.9 0.00 0.0 0.00 26

Business regulationsa,b 6.1 17.2 0.12 2.7 0.43 26Time with bureaucracy a 5.7 16.3 0.26 5.5 0.43 26Starting a new businessa,b 6.9 25.4 0.12 2.9 0.47 26Irregular payments a 6.5 28.2 0.15 5.0 0.54 26

Credit market regulations a,b 6.1 13.5 0.21 4.4 0.38 26Bank ownershipa,b 7.4 32.5 0.02 0.8 0.34 27Bank competitiona,b 6.9 19.0 0.10 2.1 0.41 26Credit extensiona,b 6.7 20.4 0.11 2.2 0.46 27Int. rate controlsa,b 6.0 13.1 0.22 3.6 0.61 26

OthersBarriers to entrep.ship 8.6 22.3 �0.45 2.6 0.30 24

Income diversionLegal structurea 4.7 13.4 0.34 7.8 0.68 27

Impartiality of courtsa 6.0 20.2 0.20 5.5 0.40 27Judiciary independencea 5.8 13.8 0.22 4.4 0.26 26Military in politicsa 4.9 19.9 0.29 10.2 0.63 27Protection of property rightsa,b 6.1 21.0 0.22 4.8 0.48 25

Impediments to tradea 6.0 7.0 0.19 1.8 0.09 27Tariffsa 7.7 9.0 �0.02 �0.2 0.00 27Hidden import barriera 6.8 12.3 0.09 1.4 0.07 26Regulatory trade restrictionsa 6.8 12.3 0.09 1.4 0.07 26Actual versus expected size tradea 7.7 31.0 �0.05 �1.0 0.03 27Capital controls a 6.0 16.1 0.19 4.3 0.44 27

Regulations a 5.6 13.7 0.29 4.6 0.35 26Labor market regulations b,a 7.0 29.5 0.12 2.7 0.43 26

Minimum wagea,b 7.5 28.8 0.01 0.2 0.32 26Hiring and firing restrictionsa,b 7.0 30.0 0.10 2.1 0.43 26Collective bargaininga,b 6.8 21.8 0.13 2.5 0.40 26Unemployment insurancea 7.4 35.7 0.03 0.6 0.33 26

Milken index (rank) 7.9 62.7 �0.02 4.5 0.54 25Number of listed companies 7.2 57.4 0.02 5.0 0.56 15Market capitalization 7.4 54.8 0.36 2.4 0.27 15Neo-corporatist institutions

Number of lawyers 7.6 56.8 �12.75 0.2 0.00 26Corruption (Heritage Foundation)a 6.2 24.3 0.18 5.5 0.60 27

a A higher number denotes more economic freedom. t-Statistics written in bold letters indicate significance at the 5% level.b Dummy variable for Turkey included.


rate of unemployment is positively correlated with the degree of coverage of collective bargaining and negatively correlatedwith the degree of union co-ordination and employer co-ordination. This is a typical result in the labor economics literature,unions are bad for employment but centralized bargaining reduces the damage. These authors find that the effect of unem-ployment on wage setting is greater – there is greater real wage flexibility – in countries with centralized bargaining or whatthey call corporatism. Calmfors and Driffill (1988) detect a hump-shaped relationship between unemployment and central-ization so that both countries with centralized wage bargaining and those with the most decentralized labor markets per-form better than those in the middle. We find a negative relationship between job satisfaction and the extent of collectivebargaining in our data, see Table 2 above. However, we do not detect any relationship between measures of the centraliza-tion of bargaining and job satisfaction, see Fig. A1 in Appendix. However, this should not come as a surprise since no obviousrelationship should be between the centralization of collective bargaining and the extent to which firms provide opportu-nities for entrepreneurship, challenges and personal growth.

We now move on and take a look at our four aggregate measures of the institutional setup; the size of government, a mea-sure of the quality of the legal framework (measuring the protection of property rights and the enforcement of contracts), thefreedom to trade internationally and the volume of regulations and relate these measures to reported job satisfaction in Fig. 2below. Each of the four aggregate measures is numbered from 0 to 10 where ten denotes maximum freedom. Thus the higherthe measure the smaller is the size of the government, the better is the protection of property rights, the better is the qualityof the judicial system, the greater is the freedom to trade internationally and the less restrictive is the regulatory framework.

All four measures affect both resource allocation and the distribution of income to a varying extent. Of the four, the size ofgovernment may be an indication of resource allocation in the corporatist fashion, as is market regulation. The measure of

6.0

6.5

7.0

7.5

8.0

8.5

1 2 3 4 5 6 7 8

tur

swi

us

den

ausnor

finswe

sln

slk

pol

czr fra

ita

hun spa

bel

ger

jappor

uk

ireice can

gre

netlux

Freedom from government

Jobsatisfaction

6.0

6.5

7.0

7.5

8.0

8.5

4 5 6 7 8 9 10

Legal structure: Right to contract

Jobsatisfaction

tur

slkgre

czr

itapol

spahunfra

belporjap

ice auscannorusire

finswe luxnet ger

uk

sln

den

swi

6.0

6.5

7.0

7.5

8.0

8.5

5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0

jap

lux

pol

greslk

tur

slnhun

fra

swi

can

den

norus aus

czrspa

ita

porfin

geruk

irebel

sweluxnet

Jobsatisfaction

Freedom to trade internationally

6.0

6.5

7.0

7.5

8.0

8.5

4 5 6 7 8 9

Paucity of regulations

Jobsatisfaction

tur

pol

slk

ita

greczr

spa hunfra

sln

porger

swe

aus

beljap finluxnet

norice

swi

den

canire

us

uk

Job satisfaction and freedom from government Job satisfaction and legal structure

Job satisfaction and the freedom to trade Job satisfaction and the paucity of regulations

Fig. 2. Job satisfaction and aggregate measures of institutions.


the quality of the legal framework can clearly be taken to capture institutions that have to do with income diversion whileimpediments to free trade affect both resource allocation and the distribution of income.

A clear positive relationship appears between the quality of the legal framework and job satisfaction, the paucity of reg-ulation and job satisfaction as well to a lesser extent between the size (that is smallness) of government and job satisfaction.In contrast, job satisfaction exhibits no clear relationship with the freedom to trade. The relationships between the variablesare documented in Table 2 (as measured by estimated coefficients and the R-squared). Note that differences in the size ofgovernment explain 44% of the cross-country variation in the mean level of job satisfaction; differences in legal structure(i.e. protection of property rights) can explain even more or 68% of the variation; and differences in the volume of regulationof business explain 35% of the variation. In all three cases we have a positive relationship such that a smaller government,better legal structure and less restrictive business regulation may contribute to greater job satisfaction.

Of the four measures, the regulation of business is the clearest measure of corporatist influences. The quality of the legalstructure also has the flavor of measuring corporatism. Taken together, a small government, efficient legal structure and lowlevels of regulation are characteristic of capitalist economies, which according to Table 2 and Fig. 2 may promote jobsatisfaction.

In an appendix we show the subcomponents of the four measures: the size of government, the legal structure, impedi-ments to trade and the volume of regulation. Fig. A2 shows the four subcomponents of the size of government: governmentconsumption, government investment, the top marginal income tax rate and transfers and subsidies. We note a rather weakrelationship in the case of the subcomponents of the size of government variable. Turning to the legal structure in Fig. A3 inthe appendix we find much stronger relationships. There is a clear upward sloping relationship between job satisfaction anda measure of the impartiality of courts; judiciary independence; a measure of the involvement of the military in politics anda measure of the protection of intellectual property rights.6

6 The correlation between job satisfaction and the involvement of the military in politics is mainly due to the inclusion of Turkey which has both low jobsatisfaction and a high degree of involvement of the military in politics.


Turning to impediments to international trade in Fig. A4 we find that a positive relationship between job satisfaction and ameasure of capital controls (absence of) and also, although a much weaker one, between job satisfaction and the paucity ofregulatory trade barriers. Finally, Fig. A5 shows the relationship between job satisfaction and the subcomponents of theaggregate measure of regulation. A positive relationship is visible for all three subcomponents; the paucity of regulationof credit market, regulation of labor markets and regulation of businesses.

We have also considered six additional variables that measure interference in resource allocation. The first three measureaccess to capital: market capitalization, the number of listed companies (both variables measured in 1990) and the MilkenInstitute Capital Access Index7 (values from 2004). The fourth measures barriers to entrepreneurship (1998 values). In addition,we use measures of corruption from the Fraser Institute8 (2000 values) and the number of lawyers per capita (2006 values) tomeasure neo-corporatist institutions. The scatter plots between each of these six variables and job satisfaction are shown inFig. 3. There is a clear positive relationship between job satisfaction, on the one hand, and the number of listed companies, mar-ket capitalization, access to capital (a higher rank implies less access explaining the negative slope of the relationship in thefigure) and the Fraser Institute index of corruption (implying that job satisfaction goes together with less corruption) and a neg-ative relationship between job satisfaction and barriers to entrepreneurship.

In Table 2, which shows the statistical correlations, there appears to be a surprisingly large number of significant corre-lations between job satisfaction and various institutional measures. Job satisfaction is negatively related to various measuresof business regulation (a positive coefficient implies that less regulation and greater job satisfaction go together) – time withbureaucracy, starting a new business, irregular payments (corruption) – as well as with some credit market regulations –regulation on credit extension and interest rate controls. Regulations on hiring and firing labor are negatively related tojob satisfaction (a positive coefficient), as are barriers to entrepreneurship. The legal structure is very significant; job satis-faction is positively correlated with the impartiality of courts, the judiciary independence and the (absence of) involvementof the military in politics and the protection of property rights, all with the expected sign. However, we did not find any rela-tionship between job satisfaction and measures of self employment, social expenditures or tax revenues as a share of GDP.We omit these variables from Table 2.

The various impediments to trade – tariffs, hidden import barriers, regulatory trade restrictions and actual versus expectedsize of tread – do not appear to be much related to job satisfaction. However, capital controls are inversely related to job sat-isfaction. The extent of collective bargaining in the labor market – the narrow type of corporatism emphasized by labor econ-omists – affects job satisfaction adversely, as do restrictions on hiring and firing. However, the level of the minimum wage, aswell as the level of unemployment benefits, is not correlated with job satisfaction. The various measures of access to capital andthe development of the capital market are highly significant: the greater the number of listed companies, the higher is marketcapitalization and the greater is the access to capital, the higher is reported job satisfaction. Finally, corruption – both as mea-sured by the Fraser Institute as ‘‘irregular payments’’ and the Heritage Foundation as ‘‘freedom from corruption’’ – is negativelyrelated to job satisfaction while the number of lawyers, when entered alone, is not significantly related to job satisfaction.

We also need to explore whether our institutional variables may only be capturing the relationship between job satisfac-tion and income per capita since jobs may be more interesting in more advanced societies as the number of people havingbad jobs decreases. For this reason, the effect of institutions in Table 2 may conceivably be exaggerated. However, we shouldnote that in our thesis the institutions affect economic performance in a broad sense; output per capita, unemployment, la-bor force participation and job satisfaction. Thus finding that job satisfaction is only statistically related to income per capitaand not at all to institutions would go against our thesis while a finding that job satisfaction is correlated with both incomeper capita and institutions would provide further support for the thesis. In order to control for this possibility we took dataon GDP per capita9 for 1995 from the Penn-World table and re-estimated the coefficients in Table 2 by always including GDPper capita alongside each of the institutional variables. The inclusion of output per capita did not affect the sign or statisticalsignificance of the large majority of institutional variables in the table, which tells us that the correlations between job satis-faction and the institutional variables are not caused by institutions being correlated with output per capita and higher outputbeing correlated with greater job satisfaction.10 Interestingly, when the number of lawyers per capita is included alongside realGDP per capita the coefficient of lawyers becomes negative and statistically significant from zero, implying that more lawyersand less job satisfaction go together.

6. The data summarized

The small sample size and the large number of potential explanatory variables for job satisfaction make multiple regres-sions difficult to implement. For this reason we calculate principal components in order to summarize the data. In order to

7 http://www.milkeninstitute.org/. The CAI measures the breadth, depth and vitality of capital markets.8 This variable is derived primarily from the Transparency International’s Corruption Perceptions Index (CPI).9 PPP converted GDP per capita, 2005 international dollars per person.

10 The following institutional variables retained their sign and statistical significance: business regulations, time with bureaucracy, starting a new business,irregular payments, credit extension, interest rate controls, legal structure, impartiality of courts, judiciary independence, military in politics, protection ofproperty rights, tariffs, capital controls, regulations, labor market regulations, hiring and firing restrictions, collective bargaining, the Milken index, marketcapitalization and the Heritage index of corruption. The following variables became less significant: size of government, credit market regulations and barriersto entrepreneurship. In contrast, the number of lawyers became statistically significant when controlling for output per capita, so that a higher number oflawyers is correlated with lower job satisfaction.

http://www.milkeninstitute.org/

6.0

6.5

7.0

7.5

8.0

8.5

0 10 20 30 40 50 60 70

ita

Number of listed companies

Job

satis

fact

ion

Job

satis

fact

ion

Job

satis

fact

ion

spafra

geruk

netjapswebel

aus usa can nor

den

swi

6.0

6.5

7.0

7.5

8.0

8.5

0.0 0.4 0.8 1.2 1.6 2.0

Market capitalization (% of GDP)

fra

ita

spa

ger netuk

swebel

aus nor usa

den

can

jap

swi

6.0

6.5

7.0

7.5

8.0

8.5

0 10 20 30 40 50 60 70

Milken index (rank)

uk

swefin ire

usa cannor

den

swi

aus

jap belpor

net ger

spafra

ita

hunczr

gre

pol

skl

skv

tur6.0

6.5

7.0

7.5

8.0

8.5

1.0 1.5 2.0 2.5 3.0 3.5 4.0

Barriers to entrepreneurship (OECD)

swi

den

canire

uk

hunczr

nor usaaus ice

jap

pol

fragre

spa

ita

gernetswe por

bel fin

tur

6.0

6.5

7.0

7.5

8.0

8.5

2 3 4 5 6 7 8 9 10 11

Absence of corruption

itapol

hunczr

greslk

tur

spafra

jap belpor lux

ger netuk

aus usa cannor icenet

swefin

den

swi

6.0

6.5

7.0

7.5

8.0

8.5

.000 .001 .002 .003 .004

Lawyers per capita

swi

den

ausire

japfin belswe

pol slkhun

fra czr

slk

nor

porgernet

can ice usa

lux

ukita

spa

gre

Job satisfaction and the number of listed companies Job satisfaction and market capitalisation

Job satisfaction and access to capital Job satisfaction and barriers to entrepreneurship

Job satisfaction and the absence of corrruption Job satisfaction and number of lawyers

Job

satis

fact

ion

Job

satis

fact

ion

Job

satis

fact

ion

Fig. 3. Job satisfaction and institutions continued.


capture the information contained in Table 2 we can calculate the principal components of the 27 (country) by 39 (variables)matrix and relate them to our measure of job satisfaction. In Table 3, we start by showing the eigenvalues for a matrix thatalso includes our measures of economic performance: job satisfaction, productivity, and unemployment and labor force par-ticipation in addition to the 39 institutional variables. The first two principal components each explain around 1/3 of thevariance in the matrices and the third one 12% of the variation.

The first principal component for institution and performance combined corresponds to the eigenvector in the first col-umn in Table A1 in Appendix. The values of the eigenvector are positive for job satisfaction, productivity and participationand negative for unemployment. The positive value for job satisfaction goes with a positive value for various institutionalmeasures – indicating more economic freedom so that a positive value implies lower spending, taxes and so forth – for

Table 3Eigenvalues for institutions and economic performance.

Number Value Proportion Cumulative proportion

1 12.18 0.36 0.362 8.07 0.24 0.603 3.99 0.12 0.714 2.23 0.07 0.785 1.95 0.06 0.846 1.39 0.04 0.88

6.0

6.5

7.0

7.5

8.0

8.5

-8 -6 -4 -2 0 2 4 6

First principal component of matrix of institutions

Job

satis

fact

ion

pol

gre

ita

jap

hunfraspa

porbel

aus canusa nor

den

irefin

gerswe

netuk

Fig. 4. Job satisfaction and an index of corporatist institutions.


government investment, time with bureaucracy, time required to start a business, irregular payments (corruption), the pri-vate ownership of banks, competition between banks and interest rate controls. Similarly, variables of economic freedomthat measure hiring- and firing-restriction, the impartiality of courts, judicial independence, the absence of the military inpolitics, the protection of intellectual property rights, hidden import barriers, regulatory trade restrictions, the absence ofcapital controls, the number of listed companies and the absence of corruption as measured by the Heritage Foundation ap-pear with a positive sign in the eigenvector – again implying less regulation and capital controls and so forth – while barriersto entrepreneurship and the country ranking in the Milken index of access to capital (indicating less access to capital) appearwith a negative sign.

Table A2 has the eigenvalues and eigenvectors for the same matrix once the four measures of economic performance – jobsatisfaction, participation, unemployment and productivity – have been omitted and, due to missing observations for theMilken index (25 observations) and the number of listed companies and market capitalization (15 observations), these threevariables are also omitted. In Fig. 4 we use the first principal component from the matrix of institutions (performance vari-ables and capital market excluded) as a proxy for corporatist institutions and plot it against our job satisfaction variable. Aclear positive relationship appears so that greater corporatism goes with lower job satisfaction. The most corporatist countryby this measure is Poland, then Greece, followed by Italy, Japan, Japan, Hungary, France, Spain and Portugal. The least cor-poratist economies are those of Denmark, Norway, Finland, the US, Ireland, the Netherlands, the UK, Canada and Sweden. Theranking of Denmark, Norway and Sweden may come as a surprise but note that the institutional variables reported in Table 2are not confined to the labor market – the centralization of the labor market in these countries takes place in economies thatare not as corporatist in other spheres as many other European countries.

7. Behind the aggregates

In the introductory chapter we noted a set of stylized facts from the study of job satisfaction using microeconomic data.There is some indication that job satisfaction has been falling over time in some countries; job satisfaction is higher amongwomen, the self-employed, the young and the old, supervisors and those with secure jobs; it is smaller for union membersand greater for the well paid workers. In this section we explore the robustness of our earlier results by correcting for dif-ferences between countries in the structure of their populations in light of the stylized facts. We use 47,418 observationsfrom the World Values Survey; taken from the 1980–1981 survey, the 1990–1991 survey and the 1999–2000 survey. We

Table 4Job satisfaction and individual attributes.

Constant term 6.93 (59.0) Income deciles 0.07 (16.6)Sex 0.02 (0.7) Union membership �0.11 (4.5)Age 0.01 (1.6) Time, 1990–91 0.03 (0.9)Age-squared 0.0001 (2.3) Time, 1999–2000 �0.05 (1.5)Self-employed 0.32 (8.9)Observations 37,810, R-squared = 0.07

t-Statistics in parentheses. The variable sex takes the value 1 for men and 2 for women, thevariable self-employed takes the value 0 if not self-employed and 1 if self-employment andsimilarly for union membership.

6.0

6.5

7.0

7.5

8.0

8.5

-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0

Country fixed effects

Job

satis

fact

ion

tur

slkgrefra czrspa hun

sln polita

uknet

sweger

porluxfin jap bel

ireusaice

auscannor

den

swi

-2.0

-1.6

-1.2

-0.8

-0.4

0.0

0.4

0.8

-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0

Country fixed effects

Adj

uste

d co

untr

y fi

xed

effe

cts

ausbel

can

den

fin

fra

ger

gre

ire

ita

japnet

nor

por

spa

swe

uk

usa

czrhun

ice

lux

polslk

slv

swi

tur

Fig. 5. Job satisfaction and country fixed effects.


regress reported job satisfaction on individual attributes and country and time dummies – one for the 1990–1991 wave andanother for the 1999–2000 wave of the WWS – and show the results in Table 4.

The signs of the estimated coefficients are mostly in line with the stylized facts; women and the self-employed havegreater job satisfaction, union members have lower job satisfaction and higher income brings greater satisfaction. Of these,self employment, income and union membership have statistically significant coefficients. However the U-shaped age pat-tern does not arise in these data, instead satisfaction is estimated to increase with age. The point estimates for the time dum-mies indicate declining job satisfaction over time although the coefficients are not significantly different from zero.

Fig. 5 above shows the relationship between the job satisfaction variable used in previous sections of this paper (averageof mean 1990–1991 and 1999–2000 values taken from the WWS) and the country fixed effects calculated by not correctingfor individual characteristics, on the one hand, and the relationship between the unadjusted and adjusted (for individualcharacteristics) country fixed effect, on the other hand. The left-hand panel shows that the estimated country fixed effectsare closely correlated with the average values for the two waves 1990–1991 and 1999–2000 reported by WWS and used inthe cross sections in previous sections. The correlation is 0.95. There is also a close relationship in the right-hand panel be-tween the unadjusted country fixed effects and the country fixed effects found when controlling for the individual charac-teristics in Table 4. The correlation between the two is 0.93.11

We conclude that using mean values for job satisfaction is justifiable in the light of the finding that this would not besignificantly altered once account is taken of various individual attributes.

8. Conclusions

Reported job satisfaction is correlated with observable measures of economic performance, such as labor market partic-ipation, unemployment and output per capita across the OECD countries. Job satisfaction may therefore not only be impor-tant for life satisfaction but also for employment and economic output.

Many institutions of corporatism tend to hamper job satisfaction. Job satisfaction is positively correlated with the FraserInstitute measure of the legal structure (protection of property rights) and negatively with the volume of regulation of creditmarkets, labor markets and businesses. These regulations go beyond what would be needed to facilitate the smooth work-ings of the market economy, such as macro- and micro-prudential regulations in credit markets and the regulation of food

11 The only outlier is Poland that has a higher adjusted job satisfaction (�0.41) than unadjusted (�0.99).


production in goods markets. Moreover, job satisfaction is negatively related to barriers to entrepreneurship, corruption andwith limited access to capital. The number of listed companies and market capitalization are positively correlated with jobsatisfaction.

Overall, we conclude that many of the institutions we have labeled as corporatist may hamper economic performance andjob satisfaction. Societies may have, to use the terminology of Thrainn Eggertsson, picked a set of institutions from the set offeasible institutions that fail to maximize economic performance.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jce.2013.01.005.

References

Borjas, George.J., 1979. Job satisfaction, wages and unions. The Journal of Human Resources 14 (1), 21–40.Bruno, Michael., Sachs, Jeffrey., 1985. The Economics of Worldwide Stagflation. Basil Blackwell, Oxford.Calmfors, Lars, Driffill, John, 1988. Bargaining structure, corporatism and macroeconomic performance. Economic Policy 3 (6), 13–61.Clark, Andrew.E., 1996. Job satisfaction in Britain. British Journal of Industrial Relations 34, 189–217.Clark, Andrew.E., 1997. Job Satisfaction and Gender: Why are Women so Happy at Work? Labour Economics 4 (4), 341–372.Clark, Andrew E., 1998. What Makes a Good Job? Evidence from OECD Countries. Discussion Paper. LEO-CRESEP, University of Orleans (December).Clark, Andrew E., Oswald, Andrew J., Warr, Peter B., 1995. Is job satisfaction U-shaped in age? Journal of Occupational and Organizational Psychology 69, 57–

81.Crouch, Colin, 1985. Conditions for trade union wage restraint. In: Lindberg, L.N., Maier, C.S. (Eds.), The Politics of Inflation and Economic Stagnation. The

Brookings Institution, Washington, DC.Durkheim, Emile, 1997. The Division of Labor in Society. Free Press, New York.Eggertsson, Thrainn, 2005. Imperfect Institutions: Possibilities and Limits of Reform. Cambridge University Press, New York.Freeman, Richard B., 1978. Job satisfaction as an economic variable. The American Economic Review 68 (2), 135–141.Hamermesh, Daniel.S., 2001. The changing distribution of job satisfaction. Journal of Human Resources 36 (1), 1–30.Huppert, Felicia A., So, Timothy T.C., 2009. What Percentage of People in Europe are Flourishing and What Characterises Them?. The Well-being Institute,

University of Cambridge.Ingelhart, Ronald, 1990. World Value Surveys and European Values Surveys, 1981–1984, 1990–1993, and 1995–1997. Inter-University for Political and

Social Research.Judge, Timothy A., Watanabe, Shinichiro, 1993. Another look at the job satisfaction-life satisfaction relationship. Journal of Applied Psychology 78 (6), 939–

948.Layard, Richard, Nickell, Stephen, Jackman, Richard, 2005. Unemployment: Macroeconomic Performance and the Labour Market, second ed. Oxford

University Press, Oxford.Phelps, Edmund S., 1968. Money-wage dynamics and labor-market equilibrium. Journal of Political Economy 76 (4), 678–711.Phelps, Edmund S., 2013. Mass Flourishing: When Modern Values: When Modern Values Sparked Grassroots Dynamism. Princeton University Press.Phelps, Edmund S., Zoega, Gylfi, 2001. Structural booms. Economic Policy 16 (32), 83–126.Phelps, Edmund S., Zoega, Gylfi, 2004. The European labour markets – the search for routes to better economic performance in continental Europe. CESifo

Forum 5 (1), 3–11.Sousa-Poza, Alfonso, Sousa-Poza, Andés A., 2000. Well-being at work: a cross-national analysis of the levels and determinants of job satisfaction. Journal of

Socio-economics 29, 517–538.Tarantelli, Ezio, 1986. The regulation of inflation and unemployment. Industrical Relations: A Journal of economy and Society 25 (1), 1–15.Tönnies, Ferdinand, 2001. Community and Civil Society, Cambridge University Press.



Documents

Corporatism and job satisfaction