Munich Personal RePEc Archive
The redistributive impact of public and private social expenditure
Goudswaard, Kees and Caminada, Koen
Department of Economics, Leiden University
2008
Online at https://mpra.ub.uni-muenchen.de/20178/
MPRA Paper No. 20178, posted 27 Jan 2010 17:00 UTC
Ke e s Goudsw a a r d
Leiden Law School
Econom ics Depart m ent , Leiden Universit y P.O. Box 9520, 2300 RA, The Net her lands Phone: + + 31 71 527 7756
URL: w w w .hsz.leidenuniv .nl
E- m ail: k .p.goudswaard@law .leidenuniv .nl
Koe n Ca m ina da
Leiden Law School
Econom ics Depart m ent , Leiden Universit y P.O. Box 9520, 2300 RA, The Net her lands Phone: + + 31 71 527 7756
URL: w w w .hsz.leidenuniv .nl
E- m ail: c.l.j .cam inada@law.leidenuniv.nl
Abst r a ct
Most analyses of social prot ect ion are focussed on public arrangem ent s. However, social effort is not rest rict ed t o t he public dom ain; all kinds of privat e arrangem ent s can be subst it ut es t o public program s. I n fact , in several count ries t here has been a shift from public t owards privat e social arrangem ent s. OECD- dat a indicat e t hat account ing for privat e social benefit s has an equalising effect on levels of social effort across a num ber of count ries. This suggest s t hat public and privat e social expendit ures are com plem ent ary t o som e ext ent . But t heir dist ribut ional effect s differ. I n all OECD count ries, t he social prot ect ion syst em causes a m ore equal dist ribut ion of incom es. I ndeed, using cross- count ry dat a, we find a negat ive relat ionship bet ween public social expendit ures and incom e inequalit y and a posit ive relat ionship bet ween public social expendit ure and incom e redist ribut ion. But we do not find a significant posit iv e relat ionship bet w een privat e social expendit ures and incom e inequalit y or incom e redist ribut ion. Consequent ly, changes in t he public/ privat e- m ix in t he provision of social prot ect ion m ay affect t he redist ribut ive im pact of t he welfare st at e.
JEL- cla ssifica t ion: D3, H22, and H55
Ke yw or ds: Social Prot ect ion, Privat e Social Expendit ure, I ncom e Dist ribut ion
* The paper elaborat es on a lect ure by Goudsw aard at t he Univer sit y of Fribourg, Sw it zer land, May 13t h 2008.
This st udy is part of t he research program ‘Reform ing Social Secur it y’: w w w .hsz.leidenuniv.nl. Financial support of St icht ing I nst it uut GAK is grat efully acknow ledged.
Th e r e dist r ibu t ive im pa ct of pu blic a n d pr iva t e
socia l e x pe n dit u r e ∗
1 I nt r oduct ion
The welfare st at e aim s t o reduce incom e inequalit y bet ween individuals. People differ in t heir t alent , socio- econom ic background and opport unit ies. The m arket process t hen result s in large differences in incom e levels. Governm ent s achieve int erpersonal redist ribut ion of m arket incom es t hrough t axes and social benefit s ( cash benefit s and benefit s in kind) . I n recent years considerable progress has been m ade in em pirical research on t he im pact of social prot ect ion syst em s on incom e inequalit y. But m ost analyses of social prot ect ion are focussed on public arrangem ent s. However, social effort is not rest rict ed t o t he public dom ain; all kinds of privat e arrangem ent s can be subst it ut es t o public program s. The OECD has recent ly done a com prehensive st udy on public and privat e social expendit ure ( Adem a, 2001; Adem a and Ledaique, 2005) . They define privat e program s as ‘social’
when t hey serve a social purpose, ar e subj ect t o governm ent int ervent ion and cont ain an elem ent of int erpersonal redist ribut ion. The dat a gat hered by t he OECD show t hat such privat e social arrangem ent s have a subst ant ial size in m any count ries. I n fact , in several count ries welfare st at e reform s have caused a shift from public t owards privat e social expendit ures. Our quest ion is whet her such a shift from public t o privat e social arrangem ent s affect s t he redist ribut ive im pact of t he w elfare st at e. Theoret ically, it is plausible t hat public and privat e arrangem ent s in social prot ect ion syst em s have dissim ilar dist ribut ional effect s. I n t his paper, we will em pirically invest igat e t he relat ionship, if any, bet ween cross- count ry differences in public and privat e social expendit ure and t he dist ribut ion of incom e in a num ber of wealt hy nat ions. To t hat end, we will use m ore recent dat a on privat e social expendit ure t han in earlier work ( Cam inada and Goudswaard, 2005) . We analyse t he effect s of account ing for privat e social benefit s on social prot ect ion st at ist ics, and link bot h public social spending and privat e social spending t o indicat ors of incom e inequalit y and incom e redist ribut ion. Our purpose is t o present a sim ple and int uit ive analysis which elaborat es on previous work. The aim of t he paper is not t o explain t he household incom e dist ribut ion across count ries, nor will we discuss t he direct ion of t he causalit y of t he relat ionship bet ween cross- count ry differences in incom e inequalit y and t he lev els of social spending. Such an analysis should be based on a t heory which would have t o address several cross- nat ional differences explaining t he household incom e dist ribut ion (cf. Got t schalk and Sm eeding, 1997) . Such a com prehensive approach is far beyond t he scope of t his paper.
This cont ribut ion is st ruct ured as follows. Sect ion 2 sum m arises em pirical result s of t he level of incom e inequalit y and incom e redist ribut ion t hrough t he welfare st at e across count ries. I n sect ion 3 we discuss t he nat ure of privat e social expendit ures and show recent dat a on t hese expendit ures. I n sect ion 4 we perform several em pirical analyses on public and privat e social expendit ures, and t he dist ribut ion of incom e. Sect ion 5 concludes.
2 Em pir ica l e vide n ce on in com e in e qu a lit y a n d in com e r e dist r ibu t ion
The best cross- nat ionally com parable collect ion of incom e dat a is t he Luxem bourg I ncom e St udy ( LI S) . LI S was creat ed specifically t o im prove consist ency across count ries. The LI S dat a are a collect ion of m icro dat a- set s obt ained from a range of incom e surveys in various count ries. The advant age of t hese dat a is t hat ext ensive effort s have been m ade by count r y specialist s t o m ake inform at ion on incom e and household charact erist ics as com parable as possible across a large num ber of count ries. The approach adopt ed by LI S overcom es m ost , but not all, of t he problem s of m aking com parisons across count ries t hat plagued earlier st udies ( Sm eeding, 2002) .
This sect ion sum m arises t he evidence on cross nat ional com parisons of incom e inequalit y over 29 nat ions based on em pirical evidence from LI S ( Mahler and Jesuit , 2006) . Levels of inequalit y can be shown in several ways, e.g. by Lorenz curves, specific point s on t he percent ile dist ribut ion ( P10 or P90) , decile rat ios ( P90/ P10) , and Gini coefficient s or m any ot her sum m ary st at ist ics of inequalit y.
All ( sum m ary) st at ist ics of inequalit y can be used t o rank incom e inequalit y in OECD count ries, but t hey do not always t ell t he sam e st ory. The obvious advant age of t he present at ion of inequalit y by sum m ary st at ist ics is it s abilit y t o sum m arise several nat ions in one pict ure.
Figure 1 show s t he m ost oft en used sum m ary m easure of t he incom e dist ribut ion - t he Gini coefficient of equivalized disposable incom e. The Gini coefficient lies bet ween 0 ( no inequalit y and 1 ( m axim um inequalit y) . The figure indicat es t hat a w ide range of inequalit y of disposable incom e exist s across developed nat ions, wit h t he nat ion wit h t he highest inequalit y coefficient ( Mexico) over t wice as high as t he nat ion wit h t he lowest coefficient ( Denm ark) . I n anot her st udy ( Cam inada and Goudsw aard, 2001) we have shown t hat incom e inequalit y has risen since t he early 1980’s in t he m aj orit y of t he OECD count ries, alt hough it is wrong t o t hink in t erm s of a world- wide t rend ( At kinson, 2000) .
Figure 1: Gini index of equivalized disposable incom e
0,0 0,1 0,2 0,3 0,4 0,5
Denmark 1992 Slovak Rep. 1996 Finland 2000 Netherlands 1999 Norway 2000 Sweden 2000 Germany 2000 Czech Rep. 1996 Luxembourg 2000 Austria 2000 Belgium 2000 France 1994 Poland 1999 Canada 2000 Spain 1990 Switzerland 1992 Australia 1994 Ireland 2000 Italy 2000 UK 1999 USA 2000 Mexico 2002 OECD-22
Source: Brandolini and Sm eeding ( 2005) ; dat a t o be found at w w w .lispr oj ect .org/
But what is t he im pact of welfare st at es on inequalit y? Sm eeding ( 2002) showed t hat social policies, wage dist ribut ions, t im e worked, social and labour m arket inst it ut ions and dem ographic differences all have som e influence on why t here are large differences in inequalit y am ong rich nat ions at any point in t im e. However, in t his paper we focus on social prot ect ion syst em s only.
Most nat ions have designed syst em s of social prot ect ion t o shield t heir cit izens against t he risk of a fall in econom ic st at us due t o unem ploym ent , divorce, disabilit y, ret irem ent , and deat h of a spouse. But t hese social prot ect ion syst em s also aim t o reduce inequalit y bet ween individuals and households. The subst ant ial differences in incom e inequalit y across w elfare dem ocracies are w ell docum ent ed ( e.g. Först er , 2000; At kinson et al, 1995; Got t schalk and Sm eeding, 1997; Först er and d'Ercole 2005) . These differences are oft en at t ribut ed t o social policies. Först er’s em pirical analyses showed t hat in m ost developed count ries, bet ween one t hird and 45 percent of all public t ransfers goes t o t he lower incom es. Korpi and Palm e ( 1998) , for exam ple, showed t hat welfare st at es wit h generous social insurance program s r edist ribut e econom ic resources m ore effect ively and have a m ore equal dist ribut ion of incom es t han welfare st at es wit h less generous insurance schem es. I n general, t ax/ t ransfer syst em s as a whole reduce m arket - incom e inequalit y in all OECD count ries.
Usually t he im pact of social policy on t he dist ribut ions of incom e is calculat ed in line wit h t he work of Musgrave, Case and Leonard ( 1974) , i.e. st at ut ory or budget incidence analysis. That is,
im port ant issues of t ax/ t r ansfer shift ing and behavioural responses are ignored.1 The m easure of t he redist ribut ive im pact of social prot ect ion on inequalit y is st raight forw ardly based on form ulas developed by Kakwani ( 1986) and Ringen ( 1991) :
Redist r ibut ion by gover nm ent = ( pr im ar y incom e – disposable incom e) / ( pr im ar y incom e) x 100.
Figure 2 shows t he reduct ion in inequalit y caused by social prot ect ion, where prim ary incom e inequalit y is given by a sum m ary st at ist ic of pre- t ax, pre- t ransfer m arket incom es and disposable incom e inequalit y is given by t he sam e sum m ary st at ist ic of disposable equivalent incom es. The figure shows t he Gini incom e inequalit y before and aft er t ax es/ t ransfers and t he inequalit y reduct ion coefficient in 22 count ries around t he year 2000. Tax es and t ransfers reduce t he Gini by on average 26 percent . For exam ple Sw eden, Denm ark and Germ any achieve a great er redist ribut ion of econom ic resources ( m ore t han 40 percent ) com pared t o Canada, Spain, Swit zerland and t he Unit ed St at es. I t t urns out t hat t he lat t er count ries are in fact t hose wit h t he least equalit y, while Sweden, Denm ark, and Norway are count ries wit h a rat her low degree of incom e inequalit y.
Figure 2: Gini indices of m arket incom e and equivalized disposable incom e
0,0 0,1 0,2 0,3 0,4 0,5 0,6
Denmark 1992 Slovak Rep. 1996 Finland 2000 Netherlands 1999 Norway 2000 Sweden 2000 Germany 2000 Czech Rep. 1996 Luxembourg 2000 Austria 2000 Belgium 2000 France 1994 Poland 1999 Canada 2000 Spain 1990 Switzerland 1992 Australia 1994 Ireland 2000 Italy 2000 UK 1999 USA 2000 Mexico 2002 OECD-22
Gini Disposable I ncom e Effect Tr ansfer s and Tax es
Source: Brandolini and Sm eeding ( 2005) ; dat a t o be found at w w w .lispr oj ect .org/
However, t he result s in Figure 2 do not show t he redist ribut ive im pact of separat e part s of social prot ect ion syst em s. Recent lit erat ure suggest s t hat t he det erm inat ion of t he relat ionship bet ween social expendit ures and inequalit y should be carried out on a disaggr egat ed basis ( see Swabisch et al, 2003) . Ferrarini and Nelson ( 2002) showed t hat only a lim it ed num ber of st udies have at t em pt ed t o specify t he link bet ween specific social t ransfer program s and incom e inequalit y. Thereby, t he knowledge about t he inst it ut ional st ruct ures t hat produce cert ain dist ribut ive out com es is lim it ed.
Especially earlier st udies t hat decom pose inequalit y int o specific t ransfers do not pay sufficient
1 See for a crit ical survey of effort s t o m easure budget incidence by Sm olensky et al. ( 1987) . How ever , m odels t hat include e.g. behavioral links are beyond t he scope of exist ing em pirical work ( Got t schalk and Sm eeding, 1998, p. 3) . Therefore, r esearchers have rest rict ed t hem selves largely t o account ing exercises which decom pose changes in overall inequalit y int o a set of com ponent s.
at t ent ion t o t he problem of t axat ion of social insurance. To gain a deeper underst anding of t he redist ribut ive m echanism s of t he welfare st at e it is necessary t o disaggregat e t he social t ransfer syst em int o program specific com ponent s. Recent LI S dat a ( Mahler and Jesuit , 2006) show a rough disaggregat ion for a num ber of count ries. Table 1 indicat es t hat on average 24 percent of t he redist ribut ive im pact of t he welfare st at e can be at t ribut ed t o t axes and 76 percent t o t ransfers.
However, t he differences bet ween count ries are quit e large: in t he US 40 per cent of redist ribut ion com es from t axes, while in Swit zerland t axes account for only 2 percent of t ot al redist ribut ion and t ransfers for 98 percent . At t he program level, pensions have by far t he st rongest redist ribut ive im pact . More t han half of t he redist ribut ive im pact of social t ransfers com es from pensions.
Unem ploym ent program s do not cont ribut e very subst ant ially t o incom e redist ribut ion, while ot her program s ( disabilit y, healt h insurance, and social assist ance) account on average for 30 percent of t ot al redist ribut ion in t he count ries present ed in Table 1.
Table 1: Gini indices of privat e sect or and disposable incom es and fiscal redist ribut ion
Count ry
Gini privat e incom e
Gini index disposable incom e
Fiscal Redist ribut ion
From t axes
From t ransfers
From pensions
From unem p.
From ot her Aust r alia 2003 0.460 0.312 32.2% 32% 68% 22% 5% 41%
Belgium 1997 0.481 0.250 48.0% 32% 68% 46% 10% 12%
Canada 2000 0.429 0.315 26.6% 35% 65% 33% 5% 26%
Denm ar k 2004 0.419 0.228 45.6% 22% 78% 35% 8% 36%
Finland 2004 0.463 0.252 45.6% 22% 78% 43% 6% 29%
France 1994 0.485 0.288 40.6% 9% 91% 59% 8% 24%
Ger m any 2000 0.473 0.275 41.9% 25% 75% 53% 5% 17%
I r eland 1987 0.500 0.328 34.4% 26% 74% 12% 10% 52%
Net her lands 1999 0.372 0.231 37.9% 31% 69% 36% 2% 30%
Nor w ay 2000 0.403 0.251 37.7% 26% 74% 37% 3% 35%
Sw eden 2000 0.447 0.252 43.6% 19% 81% 39% 10% 32%
Sw it zer land 2002 0.392 0.274 30.1% 2% 98% 81% 4% 14%
UK 1999 0.498 0.343 31.1% 20% 80% 27% 1% 52%
USA 2004 0.481 0.372 22.7% 40% 60% 33% 2% 25%
Mean 0.450 0.284 37.0% 24% 76% 40% 6% 30%
Sour ce: LI S dat a based on Mahler and Jesuit ( 2006) ; updat ed dat a ar e available at :
ht t p: / / www .lisproj ect .org/ publicat ions/ fiscalredist dat a/ fiscred.ht m; and ow n calculat ions
3 Pr iva t e socia l e x pe n dit u r e s
The OECD defines social expendit ures as ‘t he provision by public and priv at e inst it ut ions of benefit s t o, and financial cont ribut ions t arget ed at , households and individuals in order t o provide support during circum st ances which adversely affect t heir welfare, provided t hat t he provision of t he benefit s and financial cont ribut ions const it ut es neit her a direct paym ent for a part icular good or service nor an individual cont ract or t ransfer’ ( OECD 2007, p. 6) . Since only benefit s provided by inst it ut ions are included in t he social expendit ure definit ion, t ransfers bet w een households - albeit of a social nat ure - are not in t he social dom ain. Social benefit s include cash benefit s ( e.g. pensions, incom e support during m at ernit y leave, and social assist ance paym ent s) , social services ( e.g.
childcare, care for t he elderly and disabled) and t ax breaks wit h a social purpose ( e.g. t ax expendit ures t owards fam ilies wit h children, or favourable t ax t reat m ent of cont ribut ions t o privat e healt h plans) .
There are t wo m ain crit eria which have t o be sim ult aneously sat isfied for an expendit ure it em t o be classified as social: 1) t he benefit s have t o be int ended t o address one or m ore social purposes; and
2) program s regulat ing t he provision of benefit s have t o involve eit her int er- personal redist ribut ion, or com pulsory part icipat ion.
The dist inct ion bet ween public and privat e social prot ect ion is m ade on t he basis of whoever cont rols t he relevant financial flows; public inst it ut ions or privat e bodies.
Wit hin t he group of privat e social benefit s, t wo broad cat egories can be dist inguished: m andat ory and volunt ary privat e social expendit ure. Mandat ory privat e social expendit ure is social support st ipulat ed by legislat ion but operat ed t hrough t he privat e sect or, e.g. direct sickness paym ent s by em ployers t o t heir absent em ployees as legislat ed by public aut horit ies, or benefit s accruing from m andat ory cont ribut ions t o privat e insurance funds. I n som e count ries public disabilit y benefit s ( and som et im es unem ploym ent benefit s) can be supplem ent ed by privat e benefit s wit h m andat ory cont ribut ions, agreed upon in collect ive negot iat ions bet ween em ployers and em ployees.
Volunt ary privat e social expendit ure concerns benefit s accruing from privat ely operat ed program s t hat involve t he redist ribut ion of resources across households and include benefit s provided by NGOs, and benefit accruing from t ax advant aged individual plans and collect ive ( oft en em ploym ent - relat ed) support arrangem ent s, such as for exam ple, pensions, and, in t he US, em ploym ent - relat ed healt h plans.
Table 2 sum m arizes which expendit ures are social and which are not .
Table 2 Cat egorizat ion of benefit s wit h a social purpose a, b
Public Privat e
Mandat ory Volunt ary Mandat ory Volunt ary
Redist ribut ion
Means- t est ed benefit s, social insurance benefit s
Volunt ary
part icipat ion in public insurance program s.
Self- em ployed
‘opt ing in’ t o obt ain insurance coverage.
Em ployer- provided sickness benefit s, benefit s accruing from m andat ory cont ribut ions, t o e.g.
pension or disabilit y insurance.
Tax- advant aged benefit s, e.g.
individual ret irem ent account s,
occupat ional pensions, em ployer- provided healt h plans
No redist r ibut ion
Benefit s from governm ent m anaged individual saving schem es
Non t ax- advant aged act uarially fair pension benefit s
Exclusively privat e:
Benefit s accruing from insurance plans bought at m arket prices given individual preferences.
Not es:
a By definit ion t ransfers bet w een indiv iduals, even when of a social nat ur e, are not considered t o be wit hin t he social dom ain.
b The shaded cells reflect benefit s t hat are NOT classified as social.
Sour ce: OECD, Social Expendit ur e dat abase ( 2007)
Table 3 show s public and privat e social expendit ur e as a percent age of GDP, for t he m ost recent dat a year 2003. Most social support is publicly provided. I n m ost count ries t he share of public social benefit s in t ot al social expendit ures exceeds 85 percent . However, t he role of privat e arrangem ent s of v arying nat ure in providing close subst it ut es t o public social prot ect ion expendit ure is considerable in som e OECD count ries. I n t he Net herlands, Swit zerland, t he Unit ed Kingdom and Korea, t he share of privat e social expendit ure is m ore t han 25 percent , while in t he US t his share is alm ost 40 percent . I n m ost count ries privat e volunt ary expendit ure are dom inant , but t here are except ions: m ainly in Swit zerland m andat ory privat e expendit ures are very high.
Figure 3 show s t hat in a num ber of count ries privat e social expendit ures have risen quit e rapidly over t he years.
There m ay be various explanat ions for t his increase in privat e social expendit ure in various count ries. Lower public prot ect ion m ay induce privat e social arrangem ent s of different nat ure. But a shift from public t o privat e provision of social prot ect ion can also be an explicit policy obj ect ive, t o alleviat e public budget s, or t o st rengt hen incent iv es in t he syst em . For exam ple, t he
privat isat ion of t he sickness benefit program in t he Net herlands was direct ed at increasing t he incent ives for em ployers t o reduce t he num ber of beneficiaries. Policy m akers m ay also want t o realize efficiency gains t hrough a shift from public t o priv at e prov ision, because priv at e pr oviders have st ronger incent ives t o reduce cost s. Anyway, account ing for privat e social expendit ures is im port ant for j udging t he social effort and t he level of social prot ect ion in count ries.
Table 3: Public and privat e social expendit ure, % GDP, 2003
Public expendit ure
( 2)
Privat e m andat ory
( 3)
Privat e volunt ar y
( 4)
Privat e expendit ure
( 5) = ( 3) + ( 4)
Tot al expendit ure
( 6) = ( 2) + ( 5)
Share privat e
( 5) / ( 6) * 100
Aust r alia 17.9 1.2 3.2 4.5 22.4 20%
Aust r ia 26.1 0.9 1.2 2.1 28.1 7%
Belgium 26.5 0.0 3.9 3.9 30.4 13%
Canada 17.3 0.0 5.4 5.4 22.7 24%
Czech Republic 21.1 0.2 0.1 0.4 21.5 2%
Denm ar k 27.6 0.2 2.3 2.5 30.1 8%
Finland 22.5 3.5 1.2 4.6 27.1 17%
France 28.7 0.4 2.3 2.7 31.4 8%
Ger m any 27.3 1.2 1.8 3.0 30.2 10%
Gr eece 21.3 0.0 2.4 2.4 23.7 10%
I celand 18.7 5.1 0.0 5.1 23.8 22%
I r eland 15.9 0.0 0.5 0.5 16.4 3%
I t aly 24.2 1.8 0.5 2.3 26.4 9%
Japan 17.7 0.7 2.6 3.3 21.0 16%
Kor ea 5.7 2.2 0.2 2.4 8.1 29%
Luxem bour g 22.2 2.6 0.1 2.7 24.9 11%
Mexico 6.8 0.0 0.2 0.2 7.0 3%
Net her lands 20.7 0.7 7.0 7.7 28.3 27%
New Zealand 18.0 0.0 0.5 0.5 18.5 2%
Nor w ay 25.1 1.6 1.0 2.6 27.7 9%
Poland 22.9 0.0 0.0 0.0 23.0 0%
Por t ugal 23.5 0.4 1.1 1.5 25.0 6%
Slovak Republic 17.3 0.2 1.1 1.3 18.6 7%
Spain 20.3 0.0 0.3 0.3 20.6 2%
Sw eden 31.3 0.6 2.4 3.0 34.3 9%
Sw it zer land 20.5 7.2 1.1 8.3 28.8 29%
Unit ed Kingdom 20.6 0.8 6.0 6.8 27.4 25%
Unit ed St at es 16.2 0.4 9.7 10.0 26.2 38%
Mean 20.9 1.1 2.1 3.2 24.1 13%
Source: OECD. Social Expendit ur e dat abase 1980- 2003 (w w w .oecd.org/ els/ social/ expendit ure) ; download February 26t h, 2008; and ow n calculat ions.
Figure 3: Change in public and privat e social expendit ure 1980- 2003, % - point s of GDP
- 6 - 4 - 2 0 2 4 6 8 10 12 14 16
Spain Ireland New Zealand Portugal Austria Italy Greece Denmark Norway Luxembourg France Germany Sweden Japan Belgium Australia Finland Iceland Canada UK Netherlands Switzerland USA OECD-22
public pr iv at e
Not e: Count ries are ranked in order of t heir level of privat e social expendit ure in 2003. Not e t hat not all count ries of Table 3 are plot t ed in Figure 3, because 1980- dat a is not available for all count ries.
Source: OECD Social Expendit ur e dat abase 1980- 2003 (w w w .oecd.org/ els/ social/ expendit ure) ; download February 26t h, 2008; and ow n calculat ions.
But what about t he redist ribut ive im pact of privat e social arrangem ent s? I t is plausible t hat t he redist ribut ive effect s of t ransfers are weaker in count ries where program s m ost ly rely on earnings- relat ed schem es com par ed t o count ries wit h m ost ly ( public) m eans- t est ed provisions of t ransfers.
Privat e insurance schem es are act uarially fair as a rule. Most privat e insurances are not earnings- relat ed. I ndividual privat e pension insurances, for exam ple, have a defined cont ribut ion charact er, and t herefor e do not cont ain any elem ent s of ( ex ant e) incom e redist ribut ion. Privat e schem es can also have earnings- relat ed benefit s. I t is som et im es argued t hat earnings- relat ed social insurance benefit s only reproduce inequalit ies in m arket incom e and t her efore do not redist ribut e econom ic resources bet ween incom e segm ent s, in case benefit s are perfect ly earnings- relat ed and t he risk of being in receipt of benefit is equally dist ribut ed in t he populat ion. So, in t hat case a higher share of privat e social prot ect ion will not have any ( part ial) effect on t he dist ribut ion of incom e. However, privat e earnings- relat ed schem es m ay not be act uar ially fair and m ay cont ain elem ent s of solidarit y.
This is oft en t he case when ( supplem ent ary) privat e schem es are negot iat ed by social part ners in collect ive labour cont ract s. These schem es are m andat ory for ( a group of) workers. Defined benefit pension schem es, for exam ple, generally redist ribut e resources bot h wit hin generat ions ( for inst ance t hrough redist ribut ive elem ent s such as t hresholds or ceilings) and across generat ions ( risk sharing, back service) . Defined benefit syst em s for early ret irem ent t end t o redist ribut e t o m em bers w ho leave before t he official ret irem ent age from t hose who st ay. I n fact , as we m ent ioned in t he form er sect ion, privat e social program s by definit ion cont ain elem ent s of int erpersonal redist ribut ion.
Also, t ax advant ages ( t o households or t o em ployers) can be used t o st im ulat e t he provision of priv at e benefit s. This is oft en t he case in supplem ent ary pension program s, where cont ribut ions are t ax exem pt . The fiscal advant ages relat ed t o, for exam ple, supplem ent ary privat e pension plans are posit ively relat ed t o incom e levels in m ost count ries. I n general, as Ferrarini and Nelson ( 2002, pp. 14- 15) show ed, social insurance is less equalising aft er t ax at ion in all count ries.
I n general, we do expect t hat privat e schem es will generat e less incom e redist ribut ion t han public program s, alt hough at t his st age t he dist ribut ional im pact of t aking account for privat e social schem es in a cross- count ry analysis is not fully clear. Privat e arrangem ent s will likely have less redist ribut ional effect s com pared t o public program s. I n addit ion, it is plausible t hat m ainly higher incom e groups will m ake use of privat e social schem es ( Casey and Yam ada, 2002) . Considering also t hat privat e schem es oft en have favourable t ax t reat m ent ( deduct ibilit y of cont ribut ions) , which benefit s t he rich, it is possible t hat privat e social expendit ure has a posit ive effect on incom e inequalit y. I n ot her w ords, w e expect incom e inequalit y t o be relat ively high ( low) in count ries where t he share of privat e arrangem ent in t he t ot al social benefit s is relat ively high ( low) .
4 The link be t w e e n public/ pr iva t e socia l pr ot e ct ion a nd incom e ine qua lit y
We perform ed various cross- count y analyses of t he relat ionship bet ween public and privat e social expendit ures and t he incom e dist ribut ion.2 Obviously, t his analysis is not very sophist icat ed. The m at erial present ed here is only descript ive and does not explain t he household incom e dist ribut ion.
Such an analysis should ideally be based on a t heory, which would have t o address at least t he following cross- nat ional differences (cf. Got t schalk and Sm eeding, 2000, p. 263) : differences in labour m arket s t hat affect earnings of individual household m em bers; difference in sources of capit al and in ret urns t o capit al; dem ographic differences, such as t he ageing of t he populat ion and growt h of single parent households, which affect bot h fam ily needs and labour m arket decisions; and differences across count ries in t ax and t ransfers policies t hat not only affect fam ily incom e direct ly, but also m ay affect work and invest m ent decisions. Such a com prehensive approach is far beyond t he scope of t his paper.
I n Figure 4 ( panel a) , we have plot t ed t he average level of public social expendit ure as percent age of GDP and t he average level of t he Gini coefficient of disposable incom e for count ries, where bot h dat a- it em s are available. Bot h averages are represent ed by t he cross of bot h axes: 21.8 percent for public expendit ures, and 0.292 for t he Gini. Several count ries show levels in social t ransfers above t his average. Ot her count ries com bine a below- average level of public social expendit ure wit h an above- average level in inequalit y.
We find a pret t y good fit of a logarit hm ic OLS- regression wit h t he lev el of t he Gini and t he lev el of public social spending as a percent age of GDP ( a sim ilar regression is done by Gouyet t e and Pest ieau, 1999) ; see Table 4. Using public expendit ure as dependent variable produces t he expect ed negat ive sign, while t he coefficient is st at ist ically significant . I n ot her words, we find a quit e st rong negat iv e relat ionship bet w een public social expendit ures and incom e inequalit y.
Obviously, public social securit y t ransfers are well- t arget ed t owards t he poor.
The pict ure alt ers when we t ake pr ivat e social securit y expendit ures int o account in our analysis;
see Figure 4 panel b. A negat ive relat ionship bet ween privat e social expendit ures and inequalit y can not be found; indeed t he relat ionship is slight ly posit iv e. This is confirm ed by a sim ple regression analysis report ed in Table 4. The est im at ed coefficient of privat e expendit ure- v ariable is posit ive, but not st at ist ically significant . These ar e indicat ions t hat suppor t our hypot hesis t hat public and privat e arrangem ent s in social prot ect ion do have opposit e dist ribut ional effect s. This posit ive ( rat her t han a negat ive) sign m ay reflect t hat higher incom e groups find it easier t o opt in t o privat e social program s.
2 We include 22 count ries in our dat a set on t he basis of dat a availabilit y on bot h incom e ( re) dist r ibut ion m easures and privat e social expendit ur e m easures: Aust ralia, Aust r ia, Belgium , Canada, t he Czech Republic, Denm ark , Finland, Fr ance, Germ any , I reland, I t aly , Luxem bourg, Mexico, t he Net her lands, Norway, Poland, Slovak Republic, Spain, Sw eden, Swit zer land, t he Unit ed Kingdom , and t he Unit ed St at es. See t he Appendix for det ails.
Figure 4: Public and privat e social expendit ure and Gini index of equivalized disposable incom e
Panel a Panel b
y = - 0.007x + 0.45 R2 = 0.519
0,2 0,3 0,4 0,5
0 5 10 15 20 25 30 35
Public Social Ex penditur e
Gini Disposable Income y = 0.001x + 0.288
R2 = 0.003
0,2 0,3 0,4 0,5
0 5 10
Pr iv ate Social Ex penditur e
Gini Disposable Income
Table 4: I m pact of public and privat e social expendit ure on incom e inequalit y around 2003
Dependent var iable I nt er cept
Tot al expendit ure
Public expendit ure
Privat e
expendit ure Adj . R2
- 0.085 - 0.329 0.391
( 1)
( - 0.71) ( - 3.81) * *
- 0.011 - 0.401 0.543
( 2)
( - 0.10) ( - 5.10) * *
- 0.006 - 0.405 0.003 0.520 ( 3)
- ( 0.06) ( - 4.80) * * ( 0.17)
- 0.534 - 0.024 - 0.009
Gini index net disposable incom e
( 4)
( - 30.95) * * ( - 0.90)
Not es: Logarit hm ic OLS- regression; t - st at ist ics in parent heses.
* * Significant at t he 0.01 lev el; * significant at 0.05 level
Source: OECD ( 2007) , LI S- dat a based on Br andolini and Sm eeding ( 2005) , and ow n calculat ions
Not e t hat privat e arrangem ent s m it igat e t he im pact of public social effort on incom e inequalit y t o som e ext ent , alt hough t he est im at ed coefficient of t he t ot al expendit ure- v ariable is st ill negat iv e and significant . I n an earlier paper wit h less recent dat a we found t hat , as a result of t he divergent effect s of public social expendit ure versus privat e social expendit ure, t he relat ionship bet ween t ot al social expendit ures and incom e inequalit y across 16 wealt hy count ries appeared t o be st at ist ically t rivial ( Cam inada and Goudswaard, 2005) .3
We perform ed a sim ilar analysis of t he relat ionship bet ween public and privat e social expendit ures wit h t he reduct ion in incom e inequalit y caused by incom e t ransfers ( incom e redist ribut ion from t axes and social benefit s as defined in sect ion 2) . I n Figure 5 panel a t he expect ed relat ionship is shown:
count ries wit h a higher level of public social expendit ure have m ore incom e redist ribut ion or m ore reduct ion of t he Gini. Panel b shows t hat t here is no obvious relat ionship bet ween privat e social expendit ures and incom e redist ribut ion. This is confirm ed by t he regression result s in Table 5.
3 Moreover, we perform ed anot her analysis as well and we find sim ilar result s in case net rat her t han gross social expendit ures are t aken int o our em pirical analysis. However, dat a on net public and net privat e social expendit ures are not ( yet ) available for Luxem bourg, Poland and Swit zerland. See t he Appendix for det ails.
Figure 5: Public and privat e social expendit ure and t he redist ribut ion of incom e
Panel a Panel b
y = 0.007x - 0.040 R2 = 0.558
0,00 0,05 0,10 0,15 0,20
0 5 10 15 20 25 30 35
Public Social Ex penditur e
Redistribution
y = 0.002x + 0.098 R2 = 0.013
0,00 0,05 0,10 0,15 0,20
0 5 10
Pr iv ate Social Ex penditur e
Redistribution
Table 5: I m pact of public and privat e social expendit ure on incom e redist ribut ion around 2003
Dependent var iable I nt er cept
Tot al expendit ure
Public expendit ure
Privat e
expendit ure adj . R2
- 2.734 1.234 0.604
( 1)
( - 9.17) * * ( 5.75) * *
- 2.738 1.295 0.605
( 2)
( - 9.17) * * ( 5.76) * *
- 2.620 1.188 0.081 0.633 ( 3)
( - 8.81) * * ( 5.23) * * ( 1.59)
- 1.073 0.159 0.149
Redist ribut ion:
( Ginipr e – Ginipost)
( 4)
( - 22.03) * * ( 2.16) *
Not es: Logarit hm ic OLS- regression; t - st at ist ics in parent heses.
* * Significant at t he 0.01 lev el; * significant at 0.05 level
- Gini index of m arket ( pre- governm ent privat e sect or) incom e: Ginipr e
- Gini index net disposable ( post - governm ent ) incom e: Ginipost
Source: OECD ( 2007) , LI S- dat a based on Br andolini and Sm eeding ( 2005) , and ow n calculat ions
5 Conclusions
Most analyses of social prot ect ion are focussed on public arrangem ent s. But social effort is not rest rict ed t o t he public dom ain; all kinds of priv at e arrangem ent s can be close subst it ut es t o public program s. I n fact , OECD- dat a indicat e t hat in several count ries t here has been a shift from public t owards privat e social arrangem ent s. Privat e arrangem ent s ar e considered as ‘social’ when t hey serve a social purpose and when t here is som e kind of governm ent involvem ent . Exam ples are supplem ent ary em ploym ent - based and t ax advant aged pension plans and privat e healt h insurance plans w it h legal st ipulat ions. I n t his cont ribut ion w e analysed t he dist ribut ional effect s of public and privat e social arrangem ent s. I ncom e redist ribut ion is one of t he im port ant obj ect ives of t he welfare st at e. I n all OECD count ries, t he social prot ect ion syst em causes a m ore equal dist ribut ion of incom es. Taxes and t ransfers reduce t he Gini coefficient by roughly 20 t o 50 percent in OECD count ries. Public pensions have t he st rongest r edist ribut ive im pact . Based on cross- count ry regressions, we find a negat ive relat ionship bet ween public social expendit ures and incom e inequalit y.
However, we expect t hat privat e schem es will generat e less redist ribut ion from rich t o poor. And indeed, w e do not find a st at ist ically significant relat ionship bet w een privat e social expendit ures and incom e inequalit y and incom e redist ribut ion. Consequent ly, changes in t he public/ privat e m ix in t he provision of social prot ect ion m ay affect t he redist ribut ive im pact of t he welfare st at e.
Account ing for privat e social arrangem ent s m at t ers as far as t he dist ribut ional im pact of t he social prot ect ion syst em is concerned.
Refer e nces
Adem a, W. ( 2001) , 'Net Social Expendit ure: 2nd edit ion', Labour Market and Social Policy - Occasional Papers 52, Par is: OECD.
Adem a, W., and M. Ledaique ( 2005) , ‘Net social expendit ure. 2005 edit ion’, OECD Social Em ploym ent and Migrat ion Working Paper 29, Paris: OECD.
At kinson, A.B. ( 2000) , ‘The Changing Dist ribut ion of I ncom e: Evidence and Explanat ion,’ Germ an Econom ic Review 1 ( 1) , pp. 3- 18.
At kinson, A.B., L. Rainwat er, and T.M. Sm eeding ( 1995) , 'I ncom e Dist ribut ion in OECD Count ries:
Evidence from t he Luxem bourg I ncom e St udy', OECD Social Policy St udies 18, Paris: OECD.
Brandolini, A., and T.M. Sm eeding ( 2005) , 'I nequalit y Pat t erns in West ern Dem ocracies: Cross- Count ry Differences and Tim e Changes', m im eo t o be found at w w w .lisproj ect .org/ w orkshop.ht m.
Cam inada, K., and K.P. Goudswaard ( 2001) , ‘I nt er nat ional Trends in I ncom e I nequalit y and Social Policy,’ I nt ernat ional Tax and Public Finance 8 ( 4) , pp. 395- 415.
Cam inada. K., and K.P. Goudswaard ( 2005) , ‘Are public and privat e social expendit ures com plem ent ary?’, I nt ernat ional Advances in Econom ic Research 11 ( 2) , pp. 175- 189.
Casey, B.H., and A. Yam ada ( 2002) , 'The Public- Privat e Mix of Ret irem ent I ncom e in Nine OECD Count ries: Som e Evidence from Micro- dat a and an Explorat ion of it s I m plicat ions', Luxem bourg I ncom e St udy Working Paper Series 311, Luxem bourg: LI S.
Ferrarini, T., and K. Nelson ( 2002) , 'The I m pact of Taxat ion on t he Equalising Effect of Social I nsurance t o I ncom e I nequalit y: A Com parat ive Analysis of Ten Welfare St at es', Luxem bourg I ncom e St udy Working Paper Series 327, Luxem bourg: LI S.
Först er, M. ( 2000) , 'Trend and Driving Fact ors in I ncom e Dist ribut ion and Povert y in t he OECDF Area', Labour Market and Social Policy Occasional Papers 42, Par is: OECD.
Först er, M., and M.M. d'Ercole ( 2005) , ‘I ncom e dist ribut ion and povert y in OECD count ries in t he second half of t he 1990s’, OECD Social Em ploym ent and Migrat ion Working Papers no. 22, Paris: OECD.
Got t schalk, P., and T.M. Sm eeding ( 1997) , ‘Cross- Nat ional Com parisons of Earnings and I ncom e I nequalit y ,’ Journal of Econom ic Lit erat ure 35 ( June) , pp. 633- 687.
Got t schalk, P., and T.M. Sm eeding ( 2000) , ‘Em pirical Evidence on I ncom e I nequalit y in I ndust rialized Count ries,’ in: A.B. At kinson and F. Bourgignon ( eds.) , Handbook of I ncom e Dist ribut ion, New York: Elsevier- Nort h Holland Publishers, vol. 1, pp. 262- 307.
Gouyet t e, C., and P. Pest ieau ( 1999) , ‘Efficiency of t he Welfare St at e,’ Kyklos 52 ( 4) , pp. 537- 553.
Kakwani, N.C. ( 1986) , Analyzing Redist r ibut ion Policies, Cam bridge: Cam bridge Universit y Press.
Korpi, W., and J. Palm e ( 1998) , ‘The Paradox of Redist ribut ion and St rat egies of Equalit y: Welfare I nst it ut ions, I nequalit y and Povert y in t he West ern Count ries,’ Am erican Sociological Review 63 ( 5) , pp. 661- 687.
Luxem bourg I ncom e St udy, LI S- w ebsit e w w w .lisproj ect .org.
Mahler, V.A., and D.K. Jesuit ( 2006) , ‘Fiscal Redist ribut ion in t he Developed Count ries: New I nsight s from t he Luxem bourg I ncom e St udy’, Socio- Econom ic Review 4, pp. 483- 511 ( see also t heir LI S Working Paper 392; Dat a of t heir research proj ect is available at ht t p: / / www .lisproj ect .org/ publicat ions/ fiscalredist dat a/ fiscred.ht m.
Musgrave, R.A., K.E. Case, and H.B. Leonard ( 1974) , ‘The Dist ribut ion of Fiscal Burdens and Benefit s,’ Public Finance Quart erly 2 ( July) , pp. 259- 311.
OECD ( 2007) , Social Expendit ure dat abase 1980- 2003 (w w w .oecd.org/ els/ social/ expendit ur e); download February 26t h, 2008.
Ringen, S. ( 1991) , ‘Households, st andard of living and inequalit y,’ Review of I ncom e and Wealt h 37 ( 1) , pp. 1- 13.
Sm eeding, T. ( 2002) , 'Globalizat ion, I nequalit y and t he Rich Count ries of t he G- 20: Evidence from t he Luxem bourg I ncom e St udy ( LI S) ', Luxem bourg I ncom e St udy Working Paper Series 320, Luxem bourg: LI S.
Sm olensky, E., W. Hoyt , and S. Danziger ( 1985) , ‘A Crit ical Survey of Effort s t o Measure Budget I ncidence,’ in: H.M. van de Kar and B.L. Wolfe ( eds.) , The Relevance of Public Finance for Policy- Making, Proceedings I I FP Congress 1985, Det roit , pp. 165- 179.
Sw abisch, J., T. Sm eeding, and L. Osberg ( 2003) , 'I ncom e Dist ribut ion and Social Expendit ures: A Cross- Nat ional Perspect ive', Luxem bourg I ncom e St udy Working Paper Series 350, Luxem bourg: LI S.
Appe n dix : D a t a a n d cor r e la t ion t e st s a cr oss cou n t r ie s
Our research hypot hesis is t hat t he level of public social expendit ur e and incom e inequalit y across count r ies are negat iv ely cor r elat ed, w hile pr ivat e social expendit ur e m ay have a non- negat ive redist r ibut ional effect . To analyze t his hypot hesis we include 22 count r ies in our dat a set on t he basis of dat a av ailabilit y on bot h I ncom e ( Re) Dist ribut ion m easures and Privat e Social Expendit ure m easures ( gross and/ or net ) . For 19 count r ies all dat a it em s are available. For anot her t hree count r ies - Luxem bourg, Poland and Sw it zerland - only dat a of net social expendit ure are m issing ( alt hough dat a on gross privat e social ex pendit ur e is available) , so w e put t hese count ries also in our dat a set . As a result w e have perform ed regression analyses w it h t w o dat a set s cont aining 22 respect ively 19 OECD- count r ies.
Dat a set 1 I nequalit y of m arket incom e and net disposable incom e Gross public and privat e social expendit ur e, % GDP, 2003
Gini m arket incom e
Gini disposable
incom e
Effect t axes and t ransfers
Redist ribut ion m easure
Most recent
dat a year Public
Privat e m andat ory
Privat e
volunt ary Privat e Tot al
Share privat e
Most recent dat a year Aust r alia 0.406 0.311 0.095 23.4% 1994 17.9 1.2 3.2 4.5 22.4 20% 2003 Aust r ia 0.382 0.260 0.122 31.9% 2000 26.1 0.9 1.2 2.1 28.1 7% 2003 Belgium 0.343 0.277 0.066 19.2% 2000 26.5 0.0 3.9 3.9 30.4 13% 2003 Canada 0.391 0.302 0.089 22.8% 2000 17.3 0.0 5.4 5.4 22.7 24% 2003 Czech Republic 0.338 0.259 0.079 23.4% 1996 21.1 0.2 0.1 0.4 21.5 2% 2003 Denm ar k 0.416 0.236 0.180 43.3% 1992 27.6 0.2 2.3 2.5 30.1 8% 2003 Finland 0.361 0.247 0.114 31.6% 2000 22.5 3.5 1.2 4.6 27.1 17% 2003 France 0.462 0.288 0.174 37.7% 1994 28.7 0.4 2.3 2.7 31.4 8% 2003 Ger m any 0.438 0.252 0.186 42.5% 2000 27.3 1.2 1.8 3.0 30.2 10% 2003 I r eland 0.381 0.323 0.058 15.2% 2000 15.9 0.0 0.5 0.5 16.4 3% 2003 I t aly 0.427 0.333 0.094 22.0% 2000 24.2 1.8 0.5 2.3 26.4 9% 2003 Luxem bour g 0.352 0.260 0.092 26.1% 2000 22.2 2.6 0.1 2.7 24.9 11% 2003 Mexico 0.495 0.471 0.024 4.8% 2002 6.8 0.0 0.2 0.2 7.0 3% 2003 Net her lands 0.357 0.248 0.109 30.5% 1999 20.7 0.7 7.0 7.7 28.3 27% 2003 Nor w ay 0.408 0.251 0.157 38.5% 2000 25.1 1.6 1.0 2.6 27.7 9% 2003 Poland 0.378 0.293 0.085 22.5% 1999 22.9 0.0 0.0 0.0 23.0 0% 2003 Slovak Republic 0.303 0.241 0.062 20.5% 1996 17.3 0.2 1.1 1.3 18.6 7% 2003 Spain 0.354 0.303 0.051 14.4% 1990 20.3 0.0 0.3 0.3 20.6 2% 2003 Sw eden 0.432 0.252 0.180 41.7% 2000 31.3 0.6 2.4 3.0 34.3 9% 2003 Sw it zer land 0.358 0.307 0.051 14.2% 1992 20.5 7.2 1.1 8.3 28.8 29% 2003 Unit ed Kingdom 0.506 0.345 0.161 31.8% 1999 20.6 0.8 6.0 6.8 27.4 25% 2003 Unit ed St at es 0.451 0.368 0.083 18.4% 2000 16.2 0.4 9.7 10.0 26.2 38% 2003
Num ber of count r ies 22 22 22 22 22 22 22 22 22 22
Average 0.397 0.292 0.105 26.2% 21.8 1.1 2.3 3.4 25.2 12.7%
Sour ce: OECD ( 2007, w w w .oecd.org/ els/ social/ expendit ur e) , LI S- dat a based on Br andolini and Sm eeding ( 2005, w w w .lisproj ect .or g/ w orkshop.ht m; and ow n calculat ions
Dat a set 2 I nequalit y of m arket incom e and net disposable incom e Net public and privat e social expendit ur e, % GDP, 2003
Gini Market I ncom e
Gini Disposable
I ncom e
Effect t axes and t ransfers
Redist ribut ion Measure
Most recent
dat a year Public
Privat e m andat ory
Privat e
volunt ary Privat e Tot al Share privat e
Most recent dat a year Aust r alia 0.406 0.311 0.095 23.4% 1994 17.2 1.0 2.7 3.7 20.9 18% 2003 Aust r ia 0.382 0.260 0.122 31.9% 2000 20.6 0.5 1.0 1.5 22.2 7% 2003 Belgium 0.343 0.277 0.066 19.2% 2000 22.9 0.0 3.1 3.1 26.0 12% 2003 Canada 0.391 0.302 0.089 22.8% 2000 17.2 0.0 4.3 4.3 21.4 20% 2003 Czech Republic 0.338 0.259 0.079 23.4% 1996 19.5 0.2 0.1 0.3 19.8 2% 2003 Denm ar k 0.416 0.236 0.180 43.3% 1992 20.3 0.1 1.2 1.2 21.6 6% 2003 Finland 0.361 0.247 0.114 31.6% 2000 17.7 2.1 0.7 2.9 20.6 14% 2003 France 0.462 0.288 0.174 37.7% 1994 25.5 0.3 2.2 2.5 28.0 9% 2003 Ger m any 0.438 0.252 0.186 42.5% 2000 25.8 0.6 1.6 2.2 28.0 8% 2003 I r eland 0.381 0.323 0.058 15.2% 2000 14.0 0.0 0.5 0.5 14.4 3% 2003 I t aly 0.427 0.333 0.094 22.0% 2000 20.6 1.4 0.4 1.9 22.5 8% 2003 Mexico 0.495 0.471 0.024 4.8% 2002 7.6 0.0 0.2 0.2 7.7 2% 2003 Net her lands 0.357 0.248 0.109 30.5% 1999 17.9 0.4 5.1 5.6 23.5 24% 2003 Nor w ay 0.408 0.251 0.157 38.5% 2000 20.2 0.9 0.6 1.5 21.7 7% 2003 Slovak Republic 0.303 0.241 0.062 20.5% 1996 16.1 0.2 0.9 1.1 17.2 6% 2003 Spain 0.354 0.303 0.051 14.4% 1990 17.6 0.0 0.3 0.3 18.0 2% 2003 Sw eden 0.432 0.252 0.180 41.7% 2000 24.3 0.3 1.5 1.8 26.1 7% 2003 Unit ed Kingdom 0.506 0.345 0.161 31.8% 1999 19.3 0.6 4.8 5.4 24.7 22% 2003 Unit ed St at es 0.451 0.368 0.083 18.4% 2000 17.3 0.4 8.9 9.2 26.5 35% 2003
Num ber of count ries 19 19 19 19 19 19 19 19 19 19
Average 0.403 0.293 0.110 27.0% 19.0 0.5 2.1 2.6 21.6 11.1%
Sour ce: OECD ( 2007, w w w .oecd.org/ els/ social/ expendit ur e) , LI S- dat a based on Br andolini and Sm eeding ( 2005, w w w .lisproj ect .or g/ w orkshop.ht m; and ow n calculat ions
