Absolutely poor or relatively poor: does is matter for social policy design ? Chris de Neubourg Washington DC, November 2009.

Absolutely poor or relatively poor: does is matter for social policy design ?Chris de Neubourg

Washington DC, November 2009

Maastricht Graduate School of Governance

1. Does it matter for priority setting ?2. Does it matter for identifying the poor?3. Does it matter for identifying the long term

poor ?4. Does it matter when setting levels and

eligibility of means tested benefits ?5. Does it matter when assessing the

incidence and adequacy for targeted and non-targeted benefits ?


Basic data

• Poverty estimates for 15 EU countries and the USA for the period 1993 – 2000

• Using household surveys in the EU and the in the USA (ECHP and CNEF-PSID)

• Using relative poverty concept (Laeken) and an absolute concept (Orshansky) for the 16 countries


Does is matter for setting policy priorities ?

Table 1: Poverty incidence (in % of individuals, in 1996 and 2000) Laeken (relative) poverty Orshansky (absolute) poverty


Does it matter for identifying the poor? Netherlands United States

All1 O2 L3 4 All O L By: Gender - Female 50.4 49.4 43.7 51.8 54.5 54.5 - Male 49.6 50.6 56.3 47.8 45.0 45.0 By: Age groups - 0-15 20.1 29.5 26.4 23.3 33.7 31.0 - 16-24 10.7 21.6 18.9 12.2 16.2 14.2 - 25-49 38.6 34.3 34.9 37.0 29.5 32.6 - 50-64 17.6 9.9 16.010 15.6 8.4 9.5 - Above 65 13.0 4.7 3.8 11.5 22.0 34.6 By: Household type - Single adult 15.6 17.2 13.7 13.3 17.3 14.210 - Two adults 28.6 10.3 12.5 22.4 9.2 13.601 - Other adult households 8.3 5.9 13.7 8.1 3.9 4.3 - Two adults and children 27.9 23.0 34.0 23.6 12.9 20.301 - Other households with children 19.7 43.5 26.210 32.6 56.8 47.601 By: Main income source5 -Wage 70.3 47.2 63.910

82.1 52.9 79.101 - Self-Employment 3.1 2.1 4.5 - Pensions 15.9 6.8 8.2 11.3 17.1 17.0 - Unemployment benefits 0.7 0.9 1.2

2.8 21.3 2.301 - Other benefits 9.1 38.8 21.005 - Private income 1.0 4.3 1.210 3.9 8.8 1.601


Does it matter for identifying the long term poor ?

Figure 1: Persons at persistent risk of poverty (long term poor) as a percentage of total number of poor for 2000; poverty estimated by a relative (EU-Laeken) and an absolute (Orshansky) poverty indicator.

0102030405060708090

100

Portugal

Greece

Luxem

bour

gIta

ly

United S

tates

Irelan

d

Austria

United K

ingdo

m

France

Spain

Belgium

German

y

Finlan

d

Denmark

Netherl

ands

Ove

rlap

(%)

Relative poverty Absolute poverty

Note: The overlap is calculated dividing the persistent poverty rate by the overall poverty rates; countries are ranked according to the share of persistent poverty using the relative poverty indicator. Source: Own calculations ECHP and CNEF-PSID


Percentage of persistent “Laeken” poor who are also “Orshansky” poor , 2000

Ireland 48.5

The Netherlands 28.5

Austria 32.6

UK 41.8

USA na


Does it matter when setting levels and eligibility of means tested benefits ?

• Means tested benefits in Europe implicitly set absolute poverty line in the form of minimum income guarantees

• Social assistance and social assistance plus• Are these implicit poverty lines de facto

similar to the (US)“Orshansky” poor ?


Average social assistance level as percentage of “Orshansky” poverty line, 1993 - 2000

0

20

40

60

80

100

120

140

160

1993 1994 1995 1996 1997 1998 1999 2000

Belgium

Ireland

Sweden

Germany

UK



0

20

40

60

80

100

120

140

160

1993 1994 1995 1996 1997 1998 1999 2000

Denmark

Netherlands

Italy

Finland



0

20

40

60

80

100

120

140

160

1993 1994 1995 1996 1997 1998 1999 2000

France

Spain

Portugal

Austria



0

50

100

150

200

250

1993 1994 1995 1996 1997 1998 1999 2000

Denmark

Netherlands

Belgium

France

Ireland

Italy

Spain

Portugal

Austria

Finland

Sweden

Germany

UK


Average minimum income guarantee as percentage of “Orshanky” poverty line

0

50

100

150

200

250

1993 1994 1995 1996 1997 1998 1999 2000

Denmark

Netherlands

Belgium

France

Ireland

Italy

Spain

Portugal

Austria

Finland

Sweden

Germany

UK


Minimum income guarantee as percentage of “Laeken” poverty line, 1993 - 2000

0

50

100

150

200

250

1993 1994 1995 1996 1997 1998 1999 2000

Denmark

Netherlands

Belgium

France

Ireland

Italy

Spain

Portugal

Austria

Finland

Sweden

Germany

UK


Average yearly minimum income guarantee 1993 – 2000 (in nominal euro)

0

5000

10000

15000

20000

25000

1993

1994

1995

1996

1997

1998

1999

2000


Average yearly social assistance benefit and average yearly minimum income guarantee as percentage of “Laeken” (rel) and “Orshansky” (us) poverty lines, 2000.

0

20

40

60

80

100

120

140

160

180

200

socass%relpl

socass%ruspl

minincom%relpl

minincom%uspl


Does it matter when assessing the incidence and adequacy for targeted and non-targeted benefits ?


Table 4: Incidence of social transfers1 for different population groups (in % of individuals, 2000)

Old age pensions2

Family allowance

Other social insurance3

Social assistance Other benefits4

O5 L6 NP7 O L NP O L NP O L NP O L NP Netherlands 7.5 8.2 20.2 57.2 65.5 45.3 18.2 12.0 12.7 14.1 11.4 2.2 16.4 19.6 7.5 Wald test: O L8 L NP O NP

ns .01 .01

ns .01 .01

ns ns .10

ns .10 .01

ns .10 .01

Ireland 23.3 43.4 22.2 61.4 59.1 66.7 40.3 35.5 28.3 42.4 34.0 7.3 6.1 4.0 9.4 Wald test: O L L NP O NP

.01

.01 ns

ns ns ns

ns ns .05

ns .01 .01

ns ns .01

Austria 44.7 51.9 33.8 56.0 58.1 58.7 14.2 21.2 18.2 2.1 2.0 0.4 7.6 13.0 9.6 Wald test: O L L NP O NP

ns .01 .05

ns ns ns

ns ns ns

ns ns .10

ns ns ns

United Kingdom

28.0 41.6 28.6 53.8 52.0 46.5 25.6 21.3 17.2 na10 na na 36.4 33.2 16.9

Wald test: O L L NP O NP

.01

.01 ns

ns ns .01

ns ns .01

.

.

.

ns .01 .01

United States9

22.1 25.0 22.4 45.7 24.5 7.4

Wald test: O L L NP O NP

ns ns ns

.01

.01

.01

Notes: 1 The incidence rate represents the % of individuals living in households receiving income from a particular benefit category. 2 Pensions include social and private pensions. 3 Other social insurance includes unemployment and sickness/disability benefits. 4 Other benefits include education, housing and other allowances. 5 O = Laeken and Orshansky poor (using same equivalence scales). 6 L = Laeken poor but not Orshansky poor (using same equivalence scales). 7 NP = Not poor. 8 Wald test on difference between means of population groups (taking sample design into account). Indicated significance levels: 1% (.01), 5% (.05), 10% (.10) and not significant (ns). 9 For the United States we can only distinguish between pensions and other social transfers. 10 Not available or not calculated. Source: Own calculations ECHP and CNEF-PSID


Figure 6: Incidence of social assistance (below median income, 2000)

050

10

0

LO

Netherlands

050

10

0

LO

Ireland0

50

10

0

LO

Austria

050

10

0

LO

United States

Pe

rce

nta

ge

Equivalent income


Mean value of social assistance (below median income, 2000)

LO

LO

Netherlands

LO

LO

Ireland

LO

LO

Austria

LO

LO

United States

Eq

uiv

ale

nt

be

ne

fit

Equivalent income


Incidence of family allowances (below median income, 2000)

050

10

0

LO

Netherlands

050

10

0

LO

Ireland0

50

10

0

LO

Austria

050

10

0

LO

UK (BHPS)

Pe

rce

nta

ge

Equivalent income


Mean value of family allowances (below median income, 2000)

LO

LO

Netherlands

LO

LO

IrelandL

O

LO

Austria

LO

LO

UK (BHPS)

Eq

uiv

ale

nt

fam

ily a

llow

an

ce

Equivalent income


Absolutely poor or relatively poor: Does is matter ?• For policy priorities• For identification of the poor• For identification of the poor at risk of

persistence poverty• For setting minimum income guarantee

levels• For assessing incidence and adequacy

of policy measures

Absolutely poor or relatively poor: does is matter for social policy design ? Chris de Neubourg Washington DC, November 2009.

Documents

poverty incidence

absolute poverty line

relative poverty indicator

persistent poverty rate

share of persistent

persistent laeken poor

usorshansky poor

implicit poverty lines