here - CSI Nuffield

advertisement
How can we explain the high
Muslim levels of poverty?
Anthony Heath
Centre for Social Investigation
Nuffield College, Oxford
Poverty rates among different religious
groups
60
50
50
40
30
20
14
18
10
0
Poverty defined as equivalized household income less than 60% of the national median
Documenting the problem
• Our evidence from the UKLHS is unequivocal:
poverty is a much more common experience
among Muslims than among other religious
groups
• Long known that ethnic groups with Pakistani
or Bangladeshi origins have high rates of
poverty, so is this an ethnic rather than a
religious phenomenon?
Poverty by ethnicity and religious
affiliation
White
B Caribbean
B African
Indian
Pakistani
Bangladeshi
Chinese
Other
N
Angl
Cath
14
20
36
18
23
37
24
14
30
33
16
16
13
22
19
O C Musl Hind Sikh Jew Budd Other
30
56
38
57
49
50
12
20
30
12862 6521 8395 4452 1145
11
20
21
40
30
18
33
14
26
31
570 191
237
None
All
18
19
21
18
60
22
27
22
16
23
37
23
57
46
26
27
690 25862 60925
Possible explanations
• Transitory coincidental factors reflecting
migration history and recent arrival in Britain.
Might be expected to disappear in time
• Factors which might be intrinsic to particular
religious traditions and might persist over
time
• Factors which reflect prejudice towards
Muslims on the part of British employers and
co-workers
Transitory factors
• Recent arrival – lack of fluency in English a major
barrier to employment
• LDC origins – lower levels of educational
attainment restrict people to disadvantaged jobs
• Foreign qualifications mean lower returns to
education
• Lack of knowledge about the UK labour market
• History of migration to declining Northern
industrial areas
Possible intrinsic factors
• Family values – caring responsibilities might
lead to lower female rates of economic
activity, higher family size
• Religious involvement might lead to bonding
rather than bridging social capital
• Mismatch between available work
arrangements and religious preferences
• Mismatch between available childcare
arrangements and religious preferences
Prejudice, discrimination and other
barriers
• Discrimination by employers both in taking on
Muslim workers and promoting them
• Cold reception from non-Muslim co-workers –
the chill factor
• Pre-labour market discrimination, eg in elite
higher education institutions
Practical difficulties in researching
these potential explanations
Ideally one would use a range of research
methods, eg
• field experiments to detect discrimination,
• in-depth interviews to explore the chill factor,
• social network analysis on bridging and
bonding social capital,
• Attitude or ethnographic research to explore
family values
A first approximation using statistical
analysis of the UKLHS
Restricted to available measures in the UKLHS, which are
largely proxies rather than direct measures of the
explanatory concepts. We distinguish
• Confounding factors such as age
• Transitory factors such as generational status, fluency in
English, educational qualifications
• Possible intrinsic/religion-related factors such as religiosity,
attendance at place of worship, number of children
• Possible indicators of labour market barriers
Measures are undoubtedly sub-optimal
Transitory factors
st
1 gen
Anglican
Catholic
Other Christian
Muslim
Hindu
Sikh
Jew
Buddhist
Other
None
All
2
22
13
58
74
45
9
58
25
5
11
% with the characteristics
2 gen 3rd gen Primary
or less
nd
5
18
8
35
24
51
28
10
11
9
10
92
60
79
7
2
4
63
32
64
86
79
37
28
26
27
16
22
20
21
25
22
27
Degree
16
23
25
28
48
28
44
45
23
21
22
Means
Difficulty
with
English
0.01
0.20
0.08
1.05
0.59
0.62
0.02
1.02
0.10
0.03
0.10
Religion-related factors
Anglican
Catholic
Other Christian
Muslim
Hindu
Sikh
Jew
Buddhist
Other
None
All
Means
Children
HH
size
0.32
2.48
0.50
2.86
0.40
2.66
0.98
4.62
0.43
3.57
0.61
4.24
0.49
2.95
0.46
2.68
0.43
2.58
0.48
2.85
0.45
2.82
Civic
0.97
0.84
1.16
0.45
0.54
0.55
1.14
0.81
1.05
0.66
0.81
Female
employed
42
52
44
28
53
53
42
60
42
57
50
% reporting
Weekly
Great
attendance importance
10
17
28
28
32
39
41
59
21
32
33
39
17
25
14
46
39
54
1
3
12
16
Labour market factors
Anglican
Catholic
Other Christian
Muslim
Hindu
Sikh
Jew
Buddhist
Other
None
All
% reporting
Unemployment Working
Self-reported
class
discrimination
3
33
0.28
6
36
0.96
4
30
1.51
11
37
4.92
6
30
6.96
7
39
5.83
8
12
0.21
6
32
5.96
6
34
1.17
8
35
0.64
6
33
0.98
Means
Earnings (£)
8,910
11,334
10,162
7,399
12,800
10,961
14,267
12,484
8,108
12,257
10,934
In summary
• Muslims share many of the transitory disadvantages with
Hindus and Sikhs, so these factors unlikely to explain their
higher poverty rates
• Muslims do stand out with respect to religiosity, number of
children and women’s economic activity – NB religiosity
seems to persist across generations, the other variables less
so
• Muslims do stand out with respect to unemployment rate and
earnings, though surprisingly not with self-reported
discrimination or occupational level
Statistical analysis of the risks of
poverty
We use standard statistical techniques to attempt to see how far
these sorts of factor account for excess Muslim rates of poverty
1. We first take account of confounding factors and ethnicity
2. Then the transitional factors – generation, education and
language difficulties
3. Next number of children, household size and economic
activity
4. Separate model for attendance, religiosity (NS) and civic
engagement
5. Final model adds social class, earnings, discrimination (NS)
and unemployment
Results of the regression analysis:
average marginal effects
Model 1:
Religion+
controls
Model 2:
M1+English,
edu and gen
Model 3:
M2+HH &
inactivity
Model 4:
M3+social
capital
Model 5:
M4+labour
market
Anglican (ref)
Catholic
Other Christian
Muslim
Hindu
Sikh
Jew
Buddhist
Other
None
0.023***
-0.011
0.200***
0.073***
0.103***
-0.003
-0.006
0.036*
0.020***
0.019**
-0.002
0.160***
0.059**
0.062*
0.028
-0.025
0.041*
0.014**
0.014*
-0.005
0.092***
0.058**
0.038
-0.008
-0.015
0.031
0.014**
0.012
-0.002
0.085***
0.050**
0.032
-0.009
-0.016
0.031
0.009*
0.011
0.006
0.073***
0.033
0.019
-0.029
-0.030
0.019
0.006
Pseudo R2
N
0.052
59,319
0.097
58,765
0.171
58,398
0.179
58,398
0.275
49,075
The key findings
• After taking account of confounding factors,
the Muslim poverty rate is 20 percentage
points higher than that of Anglicans
• Each block of factors (especially the second)
explains part of this excess poverty rate
• In the case of the other religions, the
predictors successfully explain their initial
excess poverty rates, but for Muslims a 7 point
gap remains
Do the mechanisms work differently
for Muslims?
• In general processes operate similarly for all
the religious groups
• Muslims are distinctive in the size of the
effects of generational status
• But this largely re-states the problem
So how might we explain Muslim
exceptionalism?
• Geographical segregation and concentration?
But also applies to Sikhs so cannot explain the
difference
• Replenishment of Muslim communities? Largely
specific to replenishment from Pakistan, which
the measure of ethnic origins will have taken into
account
• Islamophobia and the ‘chill factor’? The most
plausible speculation – supported by our current
research on attitudes to Muslim migrants
Some grounds for optimism
• Transitional factors mean that we expect some
convergence over time and generations in
poverty rates
• Also generational change in family size and
women’s economic activity – not actually all that
intrinsic to Islam
• Considerable educational progress, with Muslim
women in particular catching up rapidly
• Religiosity itself not a cause of poverty (either
generally or among Muslims)
But important unanswered questions
• Not clear why Muslims have such low levels of
civic engagement (also shown in other data
sources)
• Not clear why Muslim earnings are so low
(again shown in other data sources)
• Not clear how to test whether the chill factor
really is a major part of the explanation
Policy responses?
• Tackling discrimination in universities and
employment
• Tackling poor economic environment,
especially in Northern England
• Tackling barriers to women’s employment (eg
mismatch with respect to childcare or
employment conditions)
• Countering Islamophobia in the media
Download