Civitas talk on social mobility, 4 November 2010
CLASS ON THE BRAIN?
Then:
'The vast majority of people still remain in their class of origin' Professor Tom Bottomore, 1965
'The race is still rigged - the upper and middle class...is still largely self-recruiting' Professor Ralph Miliband, 1969
'The chances that those born into different social classes will stay in those classes are still very high.' Professor Peter Worsley, 1970
'Virtually all movement across the non-manual/manual division is short-range... There has not been much change in rates of mobility since World War I' Professor Anthony Giddens, 1973
And now:
"Birth, not worth, has become more and more a determinant of people's life chances... Britain is a closed shop society"
(Alan Milburn, Foreword to 2009 Report on Fair Access to the Professions)
"Britain is no longer a socially mobile nation. We fall behind many of our international competitors in offering people a route up the social ladder... Social mobility has ground to a halt"
(Conservative Party report, Through the Glass Ceiling, 2009)
'Your parents' occupation will almost determine your occupation'
(Kevin MacGuire, Daily Mirror journalist, at last week-end's Battle of Ideas conference)
'Upward social mobility is a total myth'
(Neil Davenport, Sociology lecturer, at the same conference)
?
MYTH 1: BRITAIN IS A 'CLOSED SHOP SOCIETY'
'SAD' ASSUMPTIONS CHALLENGED BY GOLDTHORPE'S EVIDENCE
Social Mobility and class structure in modern Britain (1987). Survey of 10,000 men:
Mobility is common (49% move between 3 classes);
Movement down as well as up (41% prof/managerial class sons failed to emulate their fathers);
Movement across the whole range (15% prof/managerial class sons > working class; 16% working class sons > prof/managerial class).
GOLDTHORPE'S RESULTS REPLICATED:
...by Goldthorpe on 1983 data:
53% had moved;
22% of working class sons > prof/managerial class.
...by other researchers:
Marshall et al (? prof-managerial class came from working class)
Heath & Payne (throughout 20th century, more than half of each generation moved; 40% working class upwardly mobile)
...by birth cohort studies:
1958 cohort:
45% of men and 39% women upwardly mobile by age 33
27% of men and 37% of women downwardly mobile by 33
1970 cohort:
42% of men and 41% women upwardly mobile by age 30
30% of men and 35% of women downwardly mobile by 30
INTERNATIONAL COMPARISONS: DOES UK REALLY DO WORSE?
o Erikson and Goldthorpe rank UK 8th of 15 on overall fluidity (more mobile than Germany, France or Netherlands; less than Sweden, USA, Australia)
o 2001 Performance & Innovation Unit report concludes any differences are 'modest'.
o OECD puts UK in middle of ranking on downward mobility.
BUT 2010 National Equality Panel (Hills report) claims UK mobility lower than almost any other European country. Based on Sutton Trust parent-child income correlations. But...
o Hills accepts % men in higher occupational position than fathers same as Ireland, France and Germany
o Blanden admits: 'Large standard errors on the Australian, French, British and US income estimates make it unclear how these countries should be ranked' (LSE Paper 2009)
MYTH 2: THINGS ARE BAD AND GETTING WORSE
1958-1970 COHORT COMPARISONS
Sutton Trust economists look at father-son income correlations in 1958 and 1970 birth cohorts. Find apparent fall in fluidity in later cohort:
o 1958 (NCDS) 35% of kids from top income quartile got to top quartile, 17% fell to bottom quartile
o 1970 (BCS) 42% of kids from top income quartile got to top quartile, 11% fell to bottom quartile
Huge media and political attention paid to these studies. BUT...
NO LONG-TERM DECLINE IN FLUIDITY (IF ANYTHING, AN INCREASE):
o 'A trend of slowly increasing social mobility' over the last 200 years (Lambert et al, Soc Res Online 2007)
o For people born 1900-1960 odds ratios reducing (Heath & Payne, Oxford WP 1999)
NO DIFFERENCE IN CLASS MOBILITY BETWEEN 1958 AND 1970 COHORTS:
o 'There is no indication of any current downturn in the total mobility rate... The evidence points to a continuation of the essential stability that has been documented for decades' (Erikson and Goldthorpe, Br Jnl Soc, 2010)
RESEARCH ON BHPS FINDS NO DIFFERENCE IN INCOME MOBILITY FOR COHORTS BORN 1950 TO 1972:
o 'There are no strong changes in intergenerational mobility across cohorts from 1950 to 1972' (Ermisch & Nicoletti ISER WP 2005)
Suspiciously low parent-child income correlation in 1958 cohort probably explains why 1970 association appears to have strengthened
o Transitory fluctuation of earnings appears higher for 58 cohort parents > more error in estimating permanent earnings > weaker correlation with children's earnings
'This slender analysis has had more influence on public policy debate than any academic paper of the last 20 years. The lazy consensus which has decreed the end of social mobility is both wrong and damaging' (David Goodhart, Prospect, 2008)
MYTH 3: THERE SHOULD BE NO DIFFERENCE IN RATES OF SUCCESS OF CHILDREN FROM DIFFERENT CLASSES
GOLDTHORPE ON 'RELATIVE MOBILITY'
The middle class has been getting bigger every generation. Have working class children benefited from this expansion at the same rate as prof/managerial class children? Found prof/managerial class children:
o 4 x more likely to end up in prof/managerial class (3x in follow-up)
o 4 x less likely to end up in working class (5x in follow-up).
Argues that with no class advantages/disadvantages, should be 1:1.
Concludes: 'The post-war project of creating in Britain a more open society has failed'
FURTHER EVIDENCE OF UNEQUAL OUTCOMES:
Heath & Payne 1964-97 election survey data:
o Prof/managerial class child 3.3 x more likely than working class child to end up in prof/managerial class;
o Working class child 3.5 x more likely than prof/managerial class child to end up in working class.
NCDS age 33:
o Class I/II child 2.6 x more likely to be in class I/II
o Class IV/V child 3.8 x more likely to be in class IV/V
BUT WHAT ABOUT ABILITY???
Goldthorpe takes 1:1 as criterion of fairness. So do all recent government reports:
Getting On, Getting Ahead (2008)
o Wants people to 'realise their potential.'
o Assumes 'perfect mobility' = 1:1 ratio.
o Points to 'large and systematic differences in outcomes' as evidence of SAD causes.
Unleashing Aspiration (2009)
o Accepts: 'Individual success should reflect innate talent and ability'
o But assumes: 'While ability and effort do play a part, the effect of class origins on class destinations is much stronger'
Anatomy of Economic Inequality in the UK (2010)
o Assumes measured cognitive differences between classes are due to 'cultural capital'
o 'The systematic nature of the differences makes it hard to suggest there is equality of opportunity'
1:1 assumes equal distribution of talent across every class. BUT...
o in a meritocracy, talented people will be recruited to the higher classes...
o ...where they can be expected to produce more talent children (parent-child IQ correlation = 0.5).
Thus (unlike race) we should expect average ability levels to vary between classes.
MOBILITY RATES RECORDED BY GOLDTHORPE CORRESPOND TO WHAT WE SHOULD EXPECT IN A MERITOCRACY:
PREDICTED % ACTUAL %
Prof/manag class > prof/manag class 59 59
Prof/manag class > working class 21 15
Working class > prof/manag class 18 16
Working class > working class 58 57
IF YOU'RE BRIGHT, YOU'LL PROBABLY SUCCEED IRRESPECTIVE OF CLASS ORIGIN:
High ability children rarely fail irrespective of their class of origin:
o 65% of top IQ quartile get to class I/II
o Only 5% of top IQ quartile end up in class IV/V
But low ability middle class children often succeed when they 'shouldn't':
o 41% of middle class children in the lowest IQ quartile end up in class I/II
o 21% of working class children in the lowest IQ quartile end up in class I/II
The problem is not bright working class kids failing, but dull middle class kids succeeding
ABILITY (AND EFFORT) ARE THE BEST PREDICTORS OF CLASS DESTINATION:
Multiple regression:
FINAL BETA
1 ABILITY TEST SCORE 0.25
2 MOTIVATION 0.13
3 PARENTAL CLASS 0.08
Linear structural equations path model: Total variance explained = 35% of which...
Aspirations for child 1/35
Parents class 3/35
Interest in child's education 3/35
Child's motivation 5/35
Qualifications (ind effect) 6/35
Academic ability 17/35