Jo Blanden Home Ownership and Social Mobility

Home Ownership and Social Mobility Jo Blanden School of Economics, University of Surrey

with Stephen Machin, London School of Economics

15th June 2017

Wednesday, 14 June 2017

1

Context 1 Discussion of problems of low social mobility in Britain are commonplace.

The issue has magnified in several ways more recently: 1). The documented fall in intergenerational income mobility in economic status that earlier research has highlighted.

2). Irrespective of relative mobility patterns, the potential for no absolute mobility gains as real wages and living standards are squeezed.

3). The implications that 2) has had for wealth mobility across the generations.

Context 2 We study all three of these in this paper, with a particular focus on changing patterns of home ownership.

We look at:

1). Cross-time shifts in the age structure of home ownership.

2). Intergenerational correlations in home ownership.

3). How these fit into debates about movements in intergenerational mobility through time.

4). Making a start (given relatively big UK data shortcomings) on exploring intergenerational wealth correlations.

Structure 1). Cross Cohort Changes in Home Ownership

2). Intergenerational Patterns of Home Ownership 3). The Importance of Parental Home Ownership for the Cross Cohort Changes 4). Implications for Falling Intergenerational Mobility 5). Intergenerational Transmissions of Wealth 6). Conclusions

Cross Cohort Changes in Home Ownership 1 Percent owner occupiers by age group and year 81 78 77

80

Percent Owner Occupier

74

76

79 71

70 66

64 59

60 50

40

77

70 69

70

50

76

47

45 35

30 20

20 10 0 20-29

30-39

40-49

50-59

60-69

Age Groups 1993

2000

2007

Source: Survey of English Housing/English Housing Survey

2013

Cross Cohort Changes in Home Ownership 2 Percent owner occupiers by age and cohort of birth 78 79

77

80

75 76 67

Percent Owner Occupier

70

70

60

55

58

50 40 29

30 20 10 0

1930s

1940s

1950s

1970s

1960s

1980s

Cohorts 20s 60s

30s

40s

Source: Survey of English Housing/English Housing Survey

50s

Cross Cohort Changes in Home Ownership 3 Cohort effects from regression models conditioning on cohort, age and time 1958 set to 0 0

-10 -20

-30

-40

-50 1924

1934

1944

1954 1964 Cohort

Year and age controls Plus income

1974

1984

1994

Plus family and individual Xs

Note: Cohort effects are derived for each year and presented as 3 year moving averages, with 1958 normalised to 0. Family and individual controls are gender, marital status, ethnicity and number of children.

Income is calculated as quintiles within each survey year.

Cross Cohort Changes in Home Ownership 4 2000 to 2007

1993 to 1999

Cohort

Cohort

20, 86, 1931

20, 87, 1924 20, 86, 1987

e Ag

e Ag

20, 87, 1979

69, 86, 1931

69, 87, 1924

% Owner Occupier

20, 0, 1924

20, 0, 1931

% Owner Occupier

69, 86, 1987

69, 87, 1979

20, 0, 1987 20, 0, 1979 69, 0, 1931 69, 0, 1924 69, 0, 1987 69, 0, 1979

2008 to 2015 Cohort 20, 85, 1939 20, 85, 1993

e Ag

20, 2, 1939

% Owner Occupier

69, 85, 1939 69, 85, 1993

20, 2, 1993

69, 2, 1939 69, 2, 1993

Notes: Data points are in the format [age, owner occupation, cohort], e.g. age 40 with owner occupation of 76 percent for birth cohort 1960 = [40, 76, 1960].

Rapid increases in house prices, most particularly since late 1990s FTBs gross house price to earnings ratio 10.0 9.0

House price to earnings ratio

8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

2007

2009

2011

2013

2015

Year

North

14/06/2017

London

UK

9

Intergenerational Patterns of Home Ownership 1 Cohort data for those born in 1958 and 1970 provides information on both parents’ and children’s home ownership (this is rare) as well as more standard variables.

We have previously shown that intergenerational income correlations strengthened between these cohorts National Child Development Study (babies born in a week in 1958).

British Cohort Study (babies born in a week in 1970).

Started with around 18,000 observations. Followed up periodically.

Mobility literature has concentrated on social class at 11/10 and family income at 16 and related these to adult measures.

We now add information on housing tenure.

Outcomes are measured at 42, in 2012 for most recent cohort.

Intergenerational Patterns of Home Ownership 2 Owner occupancy at age 42 by parents’ owner occupancy at age 11/10

NCDS, Age 42 in 2000

BCS, Age 42 in 2012

Percent Owner Occupiers

Percent Owner Occupiers 90

90

88

80 70 60 50

59

OO Age 16

Not OO Age 16

OO Age 16

Not OO Age 16

0

10

20

30

40

mean of oog

60 50 40 30 20 10 0

mean of oog

70

80

80 74

Intergenerational Patterns of Home Ownership 3 Intergenerational models of owner occupation NCDS, Age 42 in 2000

BCS, Age 42 in 2012

Cross Cohort Change

Parental home ownership at 11/10 Unconditional (1) Plus gender, parental age, with natural/adoptive father at 11/10

0.127 (0.008) 0.124 (0.008)

0.200 (0.010) 0.190 (0.010)

0.073 (0.012) 0.066 (0.013)

(2) Plus standardised maths and reading score at 11/10

0.095 (0.008)

0.163 (0.011)

0.068 (0.013)

9618

8402

Unconditional

0.140 (0.008)

0.217 (0.013)

0.077 (0.015)

(4) Plus gender, parental age, with natural/adoptive father at 11/10

0.139 (0.008)

0.211 (0.013)

0.072 (0.015)

(5) Plus standardised maths and reading score at 11/10

0.114 (0.009)

0.185 (0.013)

0.071 (0.015)

8375

6267

Sample Size Parental home ownership at 16

Sample Size

The Importance of Parental Home Ownership for the Cross Cohort Changes 1 Parental home ownership is becoming more important, and the increase is sizable: 1). In terms of odds ratios, the likelihood of a cohort member owning their home if their parents did compared to a cohort member whose parents did not rises from 147 percent more likely to 191 percent more likely, or a 44 percent increase.

2). In regression models of home ownership including family background variables, parental home ownership is a more important determinant of home ownership than having parents from a professional or high income background.

Implications For Falling Intergenerational Mobility 1 Prior to this investigation, there were two known facts about the evolution of social mobility across these cohorts.

First, that income mobility fell, and second, that social class mobility remained static (see, inter alia, Blanden, Gregg and Macmillan, 2013, and Erikson and Goldthorpe, 2010).

For some, these two findings have proven hard to reconcile. Issues that have been considered (a non-exhaustive list) include whether differences are due to measurement error, changing income dynamics within social class, or if they are down to fundamental conceptual differences.

Implications for Falling Intergenerational Mobility 2

Comparison of intergenerational models of owner occupation, social class and earnings/income NCDS, Age 42 in 2000

Home ownership at 42 and 16

Sample size Social class at 42 and 11/10

Sample size Log(Earnings at Log(Income at 16) Sample size

42)

and

BCS, Age 42 in 2012

0.139 (0.008)

0.211 (0.013)

8375

6227

0.072 (0.005)

0.067 (0.006)

7926

6394

0.266 (0.029)

0.354 (0.026)

4340

3389

Cross Cohort Change 0.072 (0.015)

-0.004 (0.008)

0.088 (0.039)

Intergenerational Transmissions of Wealth 1 Home ownership is an important component of overall wealth.

Indeed, for individuals in middle age it is very often the largest component (excluding pension wealth).

It therefore seems likely that the stronger association between parental home ownership, parental income and later home ownership that we have uncovered in the second cohort is indicative of a strengthening link between parental background and later wealth.

Can explore this using a number of variables which are proxies for wealth, or sub-components of a total wealth measure (as do Charles and Hurst, 2003, in the US using the PSID).

Intergenerational Transmissions of Wealth 2 Comparison of intergenerational models of wealth proxies, (1) Maximum samples

Home ownership at age 42 on: Parental home ownership at 16

Sample size

Log(Parental Income at 16) Sample size Log(Family Income) at 42 on: Parental home ownership at 16

Sample size

Log(Parental Income at 16) Sample size

NCDS, Age 42 in 2000

BCS, Age 42 in 2012

0.139 (0.008)

0.211 (0.013)

8375

6267

0.108 (0.013)

0.165 (0.013)

6439

4822

0.202 (0.019)

0.343 (0.025)

7345

4794

0.262 (0.028)

0.410 (0.025)

5708

3722

Cross Cohort Change

0.072 (0.015)

0.057 (0.018)

0.140 (0.032)

0.148 (0.026)

Intergenerational Transmissions of Wealth 3 Comparison of intergenerational models of wealth proxies, (2) Maximum samples

Having savings/investments at 33 on: Parental home ownership at 16

NCDS, Age 42 in 2000

BCS, Age 42 in 2012

Cross Cohort Change

0.176 (0.016) 5296

0.044 (0.019)

Sample size

0.132 (0.011) 7028

Log(Parental Income at 16) Sample size

0.108 (0.016) 5413

0.182 (0.016) 4059

0.074 (0.022)

0.016 (0.005) 6282

-0.009 (0.005)

Sample size

0.024 (0.005) 7206

Log(Parental Income at 16) Sample size

0.027 (0.007) 5542

0.013 (0.005) 4827

-0.014 (0.007)

Owing a business at 42 on: Parental home ownership at 16

Notes: All models control for gender, parental age, and whether the child lived with their natural/adoptive father at 11/10. Standard errors in parentheses.

Potential Mechanisms 1) Human capital transmissions This drives a lot of the change in intergenerational income mobility.

2) Transmissions of preferences Not investigated yet

3) Direct inheritance Data is annoying inconsistent

4) House price rises enabling parents to transfer capital to children Some evidence for this hypothesis, but more work needed.

200 0

100

Price ÂŁk

300

400

Regional house price changes

1970

1980 North North West West Midlands Greater London South West Scotland

14/06/2017

1990 Year

2000

2010

Yorkshire and Humberside East Midlands East Anglia Rest of South East Wales

20

Regional differences

NCDS, Age 42 in 2000

14/06/2017

BCS, Age 42 in 2012

Cross Cohort Change

North West

0.160 (0.033)

0.241 (0.037)

0.081 (0.050)

North

0.106 (0.024)

0.200 (0.040)

0.094** (0.044)

East and West Riding

0.150 (0.031)

0.159 (0.042)

0.010 (0.032)

North Midlands

0.120 (0.029)

0.262 (0.050)

0.143** (0.030)

Midlands

0.136 (0.036)

0.260 (0.040)

0.124** (0.055)

East Anglia

0.129 (0.026)

0.221 (0.043)

0.093 (0.049)

London and the South East

0.153 (0.029)

0.271 (0.036)

0.118** (0.046)

Southern

0.133 (0.022)

0.250 (0.061)

0.111* (0.055)

South West

0.139 (0.030)

0.172 (0.051)

0.033 (0.057)

Wales

0.113 (0.034)

0.165 (0.050)

0.051 (0.061)

Scotland

0.164 (0.034)

0.162 (0.036)

-0.003 (0.045)

21

Conclusions 1). There has been a very rapid deterioration of the prospects of home ownership facing more recent birth cohorts, and a key feature of that has been whether individuals grew up in a home owned by their parents or not.

2). The strengthening of the intergenerational home ownership relationship across time adds more to the (sometimes controversial) academic debates that have taken place about whether or not social mobility fell recently in Britain.

3). The focus on the housing market allows adds evidence that the intergenerational transmissions of wealth strengthened.

4). Further research will expand focus upon the mechanisms that may lie behind the changing link between family background and home ownership (e.g. inheritance; “Bank of Mom and Dad�; changing pattern of intergenerational transmission preferences).