SLIDE 1
Inequality and Surnames
Gregory Clark University of California, Davis and LSE
Simple Measure of Social Mobility
Correlation between parents and children
- n any measure of social status
Surnames Gregory Clark University of California, Davis and LSE - - PowerPoint PPT Presentation
Surnames Gregory Clark University of California, Davis and LSE - - PowerPoint PPT Presentation
Inequality and Surnames Gregory Clark University of California, Davis and LSE Simple Measure of Social Mobility Correlation between parents and children on any measure of social status 0 implies complete mobility 1 implies
SLIDE 2
SLIDE 3
Social Mobility and Inequality closely linked
For a given set of shocks to income or
wealth each generation (e), long run distribution of outcomes depends on intergenerational correlation b
𝑊𝑏𝑠𝑗𝑏𝑜𝑑𝑓 𝑧
=
𝑤𝑏𝑠𝑗𝑏𝑜𝑑𝑓(𝑓) 1−𝑐2
SLIDE 4
Conventional Picture of Social Mobility
SLIDE 5
Implications, conventional estimates
Mobility Rates High
Share of social status variation inherited low – 4% Scandinavia, 22% USA
SLIDE 6
Implications, conventional estimates
There is a social mobility problem.
Mobility rates too low in some societies. Enormous human potential squandered. Social Democratic Nordic countries are achieving faster social mobility than free market USA
SLIDE 7
Implications, conventional estimates
What matters to social success? Culture Education Social networks Not genetic inheritance of abilities
SLIDE 8
Limitations
Looks just at one generation Looks just at individual aspects of status
SLIDE 9
Surname Method
Measure social mobility by tracing status by
surname lineages – e.g. Clark, Smith, Bazalgette
Surnames link us to previous generations
through the patriline – in England we can link some people to their ancestors of 1066 – 32 generations
E.g. Norman surnames Montgomery,
Baskerville, Punchard
SLIDE 10
Surname Method
Based on current mobility measures common
surnames should rapidly lose status information
The rate at which they loose that information –
the rate of social entropy - can be translated into the intergenerational correlation of status
Surnames track one line of descent, but that line
assumed representative (assortative mating)
SLIDE 11
Surprising and universal finding
Surnames move to average status very slowly – 10-
15 generations
Implied intergenerational correlation of status 0.7-
0.8
SLIDE 12
Table 2: Rare Oxbridge versus non-Oxbridge Surnames, 1800-29 Oxbridge Non-Oxbridge Agassiz Brickdale Agnerv Bodgett Anquetil Brooshooft Allbert Boolman Atthill Bunduck Arfman Bradsey Baitson Buttanshaw Bainchley Breckill Barnardiston Cantis Bante Callaly Bazalgette Casamajor Barthorn Capildi Belfour Chabot Bavey Carville Beridge Charretie Bedborne Cavet Bleeck Cheslyn Bemond Chanterfield Boinville Clarina Berrton Chesslow Boscawen Coham Bideford Chubham Bramston Conyngham Bisace Clemishaw
SLIDE 13
Relative Oxbridge Attendance for Elite Rare Surnames 1800-29
1 2 4 8 16 32 64 1530 1590 1650 1710 1770 1830 1890 1950 2010 Relative Representation
0.77
0.82
0.83 0.77
SLIDE 14
Family vs Surname Wealth Correlation, England
Period of Death Individual Wealth Correlation Surname Wealth Correlation 1888-1917 0.48 0.71 1918-59 0.41 0.69 1960-87 0.41 0.73 1999-2012 0.46 0.83
SLIDE 15
Patriline – what about daughters?
SLIDE 16
Assortative Mating Pre-1880 (richer lineages) – Wealth, England
SLIDE 17
Example of surprising persistence of status – Darwin great-great-grandchildren
10 children, but only 27 great-
great-grandchildren
11 notable enough to have
Wikipedia pages/Times Obits
6 university professors, 4 authors,
a painter, 3 medical doctors, a well-known conservationist, and a film director
SLIDE 18
Countries
England, 1300-2012 Sweden, 1700-2012 USA, 1920-2012 Chile, 1950-2012 India, 1860-2012 Japan, 1860-2012 China, 1650-2012 Taiwan, 1949-2012 Costa Rica, 1950-2014 Australia, 1870-2014 Hungary, 1860-2017 Russia, 1879-2017 Barcelona, 1500-1860
SLIDE 19
Intergenerational correlation measured through surnames
High 0.7-0.8 Little variation across societies and epochs
SLIDE 20
Conventional versus Surname Estimates of Status Persistence, 1950-2012
0.2 0.4 0.6 0.8 1 Intergenerational Correlation
SLIDE 21
Sweden as another example
Elite surnames from 1600-1800 Counts/Barons Untitled Nobility Latinized Surnames
SLIDE 22
Conventional Picture of Social Mobility
SLIDE 23
Riddarhuset, Stockholm
SLIDE 24
Figure 3: The History of Ennoblement in Sweden Source: Almenberg and Dreber, 2009, 178. 1,000 2,000 3,000 1600 1700 1800 1900 2000 Number of Families Nobles Created Net Stock Families
SLIDE 25
Aristocratic Surnames
Domestic - embodying status elements such as
Gyllen (gold), Silfver (silver), Adler (eagle), Leijon (lion), and Ehren (honor)
Leijonhufvud Gyllenstierna Oxenstierna Ehrensvärd
SLIDE 26
Latinized Surnames
Celsius Aquilonius Arrhenius Boethius Bruzelius Cnattingius
SLIDE 27
Representation of Surname Types Among Doctors, 1890-2011
1 2 4 8 1900 1920 1940 1960 1980 2000
Relative Representation
Titled Nobles Other Nobles Latinized Correlation 0.72
SLIDE 28
Elite Surnames in the Swedish Royal Academies
0.015625 0.0625 0.25 1 4 16 64 1740 1770 1800 1830 1860 1890 1920 1950 1980 2010 Relative Representation Cohort Elite Correlation 0.87 ..son Correlation 0.87
0.87
0.87
SLIDE 29
Summary Surname b Estimates by Period, Sweden
Group 1700-1900 1890-1979 1950-2012 Attorneys
- 0.71
Physicians
- 0.67
0.88 University Students 0.78 0.85 0.66 Academicians 0.89 0.75 0.84
SLIDE 30
Hungary, 1946-2017
0.3 0.5 1.0 1945 1955 1965 1975 1985 1995 2005 2015 Elite Surname Mean Generation
..y versus Top 20 Fit v top 20
ρ = 0.86
SLIDE 31
Why are the conventional and surname results so different?
Conventional estimates focus on individual
aspects of status
Surnames are capturing what happens to
underlying overall status
SLIDE 32
Social Status Across Multiple Generations
SLIDE 33
SLIDE 34
For each individual:
Status phenotype – measured status on
variety of aspects
Underlying status genotype – status that
is transmitted to next generation – can be inferred from the status of your lineage
SLIDE 35
Surname Estimates
Long run social mobility Social mobility of social groups – ethnic, racial,
religious, immigrants
SLIDE 36
More Fundamental Question – what transmits social genotype?
Family Resources? Family Culture? Social Networks? Genes?
SLIDE 37
Surprising Evidence – most social status transmission is genetic
Patterns of inheritance Adoption studies Groups that marry endogamously How elites get formed Shocks to family size in England 1800-1880
SLIDE 38
Social status USA 2012 – Doctors per 1000
2 4 6 8 10 12 14
Relative Representation
SLIDE 39
Conclusions
There is equality of opportunity in most
societies
Most social ability biologically inherited –
and gets rewarded
This is not a pessimistic result But it is an argument for limiting
inequalities.
SLIDE 40