Zolt

1
What are the determinants of intergenerational
mobility in Hungary? Comparing municipalities
and periods before and during modernization

• Social Stratification Research Seminar 2010
• Utrecht, the Netherlands
• 10 09 2010
• Zoltán Lippényi
• ICS/Utrecht University

2
Comparative intergenerational mobility back to
the past

• Towards open societies project (Marco van
  Leeuwen,
• Ineke Maas, ICS/UU/IISH)
• Long-term intergenerational mobility in several
  historical contexts
• Is there a trend towards increasing
  intergenerational occupational mobility?
• Could changes in occupational mobility be
  explained by industrialization, urbanization, and
  institutional development?

3
Intergenerational mobility in Hungary, 19th-20th
Century

• Previous research communism, transition 1989
  (Szelényi, 1998 Bukodi Róbert, 2004)
• modernization increases from second half of
  the 19th Century (Berend Ránki, 1982)
• Austro-Hungarian Monarchy (1867-1918)
• agricultural society with large farming
  underclass
• immobility of agricultural classes decline
• regional and municipal differences in
  modernization (Beluszky Gyori, 2005)
• growing distances on the urban ladder

4
Collecting mobility data from the past

• Data collection 60 Hungarian municipalities
• Marriage records (1850-1950)
• Sampling town and village typologies based on
  data from the 1930 Hungarian census
• Two-stage selection
• 1. towns from different macro-regions
• 2. two-three villages from the towns
  micro-region

5

• Municipal centers industrial centers agrarian
  centers

6
Preliminary results

• Kalocsa (Mid-Hungary) and its rural outskirts
• 15-20.000 residents, primarily agrarian
• Marriages 1895-1950
• men, aged 18 -40
• N 2,247

7
(No Transcript)
8
Data quality - coverage

• Population Kalocsa residents who lived in the
  town at the time of their marriage
• Unit non-response
• Never got married
• Hajnal-line marriage almost universal in most
  Hungary (Hajnal, 1965)
• Married outside church
• Secularization began in the 1950s
• Married in another city
• Hungary joint household formation system
  (Hajnal, 1982)
• Item non-response
• occupation of the father missing
• Fathers early death Van Poppel and Van Gaalen
  (2008) found no social status effect

9
Data quality reliability

• Comparability of fathers occupation with that of
  the son
• different life stage at time of measurement
  overestimating downward mobility/immobility
• Class differences in marital age could confound
  with change over time in mobility
• E.g. sons of farmers marry at older age
  increasingly over time relative to other classes
  ? greater mobility from agricultural origin
• Additional analyses famers and farm workers
  marry earlier, but no time effect and no
  origin-destination classtime effect on age at
  marriage
• Quality of occupational measurement
• more detailed titles ? more mobility observed

10
Occupations

• Occupational origin and destination
• HISCO (Van Leeuwen, Maas, Miles, 2002)
  historically comparable occupational coding
  scheme
• HISCLASS 6 categories
• Higher managers and professionals
• Lower managers and professionals
• Foremen and skilled workers
• Farmers
• Lower/unskilled workers
• Lower/unskilled farm workers

11
(No Transcript)
12
(No Transcript)
13
Methods and model specification

• Multinomial conditional logit model (Breen,
  1994 Hendrickx Ganzeboom, 1998 Wu Treiman,
  2007)
• Quasi-equal row and column effects 2 model
• Equal scaled metric for origins and destinations
• Origin-destination association
• Scaled association parameter
• Immobility
• Overall immobility parameter (diagonal)
• Immobility parameter for agricultural classes
• Control covariates
• Immigrants
• Rural outskirts

14
Model selection
Model LR chi2 (diff) Df (diff) P BIC (diff)
Baseline class-specific intercepts association immobility covariates 3366 11 4790
M1 Baseline time-varying intercepts -199.5 5 .000 -152
M2 M1 time-varying association and immobility parameters -33.35 3 .000 -5
M3 M2 time-varying association and immobility, separate for outskirts -7.36 3 0.06 21
M4 M1 time-varying immobility parameters -33.24 2 .000 -15
M5 M1 time-varying overall immobility -32.28 1 .000 -23
15
Parameter estimates M5Scaled metric
Higher managers and professionals -0.63
Lower managers/ professionals -0.29
Foremen/skilled workers -0.07
Lower skilled/unskilled workers -0.02
Farmers 0.50
Farm workers 0.51
16
Parameter estimates M5
Origin-destination association 4.56
Immobility 1.55
Agricultural immobility .86

Time Higher managers and professionals -.001
Time Lower managers/ professionals -.004
Time Foremen/skilled workers -.027
Time Lower skilled/unskilled workers REF
Time Farmers -.048
Time Farm workers -.062

Immobilitytime -.02
17
Modernization and mobility

• Indicators collected from gazetteers and census
  sources
• Population size (per 100 inhabitants)
• 7 censuses 1890, 1900, 1910, 1921, 1930, 1940,
  1948
• Other years linear interpolation
• Yearly modernization-index (per 1000 inhabitants)
  
• number of institutions in a given year (bank,
  hospital, tax office)
• number of schools
• number of larger industrial establishments
• High correlations separate models (time excluded)

18
Results

• Model 5 Baseline main effects immobility
  interaction

MODERNIZATION
Modernization immobility -.30
LR2 (17) 3510, plt.000, BIC4703

POPULATION
Populationimmobility -.03
LR2 (17) 3517.17, plt.000, BIC4697
19
Conclusion

• Evidence of changing mobility regime before
  communism declining immobility (1.3 per year)
• Similar decline in immobility for agricultural
  and non-agricultural classes
• No change in overall association and difference
• between city and its rural outskirts
• Modernization and population size decreasing
  inheritance

20
Future plans

• More contexts municipal level variation in
  modernity
• Multilevel modeling combine MCL-estimates from
  several municipalities in meta analysis
• Comparability of vital records with
  representative survey data for the same period
• Urban centers and villages in their
  micro-regions residential and intergenerational
  mobility
• Hungary, a multiconfessional land
  intergenerational mobility and religion

21
Thank you for your attention!
22
Controls and additional model specifications
Immigrant (Stereotyped ordered effect)
Outskirt residentagricultural occupation
Outskirt residentindustrial occupation _
Father dead at time of marriage n.s.
Aged 14 between 1914 and 1924 (1st World War) n.s.
Quadratic time-specification n.s.
Leave out post 1945 marriages identical results
23
The statistical model

• logit(pj/pk) aj ak (fj fk)(µ0 ? µtXt)
  fv (ßjt ßkt)Xt
• ? fj0, ? fj2 1
• Stereotyped ordered effect
• (fj fk) ? ßmXm
• Estimation Iteratively fj scaling metric ?? ß
  parameters