Transcript On intergenerational mobility in Italy

On intergenerational mobility in Italy
What a difficult future
Virtual Presentation Symposium Programme
3-rd International Symposium
Shaping Europe 2020: socio – economic challenges
Bucharest, 15th – 16th November 2013
Federica Roccisano
Catholic University of Milan
• Introduction
• From intergenerational transmission of poverty
• …To intergenerational mobility
• Taxonomy
• Methodology
18/10/2013
Summary
• Case study
2
Today the Developed Countries, like the members of the
European Union, are heavily prejudiced by numerous factors, like
the population aging or the slow birth rates.
In this paper we will analyse an important problem linked to this
situation: the redistribution of income (intergenerational
mobility).
Our area of study is one of the most problematic countries in
Europe: Italy. To examine the evolution of intergenerational
mobility in this Country, we will use data from the Survey of
Household Income and Wealth (SHIW). For the analysis on the
intergenerational transmission of poverty we’ll use data from the
Intergenerational Module of the EU SILC 2005 Module on
Intergenerational transmission of poverty and the EU-SILC 2011
Module on Intergenerational Transmission of Disadvantage.
18/10/2013
Introduction
3
The trend of intergenerational equity has changed following the
evolution of the society during the first half of the 20th Century:
before of the two world wars the grandparents of those born in
1940s shared many of the same experiences with their children;
while for whom born in 1960 changes in work, employment and
politics have produced a lot of benefits.
The worst situation regards people born in 1980 and 1990 when
workers began to leave their job in increasing numbers and at
earlier ages: poverty rates amongst the elder declined while
younger households and opportunities rose.
18/10/2013
From Intergenerational transmission of poverty…
4
The intergenerational transmission of poverty is from poor parents to
poor children when the living condition, the endowments and the
investments on education of parents are not able to get better the socio
economic status of their son.
That is the so called generational bargain: “the basic idea of the intergenerational bargain is a simple one: in all ‘communities’, from family
to globe, there are relationships for the transfer of resources between
generations and these relationships carry with them often un-codified
‘rights’ and obligations”.
Unfortunately it does not depend just on individual motivations
(altruistic or solidaristic), but it is also subject to the Welfare System, the
economic conjuncture and the functioning of the societal structures and
institutions, all of which can drive inequality and what is transferred to
next generation.
18/10/2013
The intergenerational bargain
5
Tab. 1: The Moore Approach to Intergenerational Transmission of Poverty
How is it transmitted
Financial,
Material,
Environmental
Capital:
Cash
Land
Debt
Common Property resources
Insurance, pensions
Bequests, dispossession
Bride wealth
Environmental conservation/degradation
Labour bondage
Human Capital:
Educational qualifications, knowledge,
skills, coping/ survival strategies
Good mental/physical health
Disease, impairment
Intelligence?
Socialisation
Investment of time/capital in
education/training; health/nutrition
Contagion, mother-to-child transmission
Genetic inheritance
Social, Cultural, Political Capital:
Traditions,
institutions,
norms
of
entitlement, value systems
Position in community
Access to key decision-makers, patrons,
organisations
‘Cultures of poverty’?
Which factor affect transmission
Demographic factors: household
structure, broader process of
fertility transmission
Nature of guardian: education and
skill level
Social,
cultural,
legal
and
governance related factors: norms
Economic Factors: labour market
Nature of living space: stigma,
care;
sense of community
18/10/2013
What is transmitted
Socialisation and education
Kinship
Locality
Genetic inheritance
6
The investigation on intergenerational mobility, was made first by
sociologists. The pioneers have studied intergenerational social mobility
on the basis of correlations of parents’ and children’s “socioeconomic
status” score. While in the last decade sociologists have deepen mostly
the persistence between parents and children’s outcomes.
Recently, also many economists have demonstrated the strong presence
of intergenerational transmission of economic status. The most reliable
reason of this connection is the job of head of household, but many
researchers, like sociologist or psychologists, have underlined the role of
the “cultural inheritance” and also the environmental and genetic
connectedness: the so called influence of Nature and Nurture. But
Zimmerman in 1992 has demonstrated that regression estimates by the
Nature and Nurture’s devotees are not capable of capturing linkage
between genetic endowment and economic status.
18/10/2013
…to Intergenerational Mobility
7
Authors
Country and data set
Measures of income
Sample size
Age of sons
Solon (1992)
US
PSID
1) Annual Earnings
2) Hourly Wage
3) Family income
348 fathers-son pairs 25-33
Estimate of β
Comments
1) TA: 0.41
1) IV: 0.53
S. showed that the TA-technique
provide a downwards biased estimate
and the IV technique an upwards bias.
2) IV: 0.45
3) IV: 0.53
Zimmerman (1992)
US
NLS
1) Wage + salaries
2) Hourly Wage
3) Duncan Index of status
876 fathers-son
pairs, but fewer in
most estimations
29-39
1) TA: 0.54
1) MM: 0.41
2) TA: 0.39
2) MM: 0.38
The presented estimate are elasticities,
which are close to the correlations
that are reported in the paper.
Z. also presented IV estimates that are
close to those obtained by TA and MM
techniques.
3) TA: 0.33
Dearden et al.
(1996
Britain, National
Child Development
Survey
Weekly wages
1565 pairs of fathers 31
and sons
IV: 0.59
TSIV: 0.39
The presented estimates are elasticities.
Corak & Heisz
(1999)
Canada , register
data
1. Annual earnings
2. Annual market income
≈ 350 000 pairs of
fathers and sons
28–31
1) TA: 0.13
2) 0.19
The estimates are elasticities.
Non-linearities implying greater
mobility at the lower end of the
distribution were found.
Jäntti & Österbacka Finland, register
(1995)
data
Annual earnings
22 324 pairs of
fathers
and sons
Average age:
34.8
TA: 0,20
Björklund &
jäntti
(1997)
1. Annual earnings
2. Market income (incl.
income of capital)
400 sons, 500 fathers 29–38
Sweden, Level of
Living Surveys
United States,
PSID
Annual earnings
About the same as
Solon
1) TSIV: 0.23
2) TSIV: 0.29
The Swedish data set lacks information
on fathers’ age which is not controlled
for in the estimations of neither the
Swedish nor the US ρ.
Adding such controls for the US
raises the correlations by around
0.05.
8
28–36
TSIV: 0.33
The method and data set differ
slightly from the one used by Solon,but
is the same for both countries.
Author
Country and dataset
Measure of income
Sample Size
Age of sons
Estimate of β
Methodology
Comments
Comi S.,
(2003)
12 EU Countries
European Community
Household Panel
1. Father earnings a
15011 pairs
By Country:
20-25
OLS
1) Classical equation about income
mobility by Solon
The link between father and son earnings is relatively
high in Italy, Belgium and Portugal.
Germany
Denmark
Netherland
Belgium
France
UK
Ireland
Italy
Greece
Spain
Portugal
Austria
Bjorklund A., Jantti Norway
M
Register data
(2000) et(2005)
Denmark
Register data
Blanden J.
(2005)
2263
450
823
395
797
1169
1476
1788
723
1852
1924
1354
0.18
0.09
0.067
0.21
0.12
0.10
0.03
0.27
0.16
0.17
0.20
0.02
34-41
0.14
1) Classical equation about income
mobility by Solon
38-44
0.14
Sweden
Register data
34-43
0.14
2) Transition matrix: the information in
the matrix would be able to tell us more
about the kind and direction of mobility
that is occurring. But this method
requires long-run incomes of both sons
and fathers.
Finland
Quinquennial census
panel
35-42
0.15
UK
British Cohort Study
1970
30
0.27
USA
Panel Study of Income
Dynamics
30
0.29
West Germany
Socio Economic Panel
37-40
0.17
31-28
0.14
Canada
Intergenerational
income data
Parental income 1980
and 1986 (average)
With the aim to understand more about the relation
between father and son she considers the
intergenerational Educational mobility.
They consider:
- International comparison
- Approach in sociology
- Class mobility
- Mobility of status (rif to Duncan)
Education has been often seen as a route to greater
intergenerational mobility.
9
Author
Country and dataset
Measure of income
Sample Size
Piraino
(2006)
Italy
SHIW
Income from labour
Age of sons
Estimate of β
Methodology
Comments
612 fathers-son 30-45
pairs
TS2SLS: 0,50
PI: 0,37
Matters of family background and educational
attainents!
231
fathers-son
pairs
1) Classical equation about income
mobility by Solon
TS2SLS Estimator
Predicted Income
Co-residing
CR: 0.35
2) Transition Matrix by income classes
Mocetti S.
(2007)
Italy
SHIW
Earnings
Nolan
(2012)
EU
EUSILC
1. Financial Distress
2. Income Poverty
3. Deprivation
4900 fathers
3200 sons
30-50
1): 0.50
2): 0,61
1) Classical equation about income
mobility by Solon
TS2SLS Estimator
The degree of intergenerational income mobility in
Italy is lower than that observed in other developed
countries.
Education obviously plays a crucial role in explaining
2) Quintile regression can provide a more social outcomes and in accounting for long term
complete statistical analysis of the
mobility.
intergenerational relationship across the School decisions are affected by family background.
distribution of sons’income.
Parental education and socioeconomic status appear
2SQR Estimator
to be the main determinants of educational choice,
and this reinforces intergenerational immobility.
About the occupational immobility: in some cases, it
is linked to the existence of entry barriers limiting
access to certain professions, or to the
intergenerational handing-down of control of the
family …
firm. In other cases, it is the natural outcome
of educational stratification.
“The EU-SILC Intergenerational Module appears to
offer an unprecedented opportunity to conduct a
comparative analysis of the relationship between
current poverty and social exclusion outcomes and
parental characteristics and childhood economic
circumstances. However, as our analysis reveals,
serious problems relating to the scale of missing
values and major reservations about the
comparability of key variables means that the results
of any such analysis must be treated with
considerable caution.”
10
About methodology:
three milestones
Paper Information
Zimmerman
Regression Toward
Mediocrity in
Economic Stature
The American
Economic Review
Vol. 82, No. 3 (Jun.,
1992
Methodology
18/10/2013
Author
11
Paper Information
Solon G.
Intergenerational
Income mobility in
the US, The
American
Economic Review
1992
Methodology
18/10/2013
Author
12
18/10/2013
The income elasticity
Where yd is the vector (in log terms) of the father’s permanent incomes while ys is the
vector of son’s permanent incomes.
The coefficient β indicates the rate of the intergenerational elasticity and his value
varies between 0 and 1. If β is high we will have a very strong impact of parental
outcomes on children’s economic status and so high level of intergenerational inequality
and less intergenerational mobility.
13
A recent alternative to the elasticity β is the intergenerational correlation (ρ) or the
correlation between the log earnings of the two individuals (father and son) that is
equals to the elasticity only if the standard deviation σ of log earnings is the same for
both generations:
18/10/2013
The intergenerational
correlation
14
At the beginning researchers like Becker and Tomes, studying the
condition of United States, expected that the better value of the
intergenerational elasticity should be 0,2 or less.
But after some years Solon and Zimmerman showed how it is
possible to talk about a mobile society also if the value of β is
bigger: considering the average of income over some years (from 4
to 10) it is possible to have a better estimation of permanent
income capacity and the value of intergenerational elasticity could
be also 0,4.
More recent studies, founded another time on the US condition,
argue that the most preferable value of β is around 0,6.
18/10/2013
About the preferable value of the intergenerational mobility
15
Country
Dataset
Sons Born
Sons Earnings Measure
Measure of Parental Status
Value of β
Britain
British Cohort Study
1970
2000 (Age 30)
Parental Income Average 1980-86
0.271
US
Panel Study of Income Dynamics
1954-70
Age 30
Parental income when son were 10-16
(average)
0.289
West Germany
Socio-Economic Panel
1960-73
2000
Parental Income 1984-88 (average)
0.171
Canada
Intergenerational Income Data
(from tax register)
1967-70
1998
Parental Income when son aged 16
0.143
Norway
Register Data
1958
1992 and 1999 (average)
Father’s earnings 1974
0.139
Denmark
Register Data
1958-60
1998 and 2000
Father’s earnings 1980
0.143
Sweden
Register Data
1962
1996 and 1999
Father’s earnings 1975
0.143
Finland
Quinquennial Census
1958-60
1995 and 2000 (average)
Father’s earnings 1975
0.147
18/10/2013
International Comparable Estimates of Intergenerational Mobility
16
18/10/2013
CASE STUDY
The situation of Italy
17
The actual situation of Italy is really difficult because of
the global financial crisis, the high level of public debt
and unemployment and the low level of GDP growth
rate. As we can said in the previous page all the
institutions (state, market and community) concur to
determine the intergenerational transmission of poverty.
18/10/2013
«Poor young»
18
Teenage in family with financial problems ordered by (*)
Our elaboration from EU-SILC Module 2005
Mostly Often
Occasionally
1. Denmark
9,6
14,5
2. Norway
9,3
15,3
3. Island
9,9
15,6
21. Italy
41,4
28,1
22. Slovenia
43,5
29,6
23. Slovakia
43,3
32,1
18/10/2013
(*)Rarely
75,9
75,4
74,5
30,5
27,9
24,6
Teenage in family describing the financial situation of the household.
Our elaboration from EU-SILC Module 2011
Italy
Value
Very bad
Bad
Moderately bad
EU 27 (Average)
Ranking
4,3
8,2
19,6
9°
17°
7°
3,9
8,6
16,9
19
Both Italy and Spain have not a complete panel with all the
information for at least two generations, like these used to study
the Country in table 3. To overcome this obstacle we decided to
follow the same method already applied by several researchers
who have studied the Italian case.
Hence, we will use the Survey of Household Income and Wealth
(SHIW) for Italy and the Household Budget Survey for Spain and,
since this two survey are too much short to obtain consistent
results, we will create two different samples for each Country and
proceed following the TS2SLS estimator (two-sample two-stage
least squares).
18/10/2013
Intergenerational Mobility: the model
20
Time-invariant and ususal disturbance
Time-invariant determinants
(geographical area, study level,
occupation, income)
18/10/2013
We have to construct a first sample with the information on our pseudo – fathers:
income, study level, occupation, geographical area. On this first sample we run a
regression:
Time-variant determinants (age)
It’s important to underline that we don’t consider gender information
because in our model we’ll not consider the income of the breadwinner but
just the income of fathers.
21
• Where Iˆ is the result of the first sample that allows us to replace
in the second sample missing fathers’ incomes with their best
linear predictions.
• We can synthetize all the disturbances and rewrite:
18/10/2013
• The second sample will comprehend the variables set of son in relation to
which one of pseudo-father. Our regression will be:
22
Pseudo-Fathers
(years 19841986)
Num
Mean Age
3.224
41,99
Sons’ report of
fathers
characteristics
(years 2008-2010)
786
41,22
Mean
LogWage
9,78
--
Mean Study
3,11
2,54
Mean
worksec
3,98
2,79
Mean
Workqual
2,65
3,15
Study :
1= no school 2= elementary school 3=lower
secondary school 4= high school 5=bachelor
6=specialization
Work Sec:
1= agriculture 2= industry 3= P.A. 4= commerce,
handcraft, services
Work Qual:
1= factory worker 2=employee 3=teacher
4=official 5=executive 6= freelancer
7=entrepreneur 8=self-employee 9=unoccupied
18/10/2013
Descriptive statistics for selected fathers
and sons
23
First-stage regression of pseudo-fathers income on five variables
N=3203
R2=0.1989
Variables
coefficient
Robust st. error
t
Study
0.106
0.061
1.75
YearStudy
0.025
0.014
1.72
WorkSect
-0.030
0.004
-7.30
WorkQual
0.034
0.004
8.31
Cons
9.387
0.068
138.55
18/10/2013
The Italian case
24
Second-stage regression with instrumental variables (2SLS) :
logwagehatpd= studydad worksecdad workqualdad
R2= 0,11
Variables
coefficient
Std. Error
t
logwagehatpd
0,57
2,067
0,27
Study
0,169
0,024
7,03
WorkSec
0,005
0,007
-5,18
WorkQual
-0,038
0,007
-2,86
Areageo
-0,091
0,032
-2,86
Cons.
4,125
20,255
0,20
18/10/2013
N=766
25
Mocetti (2007)
SHIW (2004 – 1977)
β = 0,50
Piraino (2006)
SHIW (2002 – 1977)
β = 0,47
Our elaboration (2013)
SHIW (2010 – 1984)
β = 0,44
18/10/2013
Summary
26
In this paper we try to describe the situation of young people in Italy, considering in
the firsts paragraphs the intergenerational transmission of poverty, and so the
perception of teenage about the financial situation of their family, while in the core
of the paper we deepen the intergenerational mobility.
What emerge is that to analyze the Italian situation is not easy because of the lack
of data and the difficulty to work on a survey that includes information from two
generations of the same family, like we have in the British Cohort Study or the Panel
Study of Income Dynamics of US. These differences make difficult also to compare
the intergenerational mobility rate of Italy with which one of other Countries.
Anyway we have try to minimize the disparity and to have much more complete
description of the Italian case we have compared our results with which one of
Mocetti and Piraino that in the last years have made a similar analysis of Italian
case. From this comparison we can say is that Italy is a not mobile society and that
occur to intervene to prevent the pauperization of young people.
18/10/2013
Concluding Remarks
27
18/10/2013
Thank you for your attention 
28