ArticlePDF Available

Wealth Inequality in Italy: Reconstruction of 1968-75 Data and Comparison with Recent Estimates

Authors:

Abstract and Figures

This paper provides a reconstruction of the joint distribution of Italian households’ income and wealth in the years ranging from 1968 to 1975. Exploiting the information available in some historical reports recently published by the Bank of Italy, the paper reconstructs synthetic microdata compatible with the aggregate results of sample surveys carried out in those years. In this way, inequality and poverty can be estimated by using the same statistical criteria that are used today, making an intertemporal comparison of the estimates possible. The concentration of household wealth shows a downward trend in the 1970s and ’80s, an increase in the years following the 1992-93 crisis and relative stability in the new century. In the period 1968-75 the concentration of wealth turns out to be greater than in recent years. The estimates of relative poverty show a decreasing trend until the 1990s and a subsequent increase; the upward trend of these indicators in recent years is steeper than that of the concentration indices. Migration flows have contributed significantly to the recent growth in the poverty indices.
Content may be subject to copyright.
Questioni di Economia e Finanza
(Occasional Papers)
Wealth inequality in Italy:
a reconstruction of 1968-75 data and a comparison with
recent estimates
by Luigi Cannari and Giovanni D’Alessio
Number
428
March 2018
Questioni di Economia e Finanza
(Occasional Papers)
Number 428 – March 2018
Wealth inequality in Italy:
a reconstruction of 1968-75 data and a comparison with
recent estimates
by Luigi Cannari and Giovanni D’Alessio
The series Occasional Papers presents studies and documents on issues pertaining to
the institutional tasks of the Bank of Italy and the Eurosystem. The Occasional Papers appear
alongside the Working Papers series which are specically aimed at providing original contributions
to economic research.
The Occasional Papers include studies conducted within the Bank of Italy, sometimes
in cooperation with the Eurosystem or other institutions. The views expressed in the studies are those of
the authors and do not involve the responsibility of the institutions to which they belong.
The series is available online at www.bancaditalia.it .
ISSN 1972-6627 (print)
ISSN 1972-6643 (online)
Printed by the Printing and Publishing Division of the Bank of Italy
WEALTH INEQUALITY IN ITALY: A RECONSTRUCTION OF 1968-75 DATA
AND A COMPARISON WITH RECENT ESTIMATES
by Luigi Cannari* and Giovanni D’Alessio*
Summary
This paper provides a reconstruction of the joint distribution of income and wealth
among Italians in the years ranging from 1968 to 1975. By exploiting the information
available in some historical reports recently published by the Bank of Italy, the paper
reconstructs synthetic microdata compatible with the aggregate results of sample surveys
carried out in those years. In this way, inequality and poverty can be estimated by using the
same statistical criteria that are used today, making possible an intertemporal comparison of
the estimates. The concentration of household wealth shows a downward trend in the 1970s
and 1980s, an increase in the years following the 1992-93 crisis and a relative stability in the
new century. In the period 1968-75 the concentration of wealth turns out to have been higher
than in recent years. The estimates of relative poverty, calculated by using both indicators of
equivalent income and indicators that combine income and wealth, show a decreasing trend
until the 1990s and a subsequent increase; the upward trend of these indicators in recent years
is steeper than that of the concentration indices. The poverty indicators that take wealth into
account have reached levels similar to those observed in the period 1968-75. Migration flows
have significantly contributed to the recent growth in the poverty indices.
JEL Classification: D31, D63, I32, C15.
Keywords: wealth, income, inequality, poverty, synthetic data.
Index
1 Introduction ......................................................................................................................... 5
2 Data and methods ................................................................................................................ 6
2.1 The data used ............................................................................................................... 6
2.2 The method for reconstructing synthetic microdata..................................................... 8
2.3 The Gini index estimates for 1977-1986 .................................................................... 11
3 Aggregate income and wealth trends since the post-war period ....................................... 12
4 Wealth inequality from the late 1960s to 2014 ................................................................. 14
4.1 The estimates from 1968 to the most recent years ..................................................... 14
4.2 Wealth and under-reporting ....................................................................................... 18
4.3 Inequality in the income-net worth indicator ............................................................. 19
4.4 A long-term look ........................................................................................................ 20
5 Indicators of poverty from the 1960s to the present ......................................................... 22
6 Conclusions ....................................................................................................................... 24
Statistical tables ........................................................................................................................ 26
References ................................................................................................................................ 32
_____________________
* Bank of Italy, Directorate General for Economics, Statistics and Research.
1 Introduction1
In his now famous book ‘Capital in the Twenty-First century’, Thomas Piketty begins
by writing: ‘The distribution of wealth is one of today’s most widely discussed and
controversial issues. But what do we really know about its evolution over the long term? Do
the dynamics of private capital accumulation inevitably lead to the concentration of wealth
in ever fewer hands, as Karl Marx believed in the nineteenth century? Or do the balancing
forces of growth, competition, and technological progress lead in later stages of development
to reduced inequality and greater harmony among the classes, as Simon Kuznets thought in
the twentieth century? What do we really know about how wealth and income have evolved
since the eighteenth century, and what lessons can we derive from that knowledge for the
century now under way?’ (Piketty, 2014, p. 11).
Despite the importance of the issue, information on the evolution of wealth inequality
over time is scarce. One of the main sources of information in Italy is represented by the
Survey of Household Income and Wealth (SHIW), conducted by the Bank of Italy since the
mid-1960s (Baffigi et al., 2016). On the occasion of the fiftieth anniversary of this survey,
the Bank of Italy has made some calculations available to the scientific community that
allow us to carefully reconstruct the distribution of income and wealth between the late
1960s and the first half of the 1970s.
Based on this information, the paper reconstructs a benchmark for the joint distribution
of income and wealth for the period 1968-1975 and compares these data with those of the
most recent surveys. In the paper, we reconstruct synthetic microdata, compatible with the
aggregate results published in those years. These microdata are then analysed according to
the methods used in current surveys, for which the original microdata are available instead.
Therefore, the paper is also of an experimental and methodological nature.
The analysis shows a marked increase in the household wealth to income ratio between
the end of the 1970s and the years of the recent financial crisis. The concentration of wealth
shows a marked reduction from 1968 until the beginning of the 1990s, then a recovery and a
subsequent stabilization; in 2014 the concentration of wealth was much lower than in the
years 1968-1975.
The paper is organized as follows: Section 2 shows the data and methods used for the
reconstruction of synthetic data. Section 3 illustrates the macro estimates of household
income and wealth from the post-war period to the present, comparing the trends with those
of the survey. Section 4 reports the main results concerning the inequality of wealth. Section
5 illustrates the evolution of poverty indices from the 1960s to the present, while Section 6
presents the main conclusions.
1 The opinions expressed in this paper do not necessarily reflect those of the Bank of Italy.
5
2 Data and methods
2.1 The data used
The analysis of the distribution of wealth is based on the Bank of Italy’s Survey of
Household Income and Wealth (SHIW). SHIW microdata have been collected since 1977 for
many variables but data collection on wealth- based on a comprehensive definition - only
started in 1987. For the years before 1977 the analysis can be carried out by using
publications from those years and, more recently, the information that the Bank of Italy has
made available to the scientific community in the form of statistical tables describing the
results of the surveys in detail.2
The data taken into consideration refers to the period 1968-1975. For those years, the
published reports include some two-way tables that allow us to accurately reconstruct the
distribution of household per capita income and wealth and their joint distribution (Table 1).
In particular, for the period 1968-1972, the following joint distributions are available:
household wealth and household size; household income and household size; household
wealth and household income; and household wealth and age of the household head.3
The surveys carried out between 1973 and 1975 present some methodological
differences in the questionnaire, the definition of some variables, and the sampling
methodology with respect to the previous ones (rich households were oversampled; see
Brandolini, 1999). The methodological note for the 1973 survey (Bank of Italy, 1973)
describes the oversampling of families belonging to the ‘upper and upper-middle classes’,
specifying that some adjustments were made to the weighting coefficients. Moreover, for
this three-year period only three of the abovementioned four bivariate distributions are
available (household income and household size; household income and wealth; and
household wealth and age of the household head): it will be necessary to take these aspects
into account in both the processing of data and the analysis of the results.
No information on wealth distribution is available for the years before 1968.
Information on household wealth largely comparable with that of the late 1960s was
collected for the SHIW from 1987 to 2014.4 The analyses that will be presented below will
therefore mainly refer to the comparison between the period for which data reconstruction is
carried out (1968-75) and the two decades around the end of the twentieth century. For the
years between 1977 and 1986, however, information is available on some important
components of wealth; this information makes it possible to estimate the trend of inequality
and to provide a more complete overview.
2 The information used for our reconstructions is available at:
http://www.bancaditalia.it/statistiche/tematiche/indagini-famiglie-imprese/bilanci-famiglie/documenti-
storici/index.html
3 In the original tables some data are missing for income or wealth. In some cases it would have been
possible to impute these data by exploiting the partial knowledge present in the data. However, a test
conducted for 1969 showed that this practice did not lead to a significant variation in results. Therefore, it
was decided to carry out the whole analysis only on the available data, as shown in the published tables.
4 However, complete comparability is limited to the years from 1991 onwards.
6
Table 1
Historical tables published on the Bank of Italy’s Internet website (*), 1968-1975
Year
Income and
number of
household
members
Net wealth and
number of
household
members
Net wealth and
income
Net wealth and
age of the head
of household
1965 .......................................................................................... - - - -
1966 .......................................................................................... X - - -
1967 .......................................................................................... X - - -
1968 ..........................................................................................
X X X X
1969 ..........................................................................................
X X X X
1970 ..........................................................................................
X X X X
1971 .......................................................................................... X X X X
1972 .......................................................................................... X X X X
1973 .......................................................................................... X - X X
1974 .......................................................................................... X - X X
1975 .......................................................................................... X - X X
(*) In 1971 net wealth has a less detailed breakdown by wealth brackets.
The definitions of wealth adopted in 1968-1975 are similar to those of 1991 and
subsequent years, but not entirely equivalent. In the Statistical Bulletin of 1970, which
commented on the 1968 data, we find the following definition of wealth:
‘The net worth of households can be defined as the algebraic sum of financial assets
(deposits, securities, credits) and real assets (real estate and durable assets) and liabilities, both
long (mortgages) and short-term (debts for purchases of consumer goods).
With the present survey we tried for the first time to collect data on this aggregate,
consolidating the various components for each household interviewed, taken with their sign. Some
assets have not been considered for the purpose of calculating wealth due to the difficulty of
detection, such as safe-haven assets (gold, jewelry, rare stamps, paintings, etc.) and assets
consisting of credits from other families (loans); the investments in companies and businesses,
both sole proprietorship and in the form of companies (with the sole exception of the shareholding
of listed companies, which are instead considered) were excluded from the calculation, as were
capital goods owned by craftsmen, traders, professionals (...). In addition, the insufficiency of
certain estimates (deposits and securities in particular) due to the poor collaboration of the
interviewees has to be taken into account when interpreting the data’. (Bank of Italy, 1970, page
56).
The main differences in the definition between the data for the late 1960s and the
current data are represented by the presence of durable goods in the former and the presence
of business equities and valuables in the latter. We do not believe that these differences
significantly affect the results of the comparison of inequality indicators. If 1991 data are
revised to take the main differences into account, i.e. by deducting the value of business
equities, with the exception of listed companies, adding cars and excluding valuables, the
Gini wealth index decreases from 0.591 to 0.586, a relatively small variation.
Another important issue is the underestimation of aggregates, due to the phenomena of
non-response, non-reporting, and under-reporting, which could have changed over time. By
comparing the microeconomic estimates of household wealth with the macroeconomic data
estimated by Cannari, D'Alessio and Vecchi (2017), we find that the ratio between the two
measures has not remained constant over time; this issue will be discussed in the following
sections.
7
2.2 The method for reconstructing synthetic microdata
Today, technology allows us to process large volumes of data easily and the
opportunity for researchers to access microdata is considerably higher than in the past. Yet
fifty years ago, when for example the SHIW was born, the hardware and software for data
processing were rudimentary compared with today's standards. Computer memory was
limited; the data were stored on voluminous and easily perishable punch cards; the result of
the calculations consisted of frequency tables, with data grouped in classes.
For the oldest surveys it frequently happens (and this is the case with SHIW) that the
original microdata do not survive as time goes by; what remains today for the surveys
conducted up to 1975 are only statistical calculations, frequency distributions and average
values, for data grouped in classes. The processing of these data is not easy, for example
when calculating the share of the poor in the absence of income or consumption classes that
identify the poverty line.
The construction of synthetic microdata (i.e. microdata reconstructed in such a way as
to replicate the available aggregate tables) makes data processing easier and allows us to use
current methods and standards easily.
The use of synthetic data was initially proposed to meet the needs of economic
research on microdata, while protecting the confidentiality of respondents (Rubin, 1993,
Reiter 2002). This need has grown over time, in correspondence with the increased need for
granular data availability for economic analysis. In practice, the idea is to construct a
microdata sample by simulating an extraction from a multivariate distribution equivalent to
that underlying the individual records whose privacy is to be protected (Barrientos et al.,
2017).
However, the use of synthetic data can also be justified by the need for analysis, when
the availability of microdata makes it possible to calculate the indicators we are interested in.
An example in this sense is found in the work of Shorrocks and Wan (2008), who
reconstruct synthetic samples for the analysis of poverty and inequality starting from income
data grouped in classes.
In this work we adopt a procedure similar to that proposed by some authors for small
area estimators (Tanton, 2014; Williamson, 2013). Starting from a distribution actually
observed and plausibly similar to that to be reconstructed, ‘adjustments’ are made to the
weighting coefficients to align the observed microdata to the frequency distributions that are
to be replicated, which in this case are the tables available for those years described in the
previous paragraph.5
To reconstruct the microdata on wealth, income, number of household members and
age of the head of household in the years considered, we start from the data collected by 12
surveys conducted from 1991 (the first year for which microdata on wealth are available and
in line with those of the following years) to 2014 (last year available). The size of the dataset
is about 95,000 households (Table 2).
5 As described in Tanton (2014), a further possibility, though probably more complex and arbitrary, could be
the generation of totally synthetic samples able to satisfy the constraints in terms of the joint distribution of
the phenomena examined.
8
For each year between 1991 and 2014 the data have been preliminarily reproportioned
in order to make the averages of wealth (Wt) and income (Yt) equal to those of the year to be
estimated:
Wk* = Wt * M(Wk)/M(Wt) and Yk* = Yt * M(Yk)/M(Yt)
where k=1968, …, 1975 ; t=1991, …, 2014
and M(.) represents the arithmetic mean operator.
This set of microdata, which in the distribution represents the average profile of wealth
and income between 1991 and 2014, has been subjected to a raking procedure (Deville and
Sarndal, 1992), using the bivariate distributions drawn from the statistical reports of the
surveys published at the time as constraints. The technique performs a post-stratification that
first of all allows us to satisfy the constraints imposed by the joint distribution of wealth and
the number of components; then the new weights undergo a new reweighting procedure,
aimed at satisfying the further constraint imposed by the joint distribution of income and the
number of components. Subsequently we proceed by readjusting the weights in order to
obtain the joint distribution of wealth and income; finally we proceed with a reweighting that
provides the bivariate distribution of wealth brackets and age of the household head. Since
only the last joint distribution is fully consistent with the constraints at the end of this first
cycle, the process is repeated iteratively until the four bivariate distributions are all satisfied
at the same time.6
In this way we construct microdata compatible with the constraints, which allows us to
estimate indicators based on the wealth and income distribution with great flexibility.7 For
example, synthetic microdata allow us to adopt the same equivalence scales that we use
today, making possible the reconstruction of homogeneous historical series, or the
calculation of indices based on the joint distribution of income and wealth. Using the full set
of data from 1991 to 2014, the sample size is large enough to represent the typical
characteristics of these distributions, such as the particular asymmetry and the limited
presence of negative values.
Raking is not free from possible shortcomings (Brick et al., 2003) and there is a
considerable lag between the period to which the starting data refer and the year whose
distribution is estimated; this suggests that the robustness of the results needs to be assessed
in various ways.
First we estimated the amount of variance contained in the synthetic microdata of
wealth and income attributable to the classificatory variables, and therefore attributable to
the information contained in the reports, and the residual variance (within the cells). This
assessment can be carried out by using the most recent surveys for which microdata are
6 The use of three bivariate marginal tables (income by wealth, wealth by number of components and income
by number of components) makes it possible to estimate all the second order moments of the trivariate
distribution.
7 Shorrocks and Wan (2008) propose a parametric method for estimating the distribution of phenomena
starting with the values published in the statistical tables of the reports. Their method, unlike the one
proposed here, is aimed at a univariate reconstruction of the variable of interest.
9
available. In particular we estimated a linear model where the value of household wealth
(and income) is a function of the four bivariate classification criteria used in the procedure:
W = a + b CLY*NCOMP + c CLW*NCOMP + d CLW*CLY + f CLW * CLETA
where CLY and CLW represent the income and wealth brackets respectively, NCOMP the
number of household members and CLETA the age classes of the household head; a, b and c
are parameters to be estimated.
The R2 coefficients are around 90 per cent between 2010 and 2014; similar models for
income provide R2 values of around 95 per cent on average. Thus it is likely that a very large
share (90-95 per cent) of the variance of the 1968-75 micro-synthetic data is explained by
the classificatory variables present in the historical reports and used as constraints in our
procedure; the residual variability (5-10 per cent) would be attributable to the variance
within the cells that the procedure imputes by using data from the most recent years.
Synthetic data can also provide fairly accurate information on some joint distributions.
Information on the bivariate distributions between income and wealth (and the others) is
directly inserted into the data set through the constraints; the bivariate aggregate tables are
perfectly reproduced in the synthetic data. In contrast, the trivariate distributions (for
example, between wealth, income and number of components) are only approximated by the
knowledge of the three bivariate distributions. The higher order relations are approximated
in the microdata, given the constraints on the bivariate distributions, by using the
information contained in the most recent surveys.
To assess the accuracy of this approximation, we resort to recent data and estimate the
logarithm of the frequency (log fijk) of the trivariate distribution as a function of the dummies
related to the three bivariate distributions:
log fijk = a + b CLY*NCOMP + c CLW*NCOMP + d CLW*CLY
The R2 of this model is around 90 per cent; therefore synthetic data seem to be a good
representation of the trivariate distribution.
A final robustness check was based on simulation models. In particular we generated
the 1968 microdata on income and wealth by using a bivariate lognormal distribution
(shifted to enable the generation of negative values of income and wealth).
For each household size in terms of the number of components, we extracted a sample
from a bivariate lognormal distribution with the mean and variance of income and wealth
estimated on the historical statistical reports and with a correlation between income and
wealth (by household size) estimated on the most recent microdata. We then applied raking
techniques to the synthetic sample in order to make the simulated distributions consistent
with the aggregate bivariate historical tables. The results obtained by using this method
compare very closely with the previous ones: the choice of starting data seems to have little
influence on the results, which seem to be robust.
10
Table 2
Sample size of SHIW surveys, 1968-1975 and 1987-2014
Year Sample
1968 ......................................................................................................................................................................... 3,478
1969 ......................................................................................................................................................................... 3,355
1970 ......................................................................................................................................................................... 3,026
1971 ......................................................................................................................................................................... 6,725
1972 ......................................................................................................................................................................... 5,889
1973 ......................................................................................................................................................................... 5,177
1974 ......................................................................................................................................................................... 4,605
1975 ......................................................................................................................................................................... 4,447
1987 ......................................................................................................................................................................... 7,328
1989 ......................................................................................................................................................................... 8,274
1991 ......................................................................................................................................................................... 8,188
1993 ......................................................................................................................................................................... 8,089
1995 ......................................................................................................................................................................... 8,135
1998 ......................................................................................................................................................................... 7,147
2000 ......................................................................................................................................................................... 8,001
2002 ......................................................................................................................................................................... 8,011
2004 ......................................................................................................................................................................... 8,012
2006 ......................................................................................................................................................................... 7,768
2008 ......................................................................................................................................................................... 7,977
2010 ......................................................................................................................................................................... 7,951
2012 ......................................................................................................................................................................... 8,151
2014 ......................................................................................................................................................................... 8,156
Total 1991-2014 ...................................................................................................................................................... 95,586
Given that the bivariate distribution between wealth and number of components is not
available for 1973-75, we used the synthetic microdata estimated for 1972 as a starting point,
when all three bivariate distributions are available; then we applied the raking procedure,
making the 1973-75 data consistent with the bivariate aggregate statistical tables available
for those years (i.e. with breakdowns by income and number of members, and by income and
wealth).
2.3 The Gini index estimates for 1977-1986
Between 1977 and 1986 the SHIW did not collect data on financial items. The Gini
indices of net wealth are therefore estimated starting from those relating to real assets, which
are the main component of household wealth. According to the decomposition of the Gini
index proposed by Pyatt, Chen and Fei (1980), if Wk (k = 1, ... 3) are the three components
of net wealth (real, financial and liability assets) and Gk the respective Gini indices, then:
G = k k Rk Gk where k = k/ e Rk = Cov (Wk, Rw) / Cov (Wk, Rwk).
In other words, the Gini index of net wealth is a linear combination of the Gini indices
Gk of its components, whose coefficients depend on the ratio of the average values of the
components (k) to the average wealth () and on the rank correlation ratio Rk, defined as
the ratio of the covariance between the k-th component and the ranking of the average
wealth to the covariance between the k-th component and the ranking of the component
itself. By assigning the average value observed for the years 1991-2014 to the financial
11
items, we obtain the Gini index of net wealth: the trend substantially reflects that of real
assets, while the level is modified due to the contribution of the other components.8
3 Aggregate income and wealth trends since the post-war period
The survey data show a sustained growth in the average wealth-to-income ratio (Figure
1), from values of around 3.0 measured at the end of the 1960s to values above 7.0 in 2014.
From a qualitative point of view these results are similar to those based on macroeconomic
estimates of the wealth-to-GDP ratio, although in the three years 1973-75 it is clear that the
effects of the oversampling of wealthy households were not entirely adjusted by the
reweighting procedure.
According to Cannari, D'Alessio and Vecchi (2017), this ratio has shown a downward
trend in Italy from the late nineteenth century (when household wealth was more than six
times GDP) to the first half of the 1960s (a period when the authors find the ratio to be about
3.0); the ratio then started growing again, returning to the levels of the late nineteenth
century in the first decade of the twenty-first century. As Piketty (2014) pointed out, this
trend can be found in other important western countries too.
Due to the greater asymmetry that characterizes the distribution of wealth compared
with that of income, the ratio of the median of household wealth to the median household
income is lower than the values examined so far, based on the ratio of means; we move from
values of around 1.0 at the end of the Sixties to about 5.0 in the most recent years. The
trends, however, are similar; wealth has significantly increased also for the families
belonging to the central distribution classes. At the micro level, the correlation coefficient
between income and wealth shows a significant increase over time, from values of around
0.45 in the period 1968-75 to 0.60 in the most recent years.
In Italy the growth of household wealth has been accompanied by an increase in the
number of real estate owners, and in particular of home owners, especially in the period
1971-91. The increase in house prices, up to the years of the recent financial and economic
crisis, has far outweighed the inflation rate and has also contributed to the growth of wealth
(Cannari, D'Alessio and Vecchi, 2017).
8 A similar reconstruction was carried out by D'Alessio (2012), although using the estimated shares at the
macro level as the weights of the various components and not those deriving from the sample estimates, as
we did here for continuity with the other available estimates. The results are therefore not comparable.
12
Figure 1
Wealth, GDP and income
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
1940 1950 1960 1970 1980 1990 2000 2010 2020
Wealth/GDP(macro)
Wealth/Income(ratioofaverages‐SHIW
andsyntheticdata)
Wealth/Income(ratioofmedians‐SHIW
andsyntheticdata)
Source: Data on aggregate wealth are drawn from Cannari, D’Alessio e Vecchi (2017). Sample estimates are
reconstructed in the present paper.
Starting from the 1970s and despite cycles of different duration and intensity, house
prices recorded a rapid growth in Italy: from 1970 to 2007 (maximum point of the real estate
cycle) they almost tripled in real terms. The increase in house prices was almost double that
of construction costs for residential buildings (Figure 2).
The change in house prices has therefore had a significant effect on the relationship
between wealth and GDP. To get an idea of the importance of the price factor for the growth
of the GDP-wealth ratio, we can calculate the value of the houses owned by households, net
of the real change in house prices. In this way Cannari, D'Alessio and Vecchi (2017) come to
an estimate of the household wealth to GDP ratio that is two units lower than that observed
in 2012. In other words, two thirds of the increase in the relationship between household
wealth and GDP is due to the growth in real house prices, which in turn is largely
attributable to the increase in the price of building land.
13
Figure 2
House prices and residential investment indexes
(Indexes, base 1927=1; real values*)
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020
Housepricecs
Indexofthecostofresidentialinvestments
(*) Aggregates are in real terms, using the Istat general deflator (‘valore della moneta’).
Source: Cannari, D’Alessio e Vecchi (2017).
4 Wealth inequality from the late 1960s to 2014
4.1 The estimates from 1968 to the most recent years
Tables A1-A6 in the Appendix show the estimates of the Gini indices and other
information on the distribution of wealth and income, at the household level, per capita and
equivalent (using the square root of the number of household members as an equivalence
scale) obtained on synthetic data generated for the years 1968-75. These estimates are joined
by those obtained on the microdata for the period 1987-2014.
At the end of the 1960s, household wealth was, as it is now, far more concentrated
than income. The Gini index of household assets ranged between 0.74 and 0.80, while for
income it ranged from 0.38 to 0.40. Similar orders of values emerge from the examination of
per capita figures (Figure 3).
14
Figure 3
Wealth and income inequality, 1968-2014
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
1960 1970 1980 1990 2000 2010 2020
0.25
0.30
0.35
0.40
0.45
0.50
0.55
0.60
0.65
Householdwealth
Householdwealth(reconstructed)
Realassets
Householdincome(righthandscale)
Equivalentincome(righthandscale)
Householdincome(noincomefromfinancialassets)(rightha nd
scale)
Source: The indices referring to wealth between 1991 and 2014 and to income and real activities between 1977 and
2014 are obtained from the SHIW’s Historical Archive data. Indices relative to the net wealth in 1987 and 1989 are
obtained from the annual SHIW’s archive data. The indices referring to the wealth between 1977 and 1986 were
estimated according to the method of Pyatt et al. (1980); those referring to the income and wealth of the period 1968-
75 are our calculations using the synthetic data reconstructed here.
In the period 1968-1975, about two tenths of families had negative or zero wealth,
compared with a share of around 5 per cent in the most recent years. Some 10 per cent of the
wealthiest families had between 56 and 61 per cent of the total wealth in the late 1960s; this
share falls to 40.6 per cent in 1989-91 and then goes back up with the 1992-93 recession.
The share held by the richest households has been around 45 per cent in the most recent
years.
The intermediate classes have acquired increasing shares of wealth. The household
segment between the 20th and 80th percentiles of wealth distribution held less than 25 per
cent of total wealth in the late 1960s while they held around 40 per cent in the two-year
period 1987-89 and still well above 35 per cent in the following years. In other words,
between 1968-75 and the most recent years we observe a significant increase in the wealth
held by the middle class at the expense of the share held by the richest class (Table A1).
The change in the distribution of wealth is associated with the diffusion of home
ownership in the last quarter of the past century: the share of homeowners increases from 45
per cent in the late 1960s to 63 per cent in the early 1990s and then settles at around 68 per
cent from the year 2000 onwards; on the other hand, the share of tenants has remained
almost stable at just over 20 per cent since the beginning of the new century (Table 3).
15
Table 3
Principal residence by tenure, 1968-1975 and 1987-2014
Year Owned by the
household Rented or
sublet
Occupied under
redemption
agreement
Occupied in
usufruct, free of
charge, etc. Total
1968 45.6 42.9 3.7 7.8 100.0
1969 45.4 43.0 5.4 6.2 100.0
1970 47.0 42.5 4.2 6.3 100.0
1971 43.5 44.9 4.8 6.8 100.0
1972 45.1 44.3 4.1 5.5 100.0
1973 46.4 44.4 2.3 6.9 100.0
1974 47.0 45.4 2.0 5.6 100.0
1975 46.3 46.1 2.2 5.4 100.0
1976 51.7 39.9 2.6 5.8 100.0
1977 47.3 44.8 1.9 6.0 100.0
1978 48.3 41.9 2.5 7.2 100.0
1979 51.5 40.4 2.0 6.1 100.0
1980 56.2 36.7 1.6 5.5 100.0
1981 49.6 41.3 2.0 7.1 100.0
1982 57.1 35.3 1.4 6.1 100.0
1983 57.5 33.7 1.9 6.9 100.0
1984 59.4 30.9 1.9 7.8 100.0
1986 59.5 31.5 0.9 8.2 100.0
1987 61.6 29.3 1.1 8.1 100.0
1989 62.0 27.8 1.5 8.7 100.0
1991 63.5 24.5 1.5 10.5 100.0
1993 62.6 24.9 1.0 11.5 100.0
1995 64.7 23.7 0.8 10.8 100.0
1998 65.8 22.8 0.6 10.9 100.0
2000 68.2 20.9 0.7 10.1 100.0
2002 68.4 20.9 0.5 10.2 100.0
2004 67.5 21.7 0.4 10.3 100.0
2006 68.5 21.0 0.4 10.1 100.0
2008 68.6 21.5 0.6 9.4 100.0
2010 67.7 21.6 0.3 10.4 100.0
2012 66.6 22.3 0.3 10.8 100.0
2014 67.7 20.7 0.5 11.1 100.0
Source: Estimates referring to 1968-1976 are drawn from the Reports (Supplements to the Bank of Italy’s Statistical Bulletin) available at
http://www.bancaditalia.it/pubblicazioni/indagine-famiglie/index.html; for the most recent years, estimates are obtained on data from the
SHIW’s Historical Archive.
In those years, the growth in household wealth that characterized such a large segment
of the population could be attributed to the high savings rate, a sustained GDP growth
(though lower than in the early 1960s), and to the almost uninterrupted growth in house
prices in the last quarter of the 20th century. Public spending was reflected in the
accumulation of public debt and private wealth, while the demographic trends kept demand
for housing high, especially in large cities.
Over time the concentration of wealth has strongly decreased. The Gini index of
household wealth has fallen from values of around 0.75 in the period 1968-1975 to 0.58 in
1989. The Gini indices calculated for the period 1977-1986 by using the method of Pyatt,
Chen and Fei ( 1980) appear to be in consistent with the preceding and following estimates:
they confirm the downward trend of the concentration of wealth from the late 1960s to the
end of the 1980s (Figure 3). Then the concentration returned to increase up to 0.63 in 2000
and oscillated in the subsequent years (0.64 in 2012 and 0.61 in 2014). Overall, the double
16
recession, coinciding with the global financial crisis and the European sovereign debt crisis,
has had relatively modest effects on inequality.9
The two main recessions in Italy’s economy since the post-war period have had
different effects on inequality: during the first recession, the distributions of income and
wealth polarized, while during the second one they recorded a general downward shift.
These phenomena are even more evident when the population is divided into three groups:
families with wealth below one half the value of the median, those with more than three
times the median, and the remaining families that we define as the middle class.10
At the end of the 1960s the poorest class included 40 per cent of households, holding a
wealth of around 1 per cent of the total; from 1987 onwards, the share of this segment in
terms of households decreased to around 35 per cent, while its share in terms of total wealth
increased to around 3 per cent. The middle class almost doubled, from just over 25 to about
50 per cent in terms of households; the wealth of this class rose from 12.5 to 44 per cent of
the total. The richest class halved in size (from 26 to 13 per cent), with a sharp drop in the
share of wealth, from about 85 to 50 per cent.
Table 4
Net wealth and social classes, 1968-2014
Share of households Share of net wealth
Year
Poor
segment
(wealth
lower than
half the
median)
Middle-class
segment
(wealth
higher than
half the
median but
lower than 3
times the
median
Rich
segment
(wealth
higher than 3
times the
median)
Total
Poor
segment
(wealth
lower than
half the
median)
Middle-class
segment
(wealth
higher than
half the
median but
lower than 3
times the
median
Rich
segment
(wealth
higher than 3
times the
median)
Total
1968 42.9 28.6 28.5 100.0 0.8 12.3 86.8 100.0
1969 43.1 28.5 28.4 100.0 0.7 12.2 87.0 100.0
1970 42.2 28.0 29.8 100.0 1.2 12.4 86.5 100.0
1971 45.4 24.1 30.5 100.0 1.2 11.3 87.4 100.0
1972 44.3 23.8 31.9 100.0 0.9 10.0 89.3 100.0
1973 45.4 23.0 31.6 100.0 0.5 9.8 89.7 100.0
1974 44.0 29.2 26.8 100.0 1.2 15.5 83.5 100.0
1975 43.4 30.9 25.7 100.0 1.3 18.0 80.7 100.0
1987 35.4 50.1 14.6 100.0 3.6 42.2 54.4 100.0
1989 35.6 53.8 10.6 100.0 4.6 54.0 41.3 100.0
1991 36.2 49.9 13.9 100.0 4.2 46.7 49.0 100.0
1993 36.9 48.5 14.6 100.0 3.7 43.4 52.8 100.0
1995 35.5 49.9 14.5 100.0 3.4 44.3 52.1 100.0
1998 36.4 50.3 13.4 100.0 3.6 42.9 53.6 100.0
2000 36.1 50.3 13.6 100.0 3.5 41.8 54.7 100.0
2002 35.9 50.4 13.7 100.0 3.4 44.3 52.2 100.0
2004 37.2 49.5 13.3 100.0 3.7 46.3 50.0 100.0
2006 36.4 49.9 13.6 100.0 3.4 44.0 52.5 100.0
2008 35.8 50.6 13.6 100.0 2.9 44.5 52.4 100.0
2010 36.9 49.8 13.2 100.0 2.9 44.0 52.9 100.0
2012 37.7 47.7 14.6 100.0 2.7 41.0 56.5 100.0
2014 37.4 49.5 13.2 100.0 3.1 45.3 51.8 100.0
9 According to Acciari et al. (2017), who estimate the concentration of wealth between 1995 and 2013 by
using inheritance data, the concentration of wealth in recent years has experienced greater growth than that
observed in the SHIW data.
10 Atkinson and Brandolini (2013) use a similar partition to income to identify the ‘middle class’.
17
On the whole, the concentration of income decreased as well, showing different trends
in the various sub-periods. The Gini index of per capita income decreased from 0.39-0.42 in
the period 1968-1970 to 0.33 in 1991; it then returned to increase until 1998 (0.365),
oscillating in the following years at around slightly lower levels.
4.2 Wealth and under-reporting
Sample data on household wealth are often affected by under-reporting, i.e. the
tendency of the interviewees not to declare everything they possess. With regard to the
SHIW, the phenomenon has been studied by several authors (Cannari and D'Alessio, 1990;
Cannari et al., 1990; Cannari and D'Alessio, 1993; D'Aurizio et al. 2006).
The bias determined by these behaviours can be evaluated by comparing sample
estimates with macro sources. The ratio of total SHIW estimates to total macroeconomic
wealth estimated by Cannari et al. (2017) for 1987-2014 is around 0.55; over the years 1968-
72 the ratio was around 0.6-0.7.
On the contrary, the 1973-75 sample estimates are substantially in line with macro
estimates, probably due to the oversampling carried out in those years. Although the
differences between micro and macro estimates can be partly attributed to definitional
differences (Baffigi et al., 2016), the phenomenon of under-reporting cannot be neglected,
even when the analysis concerns the relative distribution of a variable among families rather
than the absolute levels. The aforementioned studies have indeed shown that the inequality
indices calculated on data adjusted to take the under-reporting into account are generally
different from those obtained by using unadjusted data.
As the level of under-reporting may change over time, we use calibration techniques
(Deville and Sarndal, 1992). In short, we reweight the starting data so as to align the survey
estimates of total wealth with macroeconomic estimates.
The reweighting is carried out according to a statistical criterion; given the constraint,
we minimize the distance between the starting weights and the new weights. Different
distance criteria can be used. In this paper we resort to four different methods made available
in the SAS Calmar macro (linear, raking ratio, logit and linear truncated; see Sautory, 1993)
and use their average as a benchmark (Table A8 in the appendix).
As observed by D'Alessio and Neri (2015), the calibration of wealth data leads to
higher concentration levels (Figure 4). For example, in 1968 the concentration index
increases from 0.758 to 0.79 after the calibration; in 2014, the increase due to calibration is
even more marked (from 0.612 to 0.68).
The overall trend is not significantly affected by these adjustments, although the
calibration has a larger impact on the most recent inequality estimates (the impact is around
11 per cent in the period 1987-2014, compared with 4 per cent in the period 1968-72). A
significant difference, however, concerns the 1973-75 period, when unadjusted estimates are
very close to calibrated ones, due to the oversampling of richest households carried out in
those years. The calibration makes the downward trend between the periods 1968-72 and
1987-1989 more evident; after the adjustment, the 1973-1975 estimates are located in an
intermediate position between the two periods.
18
Overall the adjustment made to take the under-reporting into account does not
substantially modify the general picture of inequality. Adjusted data, however, predates the
decline in the levels of wealth inequality and reduces the gap between the 1973-75 estimates
and the more recent ones.
Figure 4
Wealth inequality: adjustments for under-reporting, 1968-2014
0.55
0.60
0.65
0.70
0.75
0.80
0.85
1960 1970 1980 1990 2000 2010 2020
0.55
0.60
0.65
0.70
0.75
0.80
0.85
Householdwealth
(unadjusted)
Householdwealth
(udjustedforunder
reporting)
Source: The indices referring to wealth between 1968 and 1975 are obtained on the synthetic data reconstructed here;
those between 1987 and 2014 are obtained on the data of the SHIW Historical Archive. The estimates adjusted for
under-reporting are obtained as the mean of calibrated estimators (Table A8 in the Appendix).
4.3 Inequality in the income-net worth indicator
Both income and wealth are relevant to economic well-being; therefore it is natural to
measure inequality by taking both aspects and not just one of them into account. This is
particularly important because of the different dynamics that, as we have seen, have
characterized the two indicators over time.
In order to carry out this evaluation, we proceeded to calculate a synthetic indicator,
obtained by adding the flow of resources that an household could perceive by alienating its
assets to the current income (Weisbrod and Hansen, 1968). In the construction of the
indicator, income is assumed to be perceived over the residual life of the household head
(according to the Istat mortality tables by age and sex), and the return on assets is set equal
to 2 per cent. It should be noted in this regard that the relationships between both wealth and
income and wealth and age are kept under control in the construction of the synthetic
microdata for the period 1968-75; therefore the composite income-net worth indicator should
be accurate.
Over the period 1968-1993 the time pattern of the Gini index calculated on this
income-net worth indicator is similar to that of household wealth; inequality decreased
markedly until 1989, and then recorded a strong increase around the 1992 crisis. After 1993,
however, the composite index shows a marked increase in inequality while household wealth
inequality remained almost stable. The increase is steeper when measured in equivalent or
19
per capita terms, reflecting the progressive improvement in the relative conditions of elderly
households (usually small in size) to the detriment of younger ones (with children). The
values of inequality observed at the end of the reference period are similar to those of the
late 1960s (Figure 5).
Figure 5
Income-net worth ineqaulity indices, 1968-2014
0.30
0.35
0.40
0.45
0.50
1960 1970 1980 1990 2000 2010 2020
0.30
0.35
0.40
0.45
0.50
Equivalentincomenetworth Percapitaincomenetworth
Householdincomenetworth
4.4 A long-term look
We have seen in in the previous section that on the whole the concentration of wealth
decreased from the end of the 1960s to 2014. But what is the long-run trend?
According to Alfani (2016), who studied the distribution of wealth between 1300 and
1800 in four pre-unification states in Italy,11 during the whole period there was a progressive
concentration of wealth, with a tendency for the richest class to move away from the
conditions of the other social strata. Between 1300 and 1800 the share of wealth held by ten
per cent of the richest individuals rose from 45-55 per cent to 70-80 per cent.12 The only
period in which the author records a reversal in this pattern is that corresponding to the Black
Death epidemic in the middle of the fourteenth century.
Gabbuti (2017), using data on inheritance taxes for the years ranging from the late
nineteenth century to 1915, estimates a share of wealth held by the richest tenth of the
population of between 64 and 81.5 per cent. The maximum is found for the years 1912-1913,
which were followed by a significant drop, to 69.6 per cent, in 1914-15. These are lower
values than but not too dissimilar to those of France.
11 These are assessments concerning real estate wealth, as inferred from land registers or similar sources of the
time, in the Kingdom of Sardinia, Florence, the Kingdom of Naples and the Republic of Venice.
12 The estimate for the end of the period considered is consistent with that provided by Piketty et al. (2006) for
the European average of 1810.
20
Comparisons with the data for the late 1960s taken from the SHIW need to be
cautious, also in consideration of the fact that the survey tends to underrepresent the richest
families and it is therefore possible that the concentration of wealth is underestimated.
However, it seems plausible that in the years between 1910 and the end of the 1960s the
concentration of wealth in Italy decreased considerably.13
The share of wealth held by the wealthiest tenth of Italian families at the end of the
1960s stood at relatively similar levels (given the uncertainty that characterizes this kind of
estimate) to those of other western countries, in particular France and the United Kingdom.
Instead it was lower than that of the United States (Figure 6).14
At the end of the period considered, the share of wealth held by the richest tenth was
lower than that in the late 1960s in France, the United Kingdom and Italy, while growth was
observed in the United States. The decline is more pronounced for Italy than other European
countries, and the recovery since the 1990s has been less evident in our country than in
others.
Figure 6
Share of wealth held by the richest tenth of households
0.35
0.45
0.55
0.65
0.75
0.85
0.95
1880 1900 1920 1940 1960 1980 2000 2020
France
UnitedKingdom
UnitedStates
Italy
Source: Wealth and Income Database (WID.world) and our calculations for Italy.
13 A rise in wealth concentration also happens in other western countries, such as France, the United
Kingdom, and Sweden. In the United States, wealth inequality declines over the period 1910-1950 (less
than in Europe) and remains relatively stable over the following twenty years (Piketty, 2014).
14 In this section the SHIW data are compared with those of the WID database (The World Wealth and
Income Database, available at WID.world).
21
5 Indicators of poverty from the 1960s to the present
The availability of microdata for both income and wealth makes it possible to compare
poverty indicators according to various definitions, taking into account both the size of the
household (for the equivalence scales) and the interaction between these two aggregates.
The relative poverty indicator based on equivalent income alone (with an equivalence
scale equal to the square root of the number of components) is significantly reduced between
the beginning of the period examined and the beginning of the 1980s; at the beginning of the
1990s it rose abruptly and then, after a certain decline until 2006, it rose again during the last
phase of the economic crisis. In 2014 the relative poverty level was only slightly lower than
that observed on average for the period 1968-75 (Figure 5). The measurements carried out
for the period 1977-1986, when income from financial assets cannot be included in income,
make it possible to complete the picture of relative poverty levels in Italy in those years,
confirming the trend for the period thereafter, already highlighted for income including any
financial returns.
With information on both income and wealth, it may be useful to examine indicators
that jointly consider these two aggregates. Hence, reference is made here to a simple
indicator that considers households with an equivalent income below the poverty line as
being poor and which have, at the same time, a net wealth lower than a fraction (or multiple)
of the same threshold. In this way, those who, by liquidating their assets would have
sufficient resources to overcome the poverty line for a certain time, are excluded from the
category of the poor.15 Obviously the more time is taken into consideration the lesser the role
assigned to the wealth component. Given a certain arbitrariness in the choice of this
parameter, the share of poor households was calculated on the basis of the hypotheses of
three months, one year and three years.16
Indicators that also consider wealth involve a significant reduction of more than half in
the estimate of the poor. The reduction is obviously less marked when wealth greater than
the poverty line for a period of three years (i.e. about €25,000 for an individual in 2014) is
considered sufficient to get out of the poverty condition. The poor are reduced to about a
third of the initial estimate calculated only on income when wealth exceeding three times the
monthly poverty threshold (€2,200 in 2014, again for an individual) is considered sufficient
to get out of the poverty condition.
However, the results of the various experiments converge on two points: a) relative
poverty levels in 2014 are close to those at the end of the 1960s, in particular when
considering indicators that also consider wealth; b) from 2008 to 2014 the share of
components in relative poverty definitely increased. The results for the levels of relative
poverty only partially reproduce those obtained for inequality, which does not appear to have
increased significantly over the last few years, except in the version of the income-wealth
indicator.
15 The lack of the wealth that would be needed by the household to cope with unforeseen events can also be
seen as a difficulty in itself (asset-poverty). See Brandolini et al. (2010) for a discussion on the point.
16 Haveman and Wolff (2004), Short and Ruggles (2005) and Brandolini et al. (2010) consider a period of
three months; Gornick et al. (2009) instead consider six months. In this paper we have considered both
these and other hypotheses, also as a robustness analysis.
22
Figure 7
Relative poverty, 1965-2014
(percent of individuals)
0.00
5.00
10.00
15.00
20.00
1960 1970 1980 1990 2000 2010 2020
Income
Income(noincomefromfinancialassets)
Incomeandwealth(1year)
Incomeandwealth(3months)
Incomeandwealth(3years)
The growth in poverty levels in recent years has been affected by the intensification of
migration. The share of immigrants, which in the survey can only be defined on the basis of
the place of birth and limited to the regularly resident part, has been increasing, from 1 per
cent in the early 1990s up to about 10 per cent in recent years (Table A9). In this segment of
the population, the share of the poor has grown steadily over the years, from about 10 per
cent in the early 1990s – a share in line with the remaining part of the population - to over 30
per cent in the last few years. This is due to a radical change in the composition of the
foreigners present in Italy, with a reduction in the proportion of people born in Western
Europe, America and Oceania and a growth in the number of people born in Eastern
European countries and especially in Africa and Asia.
The result is an increasing contribution to the spread of poverty in Italy by immigrants,
who in recent years have come to represent about a quarter of the poor in Italy. For the
population of those born in Italy alone, the spread of relative poverty has been almost
permanently decreasing from the mid-90s to 2008 and substantially stable in the following
years (Figure 8).
23
Figure 8
Relative poverty in Italy by place of birth, 1991-2014
(percentage of individuals)
0.0
5.0
10.0
15.0
20.0
1991 1993 1995 1998 2000 2002 2004 2006 2008 2010 2012 2014
BorninItaly Bornabroad
6 Conclusions
By exploiting information in some recently published reports on the Bank of Italy’s
SHIW between 1968 and 1975 on the distribution of income and wealth in Italy, this paper is
the first to estimate wealth inequality in Italy in the period 1968-75.
In particular, micro-synthetic data are reconstructed that are compatible with the
information present in the reports of the time. This made it possible to obtain estimates of the
concentration and relative poverty indices with the statistical criteria used today, allowing a
close comparison with the most recent estimates available (from 1977 onwards).
The results related to wealth concentration identify a downward trend in the 1970s and
1980s similar to that found by other authors on household income (Brandolini, 1999) up
until the recovery that characterizes the years following the 1992-93 crisis and the relative
stability of the new century. However, the estimated values for the period 1968-75 remain
higher than those of the most recent years.
Estimates of relative poverty, calculated using both indicators of equivalent income
and indicators that jointly consider income and wealth, highlight a similar decreasing trend
up to the 1990s and a subsequent growth; for these indicators, however, a more decisive
trend has been observed in recent years compared with concentration indices. In particular,
the levels of recent years are similar to those seen in the period 1968-75 for the poverty
indicators that also take wealth into account.
24
Finally, the work showed that a significant contribution to the growth in the share of
the poor in Italy in recent years has been provided by the intensification of migration flows,
and in particular those coming from emerging countries. In recent years, the relative poverty
rates estimated for the population of those born in Italy have been substantially stable.
25
Statistical tables
26
Table A1
Wealth distribution in 1968-75 and comparison with 1987-2014 – Household wealth
(Share of wealth per tenth of households and inequality indices)
Estimates on synthetic data Estimates on Historical Archive data
1968 1969 1970 1971 1972 1973 1974 1975 1987 1989 1991 1993 1995 1998 2000 2002 2004 2006 2008 2010 2012 2014
1st tenth -0.4 -0.4 -0.3 -0.5 -0.5 -0.2 -0.1 -0.2 -0.1 0.0 0.0 0.0 0.0 -0.1 0.0 0.0 0.0 -0.1 0.0 -0.1 -0.1 0.0
2nd tenth -0.1 -0.2 -0.1 0.0 -0.1 -0.2 -0.1 0.0 0.3 0.4 0.4 0.2 0.3 0.3 0.3 0.2 0.2 0.2 0.1 0.2 0.0 0.1
3rd tenth 0.1 0.1 0.1 0.3 0.3 0.0 0.1 0.0 1.4 1.6 1.4 1.0 1.1 1.1 1.3 1.2 1.2 1.1 1.0 0.9 0.7 0.9
4th tenth 0.8 0.8 1.0 0.8 0.7 0.3 0.7 0.6 3.1 3.6 3.2 2.7 3.0 2.9 3.1 3.1 3.2 3.2 3.2 2.9 2.7 3.1
5th tenth 1.9 1.8 2.1 1.8 1.8 1.5 2.1 2.1 4.8 5.8 5.3 4.7 4.9 4.8 4.8 5.0 5.2 5.1 5.3 5.2 4.7 5.3
6th tenth 3.8 3.7 3.8 4.1 3.8 3.7 4.4 4.5 6.5 7.8 7.3 6.8 6.9 6.7 6.5 6.7 7.3 7.1 7.2 7.2 6.7 7.4
7th tenth 6.4 6.3 6.7 7.0 6.9 6.7 7.0 7.4 8.4 9.6 9.7 9.2 9.1 8.9 8.6 9.1 9.5 9.3 9.3 9.1 8.9 9.5
8th tenth 10.4 10.3 11.3 10.9 10.9 10.4 10.8 11.3 11.3 12.6 13.1 12.6 12.4 11.8 11.3 12.1 12.5 12.2 12.2 11.8 12.1 12.6
9th tenth 18.6 18.4 19.5 18.5 18.6 16.9 17.6 18.2 17.4 17.9 19.1 18.5 17.9 17.1 16.6 17.6 17.8 17.2 17.3 16.9 17.4 17.6
10th tenth 58.7 59.3 55.9 57.1 57.7 60.9 57.5 56.1 46.8 40.6 40.6 44.4 44.6 46.4 47.5 45.0 43.0 44.6 44.4 46.1 47.0 43.7
P10/Median (%) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 2.1 3.4 4.0 1.2 1.8 2.2 2.4 2.0 2.0 1.4 1.0 0.9 0.4 0.7
P20/Median (%) 0.0 0.0 0.0 4.9 6.2 0.0 0.0 0.0 13.0 13.2 13.1 8.9 8.9 10.5 11.3 9.8 8.8 7.2 5.7 5.6 3.9 4.7
P80/Median (%) 497.7 513.2 517.3 476.0 517.9 531.8 425.5 429.1 243.8 214.9 245.1 255.3 248.2 243.5 237.8 246.1 235.1 225.4 227.3 221.5 242.7 227.1
P90/Median (%) 905.3 917.5 901.7 848.0 971.3 961.4 757.1 749.2 406.2 323.1 376.5 404.3 386.6 383.8 370.7 385.4 355.0 349.7 344.9 344.5 375.6 359.1
Gini index 0.758 0.764 0.739 0.747 0.754 0.772 0.744 0.737 0.628 0.577 0.591 0.624 0.619 0.629 0.631 0.619 0.604 0.616 0.615 0.627 0.643 0.613
Table A2
Wealth distribution in 1968-75 and comparison with 1987-2014 – Equivalent wealth (squared root scale)
(Share of wealth per tenth of households and inequality indices)
Estimates on synthetic data Estimates on Historical Archive data
1968 1969 1970 1971 1972 1973 1974 1975 1987 1989 1991 1993 1995 1998 2000 2002 2004 2006 2008 2010 2012 2014
1st tenth -0.3 -0.5 -0.2 0.0 -0.7 -0.1 -0.2 -0.1 -0.1 -0.1 -0.1 -0.1 0.0 0.0 0.0 0.0 0.0 0.0 -0.1 0.0 -0.2 0.0
2nd tenth -0.3 -0.2 -0.1 -0.5 0.0 -0.3 -0.1 -0.1 0.4 0.6 0.6 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.2 0.1 0.0 0.1
3rd tenth 0.1 0.2 0.2 0.3 0.3 0.0 0.1 0.1 1.6 2.0 1.8 1.4 1.5 1.4 1.5 1.5 1.4 1.4 1.2 1.0 0.8 1.1
4th tenth 0.9 0.9 1.0 0.8 0.7 0.3 0.7 0.8 3.5 4.2 3.7 3.2 3.4 3.3 3.3 3.4 3.3 3.4 3.4 3.2 2.9 3.3
5th tenth 2.0 2.0 2.2 1.8 1.8 1.5 2.2 2.5 5.1 6.2 5.5 5.1 5.2 5.0 4.9 5.2 5.3 5.2 5.2 5.2 4.8 5.3
6th tenth 3.8 3.8 3.9 4.1 3.8 3.8 4.5 4.7 6.6 7.9 7.4 7.0 7.1 6.8 6.6 6.9 7.2 7.0 7.0 6.9 6.8 7.2
7th tenth 6.5 6.4 6.8 7.0 6.9 6.7 7.0 7.5 8.5 9.8 9.7 9.4 9.2 8.9 8.5 9.1 9.4 9.2 9.1 8.9 8.9 9.4
8th tenth 10.1 10.1 11.3 11.0 10.9 10.3 10.8 11.2 11.6 12.7 13.2 12.5 12.3 11.7 11.4 12.0 12.7 12.2 12.1 11.8 11.9 12.3
9thtenth 18.2 18.0 19.4 18.1 18.4 16.8 17.4 17.8 17.5 17.7 18.7 18.1 17.9 16.7 16.5 17.3 17.7 17.4 17.4 16.9 17.2 17.6
10th tenth 58.9 59.5 55.6 57.6 57.9 60.9 57.6 55.6 45.2 38.9 39.6 43.0 43.1 45.9 46.9 44.4 42.6 43.9 44.5 45.9 46.8 43.8
P10/Median (%) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 2.6 3.8 4.6 1.7 2.1 2.7 2.2 2.1 1.8 1.9 0.9 0.9 0.3 0.7
P20/Median (%) 0.0 0.0 0.0 4.7 6.0 0.0 0.0 0.0 14.1 15.3 14.6 11.8 11.7 13.0 13.0 12.3 11.0 10.0 7.5 6.4 4.2 5.3
P80/Median (%) 464.9 463.0 500.3 477.8 511.6 488.8 399.9 381.4 235.9 203.8 237.2 242.9 239.7 228.7 236.5 234.5 234.7 236.1 229.8 223.5 244.5 229.3
P90/Median (%) 872.2 869.9 896.6 837.0 948.0 899.7 706.4 661.6 384.5 314.5 356.2 382.7 381.4 362.4 366.9 370.1 355.5 354.6 359.3 349.7 374.2 359.8
Gini index 0.758 0.763 0.736 0.750 0.755 0.770 0.742 0.729 0.612 0.556 0.575 0.604 0.601 0.618 0.623 0.607 0.599 0.607 0.613 0.624 0.639 0.611
27
Table A3
Wealth distribution in 1968-75 and comparison with 1987-2014 –Per-capita wealth
(Share of wealth per tenth of households and inequality indices)
Estimates on synthetic data Estimates on Historical Archive data
1968 1969 1970 1971 1972 1973 1974 1975 1987 1989 1991 1993 1995 1998 2000 2002 2004 2006 2008 2010 2012 2014
1st tenth -0.3 -0.5 -0.2 0.0 -0.7 -0.1 -0.1 -0.1 -0.1 0.0 -0.1 0.0 0.0 -0.1 0.0 0.0 -0.1 0.0 -0.1 -0.1 -0.1 -0.1
2nd tenth -0.3 -0.2 -0.1 -0.5 0.0 -0.2 -0.1 -0.1 0.4 0.6 0.5 0.3 0.3 0.4 0.3 0.3 0.3 0.3 0.2 0.2 0.0 0.1
3rd tenth 0.1 0.1 0.2 0.3 0.3 0.0 0.1 0.1 1.5 1.9 1.7 1.4 1.4 1.4 1.5 1.4 1.3 1.3 1.1 1.0 0.8 1.0
4th tenth 0.8 0.8 1.0 0.7 0.7 0.3 0.7 0.8 3.2 3.9 3.3 3.1 3.3 3.0 3.0 3.1 3.0 2.9 3.0 2.8 2.6 2.8
5th tenth 1.8 1.8 2.1 1.7 1.7 1.5 2.1 2.4 4.9 5.7 5.1 4.8 5.0 4.6 4.5 4.8 4.8 4.6 4.6 4.6 4.3 4.6
6th tenth 3.5 3.5 3.7 3.8 3.6 3.6 4.1 4.5 6.4 7.4 7.0 6.7 6.8 6.3 6.2 6.6 6.7 6.3 6.4 6.3 6.2 6.5
7th tenth 5.9 5.9 6.5 6.6 6.4 6.3 6.6 7.2 8.4 9.5 9.3 9.1 9.0 8.4 8.1 8.7 9.1 8.6 8.7 8.5 8.5 8.8
8th tenth 9.5 9.4 10.8 10.5 10.5 10.0 10.4 10.9 11.6 12.4 12.8 12.3 12.1 11.3 11.1 11.8 12.4 12.0 12.0 11.6 11.6 12.0
9th tenth 17.4 17.2 19.3 17.7 17.9 16.6 17.1 17.6 17.2 17.6 18.4 17.8 18.2 16.8 16.8 17.5 17.9 17.5 17.9 17.3 17.6 18.1
10th tenth 61.5 62.1 56.8 59.3 59.7 62.1 59.2 56.8 46.4 41.0 41.9 44.5 43.9 47.9 48.7 45.7 44.5 46.5 46.2 47.9 48.5 46.0
P10/Median (%) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 2.6 3.9 4.8 2.0 1.9 2.5 2.1 2.0 1.8 2.0 0.9 0.9 0.3 0.6
P20/Median (%) 0.0 0.0 0.0 4.8 5.8 0.0 0.0 0.0 14.2 16.4 15.8 12.5 12.1 14.2 13.3 12.4 11.4 11.3 9.3 7.6 4.7 6.4
P80/Median (%) 484.6 485.0 511.7 509.2 530.9 518.3 433.0 395.6 242.0 217.3 250.3 248.9 246.2 247.3 247.5 241.9 244.6 259.0 262.8 258.2 270.7 256.1
P90/Median (%) 944.4 943.7 954.7 906.2 981.3 957.8 775.6 707.1 401.7 337.5 385.2 393.5 394.8 393.6 404.8 400.0 392.8 405.7 418.9 406.6 429.4 415.5
Gini index 0.772 0.778 0.743 0.760 0.765 0.778 0.752 0.737 0.623 0.576 0.593 0.616 0.611 0.635 0.640 0.621 0.617 0.631 0.632 0.644 0.656 0.633
Table A4
Income distribution in 1968-75 and comparison with 1987-2014 – Household income
(Share of income per tenth of households and inequality indices)
Estimates on synthetic data Estimates on Historical Archive data
1968 1969 1970 1971 1972 1973 1974 1975 1987 1989 1991 1993 1995 1998 2000 2002 2004 2006 2008 2010 2012 2014
1st tenth 1.8 1.9 1.8 1.5 1.6 1.6 1.9 2.1 2.3 2.8 2.7 2.1 2.1 2.0 2.2 2.4 2.6 2.6 2.6 2.4 2.4 2.1
2nd tenth 3.5 3.7 3.6 3.3 3.6 3.4 3.6 4.0 4.1 4.4 4.4 3.8 3.9 3.8 4.0 4.1 4.3 4.3 4.2 4.2 4.1 4.2
3rd tenth 4.8 5.2 5.1 4.8 4.9 5.0 5.1 5.3 5.3 5.5 5.6 5.1 5.1 5.0 5.2 5.2 5.3 5.4 5.3 5.3 5.3 5.4
4th tenth 6.0 6.4 6.3 6.1 6.4 6.3 6.4 6.5 6.3 6.6 6.7 6.3 6.3 6.2 6.4 6.3 6.4 6.5 6.4 6.5 6.4 6.5
5th tenth 7.4 7.6 7.6 7.5 7.7 7.5 7.6 7.8 7.5 7.7 7.9 7.4 7.5 7.4 7.6 7.6 7.4 7.6 7.6 7.6 7.5 7.6
6th tenth 8.4 8.9 8.9 8.9 9.0 8.9 9.0 9.1 8.8 9.1 9.3 8.9 9.0 8.9 8.9 9.0 8.8 9.0 8.8 9.0 8.8 9.0
7th tenth 9.9 10.4 10.3 10.5 10.6 10.5 10.7 10.7 10.6 10.7 11.0 10.8 10.6 10.6 10.7 10.6 10.4 10.5 10.5 10.6 10.6 10.8
8th tenth 12.3 12.5 12.6 12.6 13.0 12.7 12.9 12.8 12.7 12.7 13.0 13.0 12.9 12.7 12.8 12.7 12.5 12.5 12.6 12.6 12.7 13.0
9th tenth 16.3 16.1 16.2 16.4 16.2 15.9 15.9 15.6 16.0 15.7 15.7 16.1 15.9 15.8 15.7 15.7 15.5 15.3 15.7 15.6 15.9 16.0
10th tenth 29.5 27.4 27.6 28.4 27.1 28.1 26.9 26.1 26.4 24.9 23.6 26.4 26.6 27.5 26.6 26.3 26.7 26.3 26.4 26.0 26.4 25.3
P10/Median (%) 36.2 36.1 34.6 31.5 32.2 32.2 34.6 38.1 42.4 44.7 42.3 38.0 38.1 37.9 40.3 41.7 44.3 45.0 43.1 43.6 43.0 41.4
P20/Median (%) 53.1 53.8 54.2 50.9 51.2 51.6 51.9 55.0 58.7 59.2 59.0 56.3 55.5 54.4 56.9 56.6 60.4 58.6 58.7 58.1 58.7 59.6
P80/Median (%) 176.3 169.7 170.7 171.6 174.0 173.5 173.0 166.3 173.1 166.0 167.0 176.2 173.1 171.5 171.5 168.0 171.1 166.3 171.1 168.3 173.0 174.0
P90/Median (%) 244.5 227.9 229.0 232.7 222.3 223.6 217.0 208.5 229.0 214.2 203.9 229.5 220.4 221.5 218.4 219.1 222.4 213.2 223.9 216.7 224.9 219.9
P80/P20 3.3 3.2 3.2 3.4 3.4 3.4 3.3 3.0 3.0 2.8 2.8 3.1 3.1 3.2 3.0 3.0 2.8 2.8 2.9 2.9 3.0 2.9
P90/P10 6.8 6.3 6.6 7.4 6.9 6.9 6.3 5.5 5.4 4.8 4.8 6.1 5.8 5.8 5.4 5.3 5.0 4.7 5.2 5.0 5.2 5.3
Gini index 0.399 0.375 0.379 0.397 0.380 0.387 0.373 0.357 0.358 0.334 0.325 0.366 0.366 0.375 0.362 0.357 0.354 0.348 0.353 0.350 0.357 0.350
28
Table A5
Income distribution in 1968-75 and comparison with 1987-2014 – Equivalent income (square root scale)
(Share of income per tenth of households and inequality indices)
Estimates on synthetic data Estimates on Historical Archive data
1968 1969 1970 1971 1972 1973 1974 1975 1987 1989 1991 1993 1995 1998 2000 2002 2004 2006 2008 2010 2012 2014
1st tenth 2.0 2.2 2.1 1.7 1.9 2.2 2.3 2.7 2.8 3.5 3.4 2.5 2.4 2.2 2.4 2.6 2.8 2.9 2.7 2.5 2.4 2.2
2nd tenth 3.9 4.3 4.1 3.9 4.1 4.2 4.1 4.5 4.6 5.1 5.0 4.5 4.4 4.4 4.6 4.6 4.5 4.6 4.6 4.5 4.5 4.5
3rd tenth 5.0 5.4 5.3 5.1 5.4 5.4 5.3 5.6 5.7 6.0 6.1 5.6 5.6 5.6 5.7 5.8 5.6 5.8 5.7 5.7 5.7 5.7
4th tenth 6.1 6.5 6.4 6.3 6.7 6.4 6.4 6.8 6.7 7.0 7.2 6.6 6.8 6.8 6.8 6.9 6.8 6.9 6.8 6.9 6.8 6.9
5th tenth 7.3 7.7 7.6 7.6 7.8 7.6 7.6 7.9 7.8 8.1 8.3 7.9 7.9 8.0 8.0 8.1 8.0 8.0 8.0 8.0 8.0 8.2
6th tenth 8.5 9.0 8.9 9.0 9.1 8.8 9.0 9.2 9.1 9.2 9.5 9.3 9.2 9.2 9.3 9.3 9.1 9.2 9.3 9.4 9.3 9.5
7th tenth 10.0 10.4 10.4 10.5 10.5 10.2 10.6 10.6 10.5 10.7 10.9 10.9 10.7 10.6 10.8 10.6 10.6 10.6 10.7 10.8 10.8 11.1
8th tenth 12.1 12.4 12.4 12.4 12.6 12.2 12.7 12.4 12.5 12.4 12.6 12.6 12.6 12.3 12.5 12.4 12.3 12.3 12.4 12.5 12.5 12.7
9th tenth 16.0 15.5 15.9 15.8 15.6 15.3 15.5 15.0 15.5 14.9 14.8 15.3 15.2 14.9 15.0 15.0 14.8 14.9 15.0 15.1 15.2 15.2
10th tenth 29.1 26.7 26.9 27.8 26.3 27.6 26.5 25.3 24.8 23.2 22.2 24.8 25.2 26.0 25.0 24.7 25.4 24.8 24.8 24.5 24.8 23.9
P10/Median (%) 41.5 43.0 40.4 37.5 39.2 41.6 41.7 45.7 48.0 52.1 50.6 44.4 43.5 41.6 45.1 45.0 46.1 46.6 46.1 44.3 43.5 42.0
P20/Median (%) 57.4 58.3 58.2 54.7 56.5 58.9 56.2 60.0 60.3 64.1 62.6 59.6 58.5 58.8 59.6 59.8 59.8 61.8 59.1 59.2 59.7 58.8
P80/Median (%) 172.5 163.9 170.3 166.4 166.4 165.8 166.8 157.8 163.0 154.4 153.0 161.4 159.0 153.8 157.8 155.0 155.5 155.1 155.2 157.6 156.5 157.4
P90/Median (%) 237.7 215.9 222.7 220.4 212.0 216.6 212.4 200.9 208.4 192.3 185.5 204.8 198.3 197.6 196.5 197.4 197.1 197.2 195.8 195.7 197.8 192.8
P80/P20 3.0 2.8 2.9 3.0 3.0 2.8 3.0 2.6 2.7 2.4 2.4 2.7 2.7 2.6 2.7 2.6 2.6 2.5 2.6 2.7 2.6 2.7
P90/P10 5.7 5.0 5.5 5.9 5.4 5.2 5.1 4.4 4.3 3.7 3.7 4.6 4.6 4.8 4.4 4.4 4.3 4.2 4.3 4.4 4.6 4.6
Gini index 0.386 0.354 0.362 0.377 0.357 0.362 0.358 0.331 0.328 0.296 0.288 0.333 0.337 0.343 0.331 0.324 0.330 0.321 0.326 0.326 0.330 0.326
Table A6
Income distribution in 1968-75 and comparison with 1987-2014 – Per capita income
(Share of income per tenth of households and inequality indices)
Estimates on synthetic data Estimates on Historical Archive data
1968 1969 1970 1971 1972 1973 1974 1975 1987 1989 1991 1993 1995 1998 2000 2002 2004 2006 2008 2010 2012 2014
1st tenth 1,7 2,0 1,8 1,5 1,7 2,0 2,0 2,5 2,5 3,1 3,0 2,2 2,1 1,9 2,1 2,3 2,4 2,5 2,3 2,1 2,0 1,8
2th tenth 3,5 3,9 3,8 3,5 3,8 3,9 3,7 4,3 4,3 4,7 4,7 4,1 4,1 4,1 4,2 4,2 4,0 4,2 4,1 4,0 3,9 3,9
3th tenth 4,6 5,0 4,9 4,8 5,2 5,0 4,9 5,3 5,4 5,8 5,9 5,4 5,4 5,4 5,4 5,5 5,3 5,5 5,3 5,3 5,2 5,2
4th tenth 5,7 6,1 6,1 6,0 6,3 6,0 6,0 6,4 6,7 6,8 7,0 6,6 6,6 6,6 6,6 6,7 6,5 6,6 6,6 6,6 6,5 6,5
5th tenth 6,9 7,3 7,3 7,3 7,4 7,1 7,2 7,5 7,8 7,9 8,1 7,9 7,8 7,8 7,8 7,9 7,7 7,8 7,8 7,8 7,8 7,9
6th tenth 8,1 8,7 8,6 8,6 8,6 8,4 8,5 8,8 9,1 9,2 9,2 9,2 9,2 9,0 9,2 9,2 9,0 9,1 9,1 9,2 9,1 9,3
7th tenth 9,8 10,3 10,3 10,2 10,3 10,0 10,4 10,3 10,6 10,5 10,6 10,7 10,6 10,3 10,6 10,5 10,4 10,5 10,6 10,7 10,7 10,8
8th tenth 12,0 12,3 12,5 12,4 12,5 12,2 12,5 12,3 12,7 12,3 12,4 12,5 12,5 12,2 12,5 12,5 12,3 12,4 12,5 12,6 12,8 12,9
9th tenth 15,9 15,7 16,1 15,8 15,6 15,5 15,9 15,3 15,6 15,1 15,1 15,5 15,4 15,1 15,3 15,4 15,2 15,4 15,4 15,7 15,7 15,8
10th tenth 31,8 28,7 28,6 29,8 28,6 29,9 29,0 27,3 25,4 24,8 24,0 25,9 26,4 27,7 26,4 25,9 27,2 26,2 26,4 26,0 26,1 25,9
P10/Median (%) 39,2 40,1 38,3 34,8 37,7 41,9 39,2 44,8 43,2 47,7 46,8 40,1 38,7 38,5 40,5 40,9 40,4 41,2 39,9 38,8 39,0 36,2
P20/Median (%) 55,2 55,3 55,2 51,9 56,6 57,8 55,0 59,6 57,8 62,2 60,9 56,5 56,4 57,3 57,2 57,1 54,5 57,8 55,8 54,3 54,9 52,6
P80/Median (%) 184,3 170,0 173,9 172,8 171,5 177,1 179,1 166,3 164,9 158,1 156,1 160,3 161,0 159,9 161,0 159,0 159,7 162,9 161,4 161,7 165,4 163,6
P90/Median (%) 257,4 231,5 235,4 229,4 227,3 238,4 239,4 220,3 211,4 205,0 196,8 206,8 208,0 208,3 208,1 207,5 209,2 210,9 210,3 208,9 214,7 208,9
P80/P20 3,3 3,1 3,2 3,3 3,0 3,1 3,3 2,8 2,9 2,5 2,6 2,8 2,9 2,8 2,8 2,8 2,9 2,8 2,9 3,0 3,0 3,1
P90/P10 6,6 5,8 6,2 6,6 6,0 5,7 6,1 4,9 4,9 4,3 4,2 5,2 5,4 5,4 5,1 5,1 5,2 5,1 5,3 5,4 5,5 5,8
Gini index 0,418 0,383 0,387 0,403 0,384 0,392 0,390 0,358 0,342 0,321 0,313 0,349 0,354 0,366 0,353 0,346 0,360 0,347 0,354 0,355 0,359 0,361
29
Table A7
Distribution of income-net worth indicator in 1968-75 and comparison with 1987-2014
(Share of income-wealth per tenth of households and inequality indices)
Estimates on synthetic data Estimates on Historical Archive data
1968 1969 1970 1971 1972 1973 1974 1975 1987 1989 1991 1993 1995 1998 2000 2002 2004 2006 2008 2010 2012 2014
1st tenth 1.7 1.9 1.8 1.7 1.8 1.6 1.7 2.0 2.3 2.6 2.5 1.9 1.9 1.8 2.0 2.1 2.3 2.1 2.0 1.9 1.8 1.7
2nd tenth 3.3 3.6 3.6 3.4 3.6 3.4 3.5 3.7 3.9 4.3 4.1 3.5 3.5 3.5 3.6 3.6 3.7 3.6 3.4 3.3 3.2 3.4
3rd tenth 4.6 5.0 5.0 4.7 4.9 4.7 4.7 4.8 4.9 5.4 5.3 4.7 4.7 4.7 4.7 4.8 4.8 4.8 4.7 4.6 4.4 4.6
4th tenth 5.8 6.1 6.1 6.0 6.2 5.7 5.8 5.9 6.0 6.5 6.4 5.9 5.9 5.8 5.8 6.0 6.0 6.0 5.9 5.8 5.6 5.8
5th tenth 6.9 7.2 7.2 7.2 7.3 6.9 7.0 7.1 7.1 7.6 7.7 7.1 7.2 7.1 7.0 7.2 7.2 7.2 7.0 7.1 6.8 7.2
6th tenth 7.9 8.4 8.4 8.5 8.6 8.3 8.4 8.5 8.4 8.9 9.0 8.6 8.6 8.5 8.3 8.5 8.5 8.5 8.3 8.4 8.2 8.7
7th tenth 9.5 9.9 10.0 10.1 10.3 9.8 10.1 10.2 10.1 10.4 10.7 10.4 10.3 10.1 9.9 10.1 10.0 10.0 10.0 9.9 9.9 10.5
8th tenth 11.8 12.1 12.3 12.2 12.6 11.9 12.2 12.3 12.2 12.3 12.8 12.6 12.4 12.1 12.0 12.3 12.2 12.1 12.1 12.0 12.2 12.6
9th tenth 16.0 15.8 15.9 15.8 15.7 15.4 15.6 15.7 15.7 15.5 15.6 16.0 15.8 15.5 15.2 15.7 15.6 15.4 15.6 15.5 15.8 15.9
10th tenth 32.4 29.8 29.6 30.4 28.9 32.2 31.0 29.9 29.3 26.4 26.0 29.4 29.7 31.0 31.5 29.8 29.6 30.3 30.9 31.5 32.2 29.6
P10/Median (%) 37.0 38.2 36.7 34.2 35.9 34.1 35.9 38.7 42.5 44.5 41.0 35.5 35.7 36.0 38.9 38.3 40.4 38.4 37.3 35.8 35.3 34.0
P20/Median (%) 53.8 56.3 56.4 53.0 53.8 54.2 53.0 54.7 57.4 59.5 56.4 53.2 52.9 53.0 54.1 54.0 54.4 53.5 52.9 51.4 50.7 50.8
P80/Median (%) 183.6 173.8 176.9 175.3 174.2 174.7 175.5 174.8 175.6 166.0 167.0 177.5 174.5 173.6 172.9 177.3 174.0 170.3 176.2 175.4 182.2 174.7
P90/Median (%) 263.4 240.1 238.8 239.7 232.9 244.8 239.5 237.6 236.3 217.0 213.5 239.9 232.2 232.2 232.4 235.7 232.5 229.8 240.2 237.6 253.7 236.2
P80/P20 3.4 3.1 3.1 3.3 3.2 3.2 3.3 3.2 3.1 2.8 3.0 3.3 3.3 3.3 3.2 3.3 3.2 3.2 3.3 3.4 3.6 3.4
P90/P10 7.1 6.3 6.5 7.0 6.5 7.2 6.7 6.2 5.6 4.9 5.2 6.8 6.5 6.5 6.0 6.2 5.8 6.0 6.4 6.6 7.2 7.0
Gini index 0.424 0.394 0.396 0.406 0.390 0.421 0.412 0.397 0.385 0.348 0.352 0.398 0.398 0.409 0.408 0.396 0.389 0.398 0.407 0.414 0.427 0.404
Table A8
Household wealth inequality in 1968-75 and comparison with 1987-2014: adjustments for under-reporting
(Gini indices)
Estimates on synthetic data Estimates on Historical Archive data
1968 1969 1970 1971 1972 1973 1974 1975 1987 1989 1991 1993 1995 1998 2000 2002 2004 2006 2008 2010 2012 2014
Household wealth 0.758 0.764 0.739 0.747 0.754 0.772 0.744 0.737 0.628 0.577 0.591 0.624 0.619 0.629 0.631 0.619 0.604 0.616 0.615 0.627 0.643 0.613
Calibration - Method 1 0.780 0.781 0.759 0.776 0.783 0.738 0.733 0.745 0.697 0.607 0.592 0.663 0.668 0.710 0.694 0.657 0.644 0.690 0.691 0.684 0.703 0.655
Calibration - Method 2 0.799 0.789 0.787 0.801 0.816 0.756 0.735 0.746 0.750 0.695 0.703 0.743 0.738 0.782 0.739 0.708 0.694 0.746 0.746 0.728 0.742 0.732
Calibration - Method 3 0.789 0.786 0.768 0.786 0.794 0.757 0.736 0.746 0.705 0.623 0.608 0.673 0.676 0.702 0.712 0.676 0.665 0.699 0.687 0.698 0.721 0.667
Calibration - Method 4 0.780 0.781 0.759 0.776 0.782 0.754 0.734 0.745 0.688 0.603 0.590 0.656 0.660 0.685 0.694 0.657 0.644 0.683 0.672 0.681 0.703 0.648
Average of calibrations 0.787 0.784 0.769 0.785 0.794 0.751 0.735 0.745 0.710 0.632 0.623 0.684 0.685 0.720 0.710 0.674 0.662 0.704 0.699 0.698 0.717 0.676
Calibration methods used (Sautory, 1993): 1=Linear - 2=Raking ratio - 3=Logit - 4= Linear truncated (with parameters 0.1 and 0.9).
30
Table A9
Relative poverty and place of birth
(Percentages)
Share of poor born in Italy, using the indicator … Share of poor born abroad, using the indicator Share of poor (total residents in Italy), using the
indicator …
Year Share of
individuals born
abroad Income Income and
wealth (1
Year)
Income and
wealth (3
months)
Income and
wealth (3
years) Income Income and
wealth (1
Year)
Income and
wealth (3
months)
Income and
wealth (3
years) Income Income and
wealth (1
Year)
Income and
wealth (3
months)
Income and
wealth (3
years)
1968 14.8 7.1 6.5 9.3
1969 14.1 7.6 7.1 9.2
1970 14.8 6.7 6.1 8.5
1971 16.8 7.6 6.6 9.2
1972 15.8 6.6 5.5 8.9
1973 14.1 6.7 6.0 8.0
1974 15.4 7.5 6.7 9.1
1975 12.6 6.4 5.7 7.9
1987 11.4 5.2 4.2 6.1
1989 8.6 4.0 2.8 4.6
1991 0.9 9.5 4.8 3.4 6.0 12.2 3.3 3.3 8.0 9.5 4.8 3.4 6.0
1993 1.1 13.4 5.9 4.7 7.6 10.5 6.6 5.7 7.6 13.4 5.9 4.7 7.6
1995 1.3 14.0 6.9 4.8 8.4 11.6 9.6 8.4 10.1 14.0 6.9 4.9 8.4
1998 1.9 14.0 6.2 4.2 7.7 22.0 11.4 8.8 13.8 14.2 6.3 4.3 7.8
2000 2.5 12.9 5.6 4.0 7.0 15.0 10.4 10.0 11.7 13.0 5.7 4.1 7.1
2002 3.5 12.7 5.5 3.7 6.9 24.3 21.1 19.8 21.1 13.1 6.0 4.3 7.4
2004 4.4 11.9 5.7 4.0 6.9 22.1 19.0 14.7 19.5 12.4 6.3 4.4 7.4
2006 5.4 11.6 5.4 3.6 6.4 23.0 16.6 11.1 17.8 12.2 6.0 4.0 7.0
2008 7.5 11.1 5.3 3.9 6.3 33.7 24.5 20.8 26.8 12.8 6.7 5.2 7.8
2010 9.0 11.8 6.4 4.5 7.4 30.4 25.1 21.6 27.0 13.5 8.1 6.0 9.1
2012 11.1 10.9 5.7 4.4 6.6 34.3 27.2 23.4 29.3 13.5 8.1 6.5 9.1
2014 10.0 12.0 6.1 4.9 6.6 34.3 27.1 23.7 28.2 14.2 8.2 6.8 8.7
31
References
Acciari P., F. Alvaredo, S. Morelli (2017), The concentration of personal wealth in Italy
1995-2013, Paper presented at the Conference How financial systems work:
evidence from financial accounts, 30 November-1 December 2017, Banca d’Italia,
Rome.
Atkinson A.B., A. Brandolini (2013), On the identification of the middle class, in J.C.
Gornick and M. Jäntti (eds.), Income Inequality. Economic Disparities and Middle
Class in Affluent Countries, Stanford (CA), Stanford University Press, pp. 77-100.
Alfani G. (2016), The rich in historical perspective. Evidence for preindustrial Europe (ca.
1300-1800), Working Paper No. 571, IGIER – Università Bocconi.
Baffigi A., L. Cannari, G. D’Alessio (2016), Cinquant’anni di indagini sui bilanci delle
famiglie italiane. Storia, metodi, prospettive. Questioni di Economia e finanza
(Occasional Papers), No. 368, Banca d’Italia.
Banca d’Italia (1970), Risparmio e struttura della ricchezza delle famiglie italiane nel 1968,
Statistical Bulletin, anno XXV, No. 1, January-February.
Banca d’Italia (1973), Nota sull'esecuzione della indagine sul reddito e il risparmio delle
famiglie italiane nell’anno 1973, disponibile al sito:
http://www.bancaditalia.it/statistiche/tematiche/indagini-famiglie-imprese/bilanci-
famiglie/documenti-storici/note-metodologiche/index.html
Barrientos A. F., Bolton A., Balmat T., Reiter J.P., Figueiredo J. M., Machanavajjhala A.,
Chen Y., Kneifel C., DeLong M. (2017), A framework for sharing confidential
research data, applied to investigating differential pay by race in the US government,
NBER Working Paper Series No. 23534.
Brandolini A. (1999), The Distribution of Personal Income in Post-War Italy: Source
Description, Data Quality, and the Time Pattern of Income Inequality, Temi di
discussione (Working Papers), No. 350, Banca d’Italia.
Brandolini A., S. Magri, T. Smeeding (2010), Asset-based measurement of poverty, Temi di
Discussione (Working Papers), No. 755, Banca d’Italia.
Brick J. M., J. Montaquila, S. Roth (2003), Identifying problems with raking estimators,
Proceedings of the Annual Meeting of the American Statistical Association pp. 710–
717.
Cannari L., G. D’Alessio (1990), Housing Assets in the Bank of Italy's Survey of Household
Income and Wealth, in Dagum and Zenga (eds.), Income and Wealth Distribution,
Inequality and Poverty, Springer Verlag, Berlin, pp. 326-334.
Cannari L., G. D’Alessio (1993), Non-reporting and Under-reporting Behavior in the Bank
of Italy's Survey of Household Income and Wealth, in Bulletin of the International
Statistical Institute, vol. LV, n. 3, Pavia, p. 395-412.
Cannari L., G. D’Alessio, G. Raimondi, A.I. Rinaldi (1990), Le attività finanziarie delle
famiglie italiane, Temi di discussione (Working Papers, No. 136, Banca d'Italia.
Cannari L., G. D’Alessio, G. Vecchi (2017), Wealth, in Measuring wellbeing. A history of
Italian living standards, edited by G. Vecchi, Oxford University Press.
32
D’Alessio G. (2012), Ricchezza e disuguaglianza in Italia, in Disuguaglianze diverse, a cura
di D. Checchi, Il Mulino, 2012.
D'Alessio G. A. Neri (2015) Income and wealth sample estimates consistent with macro
aggregates: some experiments, Questioni di Economia e Finanza (Occasional
Papers), No. 272, Banca d’Italia.
D'Aurizio L., I. Faiella, S. Iezzi, A. Neri (2006), L’under-reporting della ricchezza
finanziaria nell’indagine sui bilanci delle famiglie, Temi di discussione (Working
Papers), No. 610, Banca d’Italia.
Deville J.C., C.E. Sarndal (1992), Calibration estimators in survey sampling. Journal of the
American Statistical Association 87(418): 376–382.
Gabbuti G. (2017), A Noi! Economic Inequality and the Political Economy of Italian
Fascism, Thesis submitted for the Master of Philosophy in Economic and Social
History, Trinity Term 2017
Gornick J. C., E. Sierminska, T.M. Smeeding (2009), The income and wealth packages of
older women in cross-national perspective. Journal of Gerontology: Social Sciences,
64B, 402-414.
Haveman R., E. N. Wolff, (2004), The concept and measurement of asset poverty: Levels,
trends and composition for the U.S., 1983–2001. Journal of Economic Inequality, 2,
145–169.
Pyatt G., C.N. Chen, J. Fei (1980), The Distribution of Income by Factor Component,
Quarterly Journal of Economics, 95(3): pp. 451-473.
Piketty T. (2014), Capital in the Twenty-First Century, Harvard University Press.
Piketty T., G. Postel-Vinay, J.-L. Rosenthal (2006), Wealth Concentration in a Developing
Economy: Paris and France, 1807-1994”. American Economic Review 96(1): 236-256
Reiter J. P. (2002), Satisfying Disclosure Restrictions With Synthetic Data Sets, Journal of
Official Statistics, Vol.18, No.4, pp. 531–543.
Rubin D. B. (1993), Discussion: Statistical Disclosure Limitation, Journal of Official
Statistics, Vol. 9, No. 2, pp. 461-468.
Sautory O. (1993), Le macro CALMAR, Redressement d’un échantillon par calage sur
marges, Document No. F 9310, INSEE.
Shorrocks A., G. Wan (2008), Ungrouping Income Distributions. Synthesising Samples for
Inequality and Poverty Analysis, UNU-WIDER Research Paper No. 16.
Short K., P. Ruggles (2005), Experimental measures of poverty and net worth: 1996, Journal
of Income Distribution, 13, 8–21.
Tanton R. (2014), A Review of Spatial Microsimulation Methods, International Journal of
Microsimulation 7(1) 4-25.
Weisbrod, B. A., and W. L. Hansen (1968), An Income-Net Worth Approach to Measuring
Economic Welfare, American Economic Review, vol.58, No. 5, pp. 1315(1329).
Williamson, P. (2013), An Evaluation of Two Synthetic Small-Area Microdata Simulation
Methodologies: Synthetic Reconstruction and Combinatorial Optimisation. In R.
Tanton & K. Edwards (Eds.), Spatial Microsimulation: A Reference Guide for Users,
pp. 19–47. Springer: Netherlands.
33
... Since microdata are not available for this survey and we only have information on the results published in the final report prepared at the time (Censis, 1995), the present work proceeded by applying a microdata reconstruction technique already tested on historical data (Cannari and D'Alessio, 2018). In particular, the survey on 1992 was taken as a reference, which is the closest in time to that of 1995 for which we intend to estimate microdata by modifying the sample weights of the units using raking 8 techniques until the constraints are satisfied. ...
Article
Full-text available
The paper shows the opinions on taxes of Italian citizens based on data gathered in four different national surveys between 1992 and 2013. Through a Principal Component Analysis, the study constructs a synthetic indicator of the propensity to evade, examining its intensity across various social groups and its evolution over time. The results show that the propensity to evade taxes is greater among households whose heads have low levels of education and income, are elderly and are resident in the South. Over time, the propensity to evade taxes has been growing on average, especially in the North, which has reduced the gap compared with the South, and among young people under 30 years old. The paper also shows a link between the propensity for tax evasion and some indicators of actual evasion, such as the use of cash and the under-reporting behaviour in the Survey of Household Income and Wealth (SHIW) conducted by the Bank of Italy, confirming the association between cultural elements and evasion behaviour.
Article
Full-text available
The Bank of Italy’s Survey of Household Income and Wealth (SHIW) is widely used to study the economic behavior of Italian households. Like most similar surveys, the SHIW is biased downward in its estimates by the lesser propensity of wealthy families to participate and by the tendency to underreport income and wealth. This work assesses the various techniques for correct the bias, applying them to the period 1995-2012. Calibration techniques, which produce estimates consistent with the macro-economic information available from other sources, are also employed.
Chapter
Full-text available
Availability of data on the amount and composition of wealth is of signal importance in order to interpret the patterns of household consumption and savings in the light of a life-cycle theory or other interpretative schemas keyed to the source and destination of accumulated wealth and the significance of intergenerational transfers. More precise measurement of household wealth is, moreover, the prerequisite to better quantification of the income components that it generates. These components generally are subject to underestimation of the same order as the corresponding components of wealth.
Article
Data stewards seeking to provide access to large-scale social science data face a difficult challenge. They have to share data in ways that protect privacy and confidentiality, are informative for many analyses and purposes, and are relatively straightforward to use by data analysts. We present a framework for addressing this challenge. The framework uses an integrated system that includes fully synthetic data intended for wide access, coupled with means for approved users to access the confidential data via secure remote access solutions, glued together by verification servers that allow users to assess the quality of their analyses with the synthetic data. We apply this framework to data on the careers of employees of the U. S. federal government, studying differentials in pay by race. The integrated system performs as intended, allowing users to explore the synthetic data for potential pay differentials and learn through verifications which findings in the synthetic data hold up in the confidential data and which do not. We find differentials across races; for example, the gap between black and white female federal employees' pay increased over the time period. We present models for generating synthetic careers and differentially private algorithms for verification of regression results.
Research
This article provides an overview of long-term changes in the relative conditions of the rich in preindustrial Europe. It covers four pre-unification Italian states (Sabaudian State, Florentine State, Kingdom of Naples and Republic of Venice) as well as other areas of Europe (Low Countries, Catalonia) during the period 1300-1800. Three different kinds of indicators are measured systematically and combined in the analysis: headcount indexes, the share of the top rich, and richness indexes. Taken together, they suggest that overall, during the entirety of the early modern period the rich tended to become both more prevalent and more distanced from the other strata of society. The only period during which the opposite process took place was the late Middle Ages, following the Black Death epidemic of the mid-fourteenth century. In the period from ca. 1300 to 1800, the prevalence of the rich doubled. In the Sabaudian State, the Florentine State and the Kingdom of Naples, for which reconstructions of regional wealth distributions exist, in about the same period the share of the top 10% grew from 45-55% to 70-80% - reaching almost exactly the same level which has recently been suggested as the European average at 1810. Consequently, the time series presented here might be used to add about five centuries of wealth inequality trends to current debates on very long-term changes in the relative position of the rich.
Chapter
An essential requirement for all microsimulation models is an initial set of population microdata. If processes are to be modelled at the subregional level, these microdata must be spatially detailed. Unfortunately, confidentiality and sample size restrictions preclude the provision of such microdata via either census or social survey. As a result, a number of techniques have been developed to synthetically generate the requisite microdata. This chapter reports on two techniques developed in the UK: ‘synthetic reconstruction’ and ‘combinatorial optimisation’. The basic methodology of each approach is outlined, following which recent innovations in their detailed implementation are introduced. The performance of both approaches is necessarily affected by the nature of between-place differences, which are shown to be surprisingly resistant to conventional data reduction techniques. Having considered this spatial variability, a new framework is introduced for the evaluation and validation of synthetic microdata, which includes multiple realisations and the use of measures of fit based around the Z-score and two derivations: Σ Z 2 and RSSZ. Finally, an evaluation of the output from each approach is presented. Combinatorial optimisation is shown to produce estimates with less bias, and significantly less variance, than those produced via synthetic reconstruction. The resulting synthetic microdata are shown to have a very high degree of fit to estimation constraints and to produce good estimates for margin-constrained distributions.
Article
To avoid disclosures, Rubin proposed creating multiple, synthetic data sets for public release so that (i) no unit in the released data has sensitive data from an actual unit in the population, and (ii) statistical procedures that are valid for the original data are valid for the released data. In this article, I show through simulation studies that valid inferences can be obtained from synthetic data in a variety of settings, including simple random sampling, probability proportional to size sampling, two-stage cluster sampling, and stratied sampling. I also provide guidance on specifying the number and size of synthetic data sets and demonstrate the benet of including design variables in the released data sets.