Content uploaded by Olivier Thevenon
Author content
All content in this area was uploaded by Olivier Thevenon on Feb 15, 2014
Content may be subject to copyright.
DOCUMENTS
DE TRAVAIL202
Olivier Thévenon
and Angelica Salvi Del Pero
Gender Equality
(f)or Economic Growth?
Effects of Reducing the Gender Gap in Education
on Economic Growth in OECD Countries
1
GENDER EQUALITY (F)OR ECONOMIC GROWTH?
Effects of Reducing the Gender Gap in Education on Economic Growth in OECD Countries
Olivier Thévenon,
INED£
Angelica Salvi Del Pero
OECD¥
£INED  Institut National d’Etudes Démographiques
133, Boulevard Davout
75980 Paris Cedex 20, France
tel: +33 1 56 06 22 44
fax: +331 56 06 21 94
olivier.thevenon@ined.fr
¥OECD Organisation for Economic Cooperation & Development
Social Policy Division
2 rue André Pascal
75775 Paris Cedex 16, France
tel: +33 1 45 24 91 53
angelica.salvidelpero@oecd.org
JEL codes: J16 (Economics of Gender), J21 (Labor Force and Employment, Size, and Structure), J24
(Human Capital), O4 (Economic growth and productivity)
Key words: Gender, human capital, Solow growth model, labour force participation, economic growth.
We thank Andrea Bassanini, Willem Adema, Nabil Ali, Fabrice Murtin and Stefano Scarpetta for helpful discussions and for
providing us with valuable comments on an earlier draft of the paper. Thanks also to Martina Portanti for her contribution to data
preparation, and to Andrea Bassanini for sharing his programs. This study was part of a larger project investigating the gender
differences in education, employment and entrepreneurship (OECD, 2012b). The views expressed in this paper are those of the
authors and do not necessarily reflect those of the OECD or its member countries.
2
ABSTRACT
This paper assesses the extent to which the increase in women’s human capital, as measured by
educational attainment, has contributed to economic growth in OECD countries over the past five decades.
Using longitudinal crosscountry data covering 30 countries from 1960 to 2008 on education (the Barro
Lee dataset) and growth (update of OECD data), our results point out a positive and significant impact of
the increase in women’s educational attainment relative to men on output per capita growth – as measured
by GDP per capita. Our results are robust to the distinction between subperiods and indicate that the effect
of the equalisation of years of completed education on economic growth has been higher in the most recent
periods. Results also hold when countries with an aboveaverage increase in years of completed education
are removed from the sample.
RÉSUMÉ
Ce papier évalue dans quelle mesure la croissance du capital humain détenu par les femmes, tel qu’il
est mesuré par les années d’éducation, a contribué à la croissance économique des pays de l’OCDE au
cours des cinq dernières décennies. Mobilisant des données couvrant 30 pays de 1960 à 2008 sur
l’éducation (Base de données de Barro et Lee) et la production économique (données OCDE actualisées),
nos résultats pointent un effet significatif et positif de l’augmentation du nombre relative d’années
d’éducation des femmes par rapport aux hommes sur la croissance du PIB par tête. Nos résultats sont
robustes à la distinction de souspériodes, et semblent indiquer que l’effet de l’égalisation du nombre
d’années d’étude sur la croissance économique a été plus important dans les années plus récentes. Les
résultats sont aussi confirmés lorsque les pays ayant connu une croissance des années d’éducation
significativement supérieurs à la moyenne sont retirés de l’échantillon.
3
1. Introduction
Investment in human capital improves the economic and social opportunities of young individuals, thereby
helping to reduce poverty and foster technical progress. In addition to the direct effects of education on
economic participation, education – especially female education – also affects other societal outcomes such
as child mortality, fertility, individual health outcomes, and the investment in the education and health of
future generations. Investing in women's human capital is therefore key to economic growth and social
cohesion, especially in developing countries where the gender gap in education is still large.
A large body of theoretical and empirical analysis exists on the link between investment in human capital
and economic growth (see the literature surveyed in Thévenon et al., 2012). The empirical evidence on the
relationship between economic growth and gender equality in the distribution of education is instead not as
conclusive. This paper tests this relationship in an internationally comparative perspective using a human
capital growth model augmented to include the effect of gender inequality in educational attainment, using
pooled longitudinal crosscountry data covering 30 OECD countries over the 19602008 period.
The structure of the paper is the following. The next section provides a review of the relevant literature;
section 3 describes the growth model; section 4 discusses the econometric approach for the growth model;
data used in the growth model is reviewed in section 5; section 6 provides the results of the growth model
before summarizing our main points in the conclusion.
2. Education and growth since the 1960s
It has now become a widespread argument in developing but also in more economically advanced
countries that the economic gains from educating girls are greater than those from educating boys (Schultz,
2002). There are various reasons to believe that female and male education affect output levels and growth
in different ways. Female education, just as male education, promotes growth by expanding the skilled
workingage population and by improving the productivity of the female labour force (Mammen and
Paxson, 2000) both directly – by raising output levels – and indirectly – through the increase in physical
capital investment and technological change that follow from higher output levels (Barro and SalaiMartin,
1995). A balanced distribution of education among men and women is also likely to foster economic
growth if male and female human capital are production factors with diminishing returns and are
imperfectly substitutable (Knowles et al., 2002). Moreover, higher education levels among women are
argued to produce additional social gains by reducing fertility and infant mortality, increasing life
expectancy, and increasing the quantity and quality of investments in children education (Schultz, 1988).
This spillover effect is expected to occurr even if not all educated women enter the labour market and
female labour force participation rates remain lower than those of males.
Over the past decades women have benefited from growing access to postelementary and postsecondary
education, especially in EU and other OECD countries (OECD, 2012b). Yet, there is so far not much
empirical evidence and consensus on the extent to which the reduction in the gender gap in education has
contributed to foster economic growth.
For instance, Barro and Lee (1994) and Barro and SalaiMartin (1995) – based on a panel of 138 countries
– report the puzzling finding that years of schooling have an effect on economic growth that is positive for
men but negative for women. They suggest that this is related to the fact that a high spread between male
4
and female secondary attainment is a measure of «backwardness» in the returns of education, and that
higher female education attainment implies that countries have reached a stage of economic development
from which no rewards can be expected from an additional increase in years of education (Barro and Lee,
1994). Their interpretation of the coefficients of male and female human capital is not straightforward,
however, since they may reflect not only sexspecific educational attainment, but also the the gender gap in
attainment (see below). Persistent gender differences in the chosen field of education and gender
differences in the labour market returns to educational investment might be another explanation.
Furthermore, the relationship between education, female employment and growth changes over time and
through the process of economic development. The extent to which the aggregate economic output
responds to female education and to the increase in female labour market participation is found to be U
shaped, i.e. positive only after countries have reached the “industrial” stage of development and women
have gained access to more productive sectors of economic activity (Goldin 1995; EsteveVolart, 2000;
Mammen and Paxson, 2000; Lincove, 2008 and Tam, 2011). The increase in girls’ access to post
elementary education is certainly a driver of this process, as suggested by the evidence of a convex
relationship between the decrease in gender discrimination in primary schooling and growth reported by
EsteveVolart (2000).
The Barro and Lee (1994) results have been challenged by several other empirical analyses. Dollar and
Gatti (1999) show, for example, that the negative association between the years of schooling completed by
girls and per capita output growth disappears once countryspecific factors are controlled for. Even more
challenging are those results of Caselli et al. (1996) and Forbes (2000) who come to the opposite
conclusion. Using a Generalized Methods of Moment estimation to control for the potential endogenity of
human capital variables and measurement errors, female schooling is found to be associated with a
statistically significant positive coefficient while a negative coefficient is obtained for male schooling. The
authors interpret the positive sign for female education as the net effect of a positive impact due to the
influence of education on fertility out weighting the negative “direct” effect of human capital found by
Barro and Lee. The negative coefficient for males, by contrast, reflects only the human capital effect. Yet,
the negative human capital effect remains ‘puzzling’ as it runs against most theoretical arguments (Topel,
1999; and, Krueger and Lindhal, 2001).
EsteveVolard (2000), Klasen (2002) and Knowles et al. (2002) emphasize the problems affecting the
empirical identification of effect of male and female education due to the multicollinearity caused by their
close correlation. Their suggested empirical strategy involves using a direct measure of the gender gap in
educational attainment in the growth equation instead of separate measures of male and female education.
Klasen (2002) finds that gender inequality in the initial levels and in the expansion of education
significantly reduces economic growth; the results – which are robust to different specifications of the
education variables and control for possible simultaneous determination of human capital stocks and
economic growths – are consistent with those of Knowles et al. (2002) and the older results by Hill and
King (1995). Knowles et al. (2002) further suggest that female education has a statistically significant
positive effect on the productivity per worker, while the role of male education is less clear and depends on
the inclusion of benchmark (baseperiod) values of human capital and other control variables.
In addition to differences in the chosen data and econometric strategy, many other reasons can explain the
conflicting results in the literature. Differences in stage of economic development can be one factor: less
developed countries may experience higher growth rates in spite of lower educational attainment due to
convergence mechanisms. Thus, the Barro and Lee (1994) results may be essentially driven by the
5
inclusion of the four East Asian Tigers and countries in SubSahara Africa (Stokey, 1994; and, Lorgelly
and Owen, 1999). This drawback motivates our choice to limit our sample to OECD countries, which are
relatively homogeneous in terms of stage of development and education trends.
This paper attempts to reassess the influence of gender differences in educational attainment on the long
run steadystate of economic output. There are different reasons for doing so. First, we want to update the
conclusions and allow for longer term assessment by taking into account recent trends in education. Our
sample covers the 30 OECD countries from 1960 to 2008 and we use the education data published in the
revised and cleaned BarroLee dataset (Barro and Lee, 2010). We also use the updated version of the data
GDP per capita published by the OECD. Our goal is also to address some of the econometric issues that
were overlooked by many of the previous assessments. Firstly, we tackle the assumption that the effect of
the determinants of growth (including physical and human capital) is homogenous across countries. This
assumption can be weakened by considering that countries will converge towards the same set of economic
steadystates in the longrun, but at a different pace. This assumption should not be overly strong as we
restrict our analysis to OECD countries, which have access to common technologies and share intensive
intraindustry trade and foreign direct investment (Arnold et al., 2011, p. 6) and are therefore quite
homogeneous in terms of stage of development, specialization, technological and institutional settings
(Pesaran and Smith 1995; Durlauf et al., 2005; Pedroni, 2007; and, Eberhardt and Teal, 2011). This
assumption is supported by Kourtellos (2011) and Di Vaio and Enflo (2009) that show how two growth
regimes have emerged post World War II: a group of countries characterized by higher growth rates and
convergence of per capita income (mainly OECD countries) and another group characterized by divergent
and lower growth rates.
Against this backdrop, we model the steadystate level of output per capita as a function of the propensity
to accumulate physical and human capital, the population growth rate, the level and growth rates of
technological and economic efficiency, and the (constant) rate of depreciation of capital, as set by Mankiw
et al. (1992) in the humancapitalaugmented Solow growth model. The model is then tested with
estimation procedures based on more or less flexible assumptions regarding the convergence process
towards steadystates. We control for the incidence of countryspecific and timeconstant (but
unobservable) factors shaping economic efficiency by means of a fixedeffect panel approach. The
inclusion of countryspecific time trends also allows us to capture changes in technologies or social
institutions that affect economic efficiency, even though they are not explicitly modelled (Pedroni, 2007).
We thus follow quite closely the perspective adopted by Bassanini and Scarpetta (2002) and Arnold et al.,
(2010), but we add the gender dimension to the impact of educational attainment.
The next two sections present respectively the theoretical framework and the empirical setting. Basic
statistics and data properties are then discussed, before presenting the regression results.
3. Theoretical background and econometric approach
The humancapitalaugmented Solow growth approach first presented by Mankiw et al. (1992) provides an
adequate framework to model the influence of education on growth. Following Arnold et al. (2011), we
consider a humancapitalaugmented Solow model with a standard CobbDouglas production function and
account for shortrun components as annual data is used to estimate the model (Pesaran and Smith, 1995;
Bassanini and Scarpetta, 2002; and, Arnold et al. 2011):
6
[1]
where and are the output and physical capital
quantities per effective unit of labor; stands for the average human capital
which sums the contribution of formal education completed by men and women; is the investment rate
in physical capital; and is the growth rate of labor. Arnold et al. (2011) prove that this reduced form of
output growth is also compatible with a twosector AK model à la “UzawaLucas” model of economic
growth.
In this context, gender differences in education can be captured by including a logratio (lnR) measuring
the difference in years spent on average in education by women compared to men:
[2]
Rf/m is defined as the femaleto male ratio in the average number of years spent in the education system by
men and women aged 25 to 64 ratio and is used to capture gender gaps avoiding the multicollinearity
problems encountered when the average years of education of men and women are included separately.The
growth equation described in [2] can be rewritten within an errorcorrection model where growth rates are
expressed as a function of the longrun evolution of the steadystate and of shortrun variations, as
appropriate for an empirical estimation based on pooled crosscountry annual data (Lee et al., 1997; Bond
et al. 2004; Bassanini and Scarpetta, 2002; and, Arnold et al., 2011). An alternative common technique to
reduce the influence of shortrun variation is to take averages of the data, typically over 5 years (Islam,
1995; Caselli et al., 1999; and, Bond et al. 2001) but this would not fully take advantage of the complete
set of information provided by the annual data.
[3]
where subscripts indicate country (i) and time (t) and i is the countryspecific speed of adjustments
(
). We thus assume that the steadystate of the growth rate of per capita output depends on
countryspecific factors that may shift the longrun path of economic development and/or produces short
run differences in the convergence towards the steady state of the economy.
This approach has the desirable feature of estimating a dynamic specification with countryidiosyncratic
speed of convergence and deviations from the steadystate, while still allowing for countryspecific
production levels and growth rates. Another advantage is to overcome the fact that the output steadystates
are unobservable, and observed changes in output per capita may well depend on shifts in the steadystate
output per capita arising from other factors than technology. These circumstances make it necessary to
have an empirical setting that clearly separate the longrun evolution of the steady states levels of output
from their transitory variations.
7
The estimation of equation [3] also requires a choice on the allowed degree of heterogeneity across
countries. To account for the fact that population and productivity growth patterns may vary considerably
across countries (Lee et al., 2007; Bassanini and Scarpetta, 2002b; and, Eberhardt and Teal, 2011) the
model will be estimated with a Pooled Mean Group (PMG) estimator, which assumes that countries
converge towards the same steadystate but the speed of adjustment can differ across countries (Lee et al.,
1997; and, Pesaran et al., 1999). The PMG estimator is more restrictive than a Mean Group estimator –
which does not imposes longrun coefficients to be the same across countries – but it still allows shortrun
variations in the pace of adjustment.
1
This approach is consistent with the fact that the production functions
of OECD countries are progressively becoming more homogenous on account of access to common
technologies, intensive intraindustry trade and large foreign intraOECD direct investment (Arnold et al.
2011). Another advantage of the PMG estimator is that it is not affected by the ‘downward bias’ in the
estimated coefficients that characterizes Dynamic FixedEffects estimators and Mean Group estimators
when the lagged dependent variable is endogenous to the fixed effect in the error term (Nickell, 1981).
A few caveats are worth discussing before estimating the model. First, the main requirement to implement
the PMG estimator is to have a large Nlarge T panel. As a consequence it is necessary to use panel data
with annual observations, which will prevent us from using reduced panels with data averaged by periods
of 5 years used by other studies on economic growth . Secondly, longrun growth accounting equations are
relations between variables in levels, which may not be stationary. Regressions are likely to be spurious if
the production function relating the variables in the long run is not a cointegrating vector, and the standard
statistics used to assess the quality of adjustment (R2 and standard errors) no longer apply (Philipps and
Perron, 1999; and, Kao, 1999). For this reason, we employ an ECM which encompasses the true model:
the levels terms are dropped out if there is no cointegration whereas they form a stationary relationship if
there is cointegration. We also check that all panel residuals are stationary with the unit root test assuming
that panel units are crosssection independent (Im et al., 2003) or not (Pesaran, 2007).
Exogenous changes in the production technology and in the institutions that condition the efficiency of the
production function can also introduce crosscountry heterogeneity in the pace of convergence towards the
steadystate. These differences cannot be observed and are proxied in our model by countryspecific trends,
which we define through a sequence of 5years dummies. Time trends, however, have a limited accuracy in
accounting for the unobservable changes if technology parameters don’t vary randomly across countries
and are correlated with the regressors and/or the errors terms (Durlauf et al., 2005). This misspecification
can have serious implications if the observable and/or unobservable variables are nonstationary: spurious
results may come from a failure to account for heterogeneity in the technology parameter, which leads to a
breakdown of the cointegrating relationship between inputs and output (Kao, 1999; Smith and Fuertes,
2007; and, Eberhardt and Teale, 2011). This can occur, for example, when countries experience a common
shock or are exposed to same processes (even if not with the same strength), which creates crosssection
1
Another way to put it is that the Mean Group estimator consists of an unweighted average of countryspecific (longrun)
coefficients and yields consistent estimates but is very sensitive to outliers. The PMG can be instead thought of as a
weighted average of individual group estimators, with weights proportional to the inverse of their variance; this allows
for heterogenous shortrun coefficients, but it constraints longrun parameters to be the same across countries. The
PMG therefore exploits the efficiency of the pooled estimation while avoiding the inconsistency problem of pooling
heterogenous dynamic relationships. The poolability restriction of the longrun parameters is tested using a Hausman
type test applied to the difference between the MG and the PMG estimator (see Tables A2 and A3)
8
dependence between panel units. This can be detected with a test of the crosssection independence of
residuals provided there are enough common observations across panels (Pesaran, 2004).
2
Last, as in any panel regression using macrolevel data, attention should be paid to possible endogenity
between economic growth and the accumulation of human and/or physical capital. Using stock data for
human capital, lagged values of the explanatory variables and a rather long panel mitigates the reverse
causality problem but it does not settle the issue completely since ∆Yit1 and εit might be correlated.
However, as shown in Pesaran (1997) and further discussed in Pesaran and Shin (1999), if the independent
variables have a finite order autoregressive representation, augmenting the ARDL specification with an
adequate number of lags makes the estimation of the longrun coefficients immune to endogeneity
problems. Pesaran further suggests that a twostep strategy whereby the lag orders of the ARDL model is
first selected using either the Akaike Information Criterion (AIC) or the Schwartz Bayesian Criterion (SC),
and then the longrun coefficients are estimated on the basis of the selected model, performs reasonably
well in mediumsized samples. To satisfy this requirement, we checked the robustness of results to
different lag structures in the independent variables and results are not different from those presented here.
As shown in Annex Table A1, the best model according to the Schwartz Bayesian Information Criterion
(BIC) would include one lag of the physical and human capital and the population growth. Only minor
parameter differences emerge from this and our baseline specifications. Moreover, the key test statistics are
robust to this type of changes in the specification (see the Annex –detailed results available on request).
3
4. Data overview
One of the key features of education trends in the past decades is the drastic rise in women’s educational
attainment and the decline of gender inequalities in education that took place in most regions of the world.
In OECD countries, primary school enrolment is nowadays nearly universal and gender equal (OECD,
2012b), while the picture is more mixed at secondary and postsecondary level. In the past decades the
increase in postsecondary education graduation rates has been greater among women than among men
across OECD countries, and – except for Turkey – boys are nowadays more likely than girls to drop out
before completing secondary education. As a result, younger women increasingly have higher educational
attainment than young men in the OECD (OECD, 2012b).
Qualitative differences in education persist, however. PISA data show that boys lag behind girls in reading
skills at the end of compulsory education to the equivalent of a year of schooling, on average. In many
countries boys are ahead in mathematics, but the gender gap is overall small compared with reading.
2
An appropriate strategy to account for the incidence of the common unobserved factors is the Common Correlated Estimator
proposed by Pesaran (2006), which includes crosssection averages of the dependent and independent variations in the
regression equation. However,
3
GMM with lagged instrumental variables can also be used to address the possible endogeneity of education and/or physical
capital. Nor the difference or the system GMM, proposed respectively by the Arellano and Bond (1991) and Blundell
and Bond (1998), provide here a very convicing approach due to data specificities. On the one hand, by construction,
yearly observation for education are obtained by interpolation, which implies that the use of lagged values of education
provides invalid and/or weak instruments, which is problematic for the estimation of difference and system GMM
models (Bond et al., 2001; Bun and Windmeijer, 2010; Bazzi and Clemens, 2013). On the other hand, the same
argument and the superiority of PMG estimators over MG estimates (see next section on results and tables A2 and A3)
refrain us from using Mean Group GMM estimators to address the possible endogeneity issues, as used for example in
Bond et al. (2010) and Arnold et al. (2011).
9
Differences in the field of study are also quite large: girls are significantly less likely to choose scientific
and technological fields of study in tertiary education, and when they do are then less likely to take up a
career in these fields.
Results on the relationship between human capital accumulation and growth are sensitive to data quality
and to the variables used to measure educational attainment and gender differences. The choice of data and
indicators used to measure educational attainments and compare them across gender is crucial in driving
results. To illustrate this point, Barro (1999) uses an updated version Barro and Lee (1996) data set on
education and no longer finds a negative role for female education.
4
In this paper we measure education as
the average number of years of schooling attained by the adult male population over the average number of
years attained by the adult female population. Different datasets can be used to obtain information on the
level of education,
5
but only few of them provide education data disaggregated by gender with sufficiently
long time series: the data collected by Lutz et al. (2007) and those collected by Barro and Lee and updated
recently (Barro and Lee, 2010). In this paper we use the latter because it provides longer time series – from
1960 to 2010 instead of 19702000 in the Lutz et al. (2007) – and because the 2010 version of the data has
been revised to address most of the concerns raised by critics on the former versions, including those of
Cohen and Soto (2007) and de la Fuente and Domenech (2006).
6
The data are reported at 5year intervals
from 1950 to 2010; we use a linear interpolation between each reported value to obtain annual data.
7
Educational attainment is measured as the number of years of education completed by men and women
from 25 to 64 years (to limit the bias due to incomplete education). This indicator is preferred to enrolment
rates because it captures the stock of education and because data on enrolment rates are affected by to
crosscountry differences and changes over time in the classification of educational attainment. The quality
of crosscountry data on years of education remains, however, not fully satisfactory. Morrison and Murtin
(2013), who estimated their own database of growth rates of years of schooling, find inconsistencies with
the Barro Lee dataset for a number of countries. Most notably, the years of schooling are rather low in a
number of European countries. Unfortunately Morrison and Murtin (2013) don’t provide data by gender so
their dataset cannot be used for the purposes of this paper.
Table 1 reports the average evolution of years in education for men and women in the 30 OECD countries
8
from 1960 to 2008, i.e. the period preceding the current economic crisis.
9
The average number of years
4
To a large extent this results stems from the revision of the education data that was undertaken to improve its quality and improve
the consistency of time series. The influence of data quality is also emphasized by De la Fuente and Domenech (2006) who find
overall a positive correlation between the quality of data and the significance of human capital coefficients in growth regressions.
Crespo Cuaresma (2005) finds important differences in the distribution and evolution of education in OECD countries across
different datasets (namely the data collected respectively by BarroLee, CohenSoto and De la FuenteDomenech).
5
See in particular de la Fuente and Domenech (2006) for a review of available data, and the data collected by Cohen and Soto
(2007), Lutz et al. (2007) or Morrisson and Murtin (2009).
6
More precisely, the data on education are derived from census and survey data obtained from Unesco and Eurostat used to
contruct estimates of various levels of educational attainment. Missing observations are estimated by extrapolating backwards
and forwards from census and survey data. Thes estimates are corrected for mortality rates that are allowed to differ across
different education cohorts. Preliminary testing shows, for OECD countries, that this version provides smoother time profiles for
educational attainment in Norway and the United States than the former versions of the dataset (Barro and Lee, 2010).
7
Note that from the statistical point of view, this interpolation is likely to smooth the trends in educational attainment, and to some
extent limit the remaining erratic movements which might be due to error measurement.
8
The countries covered are: Australia, Austria, Belgium, Canada, Chile, the Czech Republic, Denmark, Estonia, Finland, France,
Germany, Greece, Hungary, Iceland, Ireland, Israel, Italy, Japan, Korea, Luxembourg, Mexico, the Netherlands, New Zealand,
10
spent in education by men and women has steeply increased, and there are small gender differences in the
level of education (11.19 years for men and 11.21 for women). Yet, the increase has been steeper for
women than for men.
Qualitative differences in educational attainment and trends are not completely captured by the evolution
of years in education. Table 1 provides descriptive statistics on the proportion of men and women with
completed secondary education.
10
By this indicator, gender differences are quite significant even though
the gap is closing on average (Annex Figures A1). Overall, the increase in completion of secondary
education has been slightly steeper for women than for men: in the OECD the percentage of women aged
25 to 64 with completed secondary education was 11.7% in 1960 (vs. 16.6% for men), while the proportion
in 2010 is at 54.5% slightly higher than for men (Table 1). Qualitative differences in educational
attainment across gender remain therefore important despite the decrease in gender gap in terms of quantity
of education and differentiating between levels of education may have an impact on results as found by
Barro and Lee (1996), who identify a strong effect of secondary and higher schooling on growth. On the
other hand, the 1997 revision of the education classification system poses significant consistency issues in
the data on the level of education. For this reason and to limit the issue of multicollinearity that would arise
by including more education variables, in this paper we will measure education only as the number of
years of completed schooling.
Data on GDP, physical capital and working age population are taken from the OECD’s Economic Outlook
(No 90) data series. GDP per capita and Gross fixed capital formation (physical capital) are expressed in
2005 USD, taking advantage of the OECD’s update of calculated time series which changed the base
reference year from 2000 to 2005.
Table 4.1 also reports the average values of main variables entering in the growth equation: the dependent
variable (ΔlnY), measured as the growth in real GDP per head of population aged 1564 years; the
convergence variable (lnYt1), measured as the lagged real GDP per head of population aged 1564 years;
the propensity to accumulate physical capital (lnSK), proxied by the ratio of real private nonresidential
fixed capital formation to real private GDP; population growth (ΔlnN), measured in the working age
population (1564 years).
Norway, Poland, Portugal, the Slovak Republic, Slovenia, Spain, Sweden, Switzerland, Turkey, the United Kingdom, and the
United States. Country level figures are provided in Annex Figure A1.1.
9
Data series on GDP per capita and capital formation show indeed important breaks for years 2009 to 2010, as expected from the
consequences of the economic crunch.
10
The same information could have been used for people with tertiary education. However, it is frequently the case that this latter
is not completed before the age of 25, therefore introducing bias in the measurement.
11
Table 1. Basic Statistics
Variables
Year
Sample mean
Standard deviation
GDP per capita (in USD at 2005 PPP)
1960
16295
5711
1990
35503
10622
2010
45202
17948
Average years of education – Men
1960
6.49
2.16
1990
9.65
1.72
2010
11.19
Average years of education – Women
1960
5.79
2.31
1990
9.01
2.02
2010
11.21
1.56
% of men age 2564 with completed secondary
1960
16.6
11.6
education and above
1990
40.4
14.8
2010
53.2
18.1
% of women age 2564 with completed secondary
1960
11.7
10.7
education and above
1990
36.2
13.9
2010
54.5
15.4
% per capita annual growth rate of capital
2.13
8.51
% annual growth of male working age population
1.00
0.85
% annual growth of female working age population
0.91
0.94
Sources: Barro and Lee (2010), OECD Economic Outlook (No 90).
Note: annual data obtained through linear interpolation of 5year intervals data points.
A very high correlation is found (0.99) between male and female years of education (Table 2); although
still high, the correlation is far weaker (0.63) between the total average years of education and the female
tomale ratio.
Table 2. Covariance matrix between education variables
Male average
years of education
Female average
years of education
Total average
years of education
Gender ratio in
average years of
education
Male average years of education
1.00



Female average years of education
0.99
1.00


Total average years of education
0.99
0.96
1.00

Gender ratio in average years of
education
0.73
0.82
0.63
1.00
The issue of multicollinearity between our independent variables is investigated further with the
computation of the Besley, Kuh and Welsh (1980)
11
statistics reported in Table 3, which clearly suggests
that multicollinearity might be serious (condition index greater than 30) because the intercept and the
variable measuring the formation of physical capital are collinear. These results confirm that collinearity
11
The implementation of this “test” in a context of modeeling economic growth was suggested by L’Angevin and Laib (2005)
12
between the average years of education and the gender ratio appears to be less of a concern despite
relatively high correlation coefficient.
Table 3. Assessment of multicollinearity between the independent variables
Condition index
Intercept
Years of
education
Physical capital
Gender gap in
education
1
1.00
0.00
0.00
0.00
0.01
2
2.18
0.00
0.00
0.00
0.37
3
24.51
0.08
0.94
0.04
0.57
4
60.08
0.92
0.05
0.96
0.06
5. Model specifications and results
Before presenting our results, we’ll briefly discuss our choice of the estimation procedure. The PMG
estimaton is preferred for a number of reasons. The PMG estimator, which imposes longrun homogeneity
to all slope coefficients but the time trend, yields lower standard errors and therefore more precise
parameters compared to MG (see Tables A2 and A3 in the appendix). The measured speed of convergence
is also significantly reduced, without changing the sign of the estimated longrun coefficients. Furthermore,
a Hausman test comparing the PMG restrictions on longrun convergence against the parameters obtained
by the MG estimation does not reject the former at the conventional statistical levels.
12
Loglikelihoods are
also lowest with PMG estimations, which suggest that data are better predicted by this procedure.
Additionally, the coefficient for education obtained with the PMG is much lower than those estimated by
the MG, but significantly positive even when time dummies are included. The coefficient for education
also remains highly positive when it is estimated with a GMM procedure to tackle the bias that reverse
causality or omitted variables can potentially induce. Last but not least, only the PMG estimates yield
stationary residuals.
Two additional features emerging from Table A2 suggest that the results are consistent with an endogenous
growth model (Arnold et al., 2011) and only partially corroborate the assumptions of the Solow human
capitalaugmented framework. First, the assumption of decreasing return to scale for factors of production
cannot be rejected by the specification without time trends, while the estimation including time trends
suggests that return to scale are constant (i.e.
is close to 1 and the test does not reject the null assumption
of constant returns to broad physical and human capital). Secondly, the estimated speed of convergence
parameter (respectively 0.21 to 0.33) is quite higher than the predicted speed predicted by the Solow
model.
13
These tests give enough confidence on the quality of the preferred PMG estimation and we now turn to
model specifications that include the ratio measuring gender inequalities in completed years of education
12
Results from a dynamic fixed effect model (DFE) are reported in Annex Table A2. The estimators obtained yield a much lower
speed of convergence, which is consistent with the expected downward bias in dynamic heterogeneous panels. Moreover,
restricting the short term dynamic to be homogenous (as with DFE) affects the sign and significance of the longrun coefficients.
13
These values are again close to those estimated by Bassanini and Scarpetta (2002) and Arnold et al. (2011) which for
range
between 0.25 and 0.36.
13
(Table 1). Columns (1) and (2) present the results of the estimation on the overall sample of 30 countries
for the years between 1960 and 2008; the baseline specification in column (1) only includes the overall
level of education while the specification in column (2) includes the ratio Rf/m of the average years of
education of women relative to men. As a robustness check, we estimate the specification in column (2) for
restricted periods of time: column (3) presents the results estimated on the 19601990 period, while column
(4) presents the results for the 19841990 period. The last column presents the results of an additional
robustness check carried out to verify whether results are driven by countries that had very high increases
in the overall average years of education during the 19602008 period.
The results of the estimations for restricted periods in columns (3) and (4) suggest that important
differences over time exist. During the first subperiod (19601990) the overall years of education have a
larger effect than the gender ratio; in the mid1980s and onwards, instead, the gender ratio has a much
higher coefficient than the overall years of education. Column (5) presents the results of the estimation on
a sample that excludes countries where the increase in the overall years of education was more than ½
standard deviation above the average between 1960 and 2008. In the restricted sample both the overall
years of education and the gender gap in education have a larger effect on growth (compared to the
specification in column 2) suggesting that gender equality matters more in countries with low to moderate
average increase in education.
All estimation procedures presented in Table 4 identify a convergence parameter with a negative sign,
which is consistent with the assumption that variables converge to a longrun equilibrium. The estimated
partial elasticity of output with respect to physical capital (
) is relatively limited, ranging between 0.23
and 0.32, and consistent with previous findings.
14
The coefficients of the average education of the
population and of the gender gap in education are positive and highly significant. The PMG estimation on
the overall sample (column 2) suggests a growthelasticity to the years of education of 0.94; since the
average number of years in education has increased on average by 1.2% per annum (from 6.1 years in 1960
to 11.1 in 2008), human capital accumulation is estimated to have induced an increase in growth of 1.1%
(=0.94*1.2) per annum. As GDP per capital actually grew by 2.1% per annum on average, the model
implies that the increase in years of education accounts for about 50% of economic growth, of which just
over half was due to the increase of educational attainment among women. The results also suggest that a
balanced gender ratio in education (Rf/m=1) increases output per capita by around 0.8% in comparison to a
scenario where women have no access to education. The estimated effect of the average years of education
is slightly smaller in this case than those obtained when the gender ratio is not included, but it is balanced
by a significant and quite large effect of greater equality in education between women and men.
14
As a benchmark, Mankiw et al. (1992) estimate
=0.48 and
=0.23 using data from a group of 98 countries over the period
1960 to 1985. For OECD countries, Bassanini and Scarpetta (2002) and Arnold et al. (2011) find values between 0.13 and 0.22
for
, and between 0.52 and 0.82 for
.
14
Table 4. Growth equation results
(1)
(2)
(3)
(4)
(5)
Dependent variable:
∆log Y
Pooled Mean
Group (PMG)
Pooled Mean
Group (PMG)
19601990
19842008
Without countries
with high increase
in years of
education(3)
Convergence coefficient
logYt1
0.28***
(0.04)
0.33***
(0.06)
0.39***
(0.07)
0.52***
(0.07)
0.30***
(0.06)
Longrun coefficients
log K
0.28***
(0.01)
0.30***
(0.01)
0.32***
(0.02)
0.23***
(0.01)
0.28***
(0.02)
log H
1.07***
(0.06)
0.94***
(0.07)
0.98***
(0.07)
0.40***
(0.03)
1.03***
(0.08)
log Rf/m
..
0.81***
(0.16)
0.60***
(0.18)
1.42***
(0.16)
1.08**
(0.19)
∆log N
1.57***
(0.70)
..
..
..
..
∆log Nm
..
0.82
(1.05)
1.71
(1.52)
0.91
(0.76)
0.69
(1.24)
∆log Nf
..
4.57***
(1.34)
3.81**
(1.67)
1.76**
(0.86)
4.29***
(1.60)
Diagnostics of residuals:
Test of cross section
independence abs. ρ (p
value)1
0.86 (0.00)
0.88 (0.00)
0.90 (0.00)
0.84 (0.00)
0.93 (0.00)
Stationarity  unit root test2
0.00
0.00
0.00
0.00
0.00
N. of countries
30
30
23
29
22
N. of obs.
1150
1127
620
617
446
Log likelihood
3184
3184
1882
2148
3185
Notes: Only longrun parameters are presented. Period effects are captured by 5years time dummies, assumed to have country
specific effect.
15
Standard errors in brackets. ***, **, *: significant at the 1%, 5% and 10% level, respectively.
1) Pesaran (2004) CD test, the null hypothesis assuming that all residuals are crosssection independent. Absolute
correlation and pvalue of the test are reported; a pvalue below 0.05 leads the rejection of crosssection independence.
2) Results of the Pesaran (2007) CIPS tests which assume crosssection dependence between panel units; a pvalue below
15
Eight (out of the potential ten) time dummies defined for sequence of 5 years are included in the set of the shortrun regressors.
This leaves enough degrees of freedom to run the regression without imposing linear trends, which would also create
collinearity with the linear interpolation years of education used to complete the time series. In fact, the regression
including linear time trends instead of our set of time dummies prove to bias the estimate of the coefficient of the years
of education as part if its influence is in fact absorbed by the time trend.
15
0.05 does not reject the assumption that all residuals are stationary.
3) Australia, Canada, Czech Republic, Ireland, Israel, New Zealand, Norway, Slovak Republic and the United States are
excluded from this specification.
All the reported estimations in Table 4 yield stationary residuals, so that there is no apparent problem of
spurious results. On the other hand, the magnitude of estimated effect of human capital appears to be
sensitive to the way time trends are modelled: models that include period effects yield significantly higher
speed of convergence. The coefficients associated with physical capital also decrease when time controls
are included, while those of human capital increase. In other words, the economic returns to education
appear to be higher once other unobservable changes in the production function are controlled for,
suggesting that economic growth is progressively more dependent on the prolongation of education. These
unobserved factors, however, might not be well captured by time trends, as there is evidence of a cross
section dependence between residuals in all models of Table 4: the absolute correlation value is very high
(ranging from 0.84 to 0.93), and the CDPesaran (2004) test constantly rejects the null assumption of cross
section independence.
16
6. Conclusions
This paper provides an assessment of the extent to which the dynamics of economic growth over the past
four decades prior to the ongoing recession are related to the increase in educational attainment of women.
Both men and women have experienced an important increase in the number of years spent in education
since the early 1960s. In many countries, the increase in the average numbers of years in education spent
by women and men aged 2564 has been roughly equivalent – the average number of years in education
increased from 6.5 in 1960 to 11.2 in 2010 for men, and from 5.8 to 11.2 for women.
There are several reasons to argue that greater gender equality in education boosts economic growth.
Assuming that boys and girls have a similar distribution of innate abilities and that children with more
abilities are more likely to receive better quality and/or longer education, gender inequality in education
implies that boys with lower abilities than girls are more likely to be enrolled in education. As a result, the
average level of human capital in the economy would be lower than in a context of equal opportunities in
education for boys and girls, which in turn might slow down economic growth. By the same reasoning,
gender inequality in education reduces the impact of male education on economic growth and raises the
impact of female education (Dollar and Gatti, 1999 and, Knowles et al., 2002). It may also hinder
economic growth by reducing returns on investments. Finally, greater gender balance in human capital also
leads to higher growth if male and female human capital are imperfect substitutes and if the marginal
returns to education are declining (Knowles et al., 2002).
Our analysis is based on new OECD data series on GDP per capita and the updated version of the Barro
and Lee (2010) data set on educational attainment of men and women. Our results support the assumption
16
Also, this assumption is rejected when crosssection averages of the dependent and independent variables are added as
regressors in the CCEMG specification (table A1). This specification, which is expected to sort out the issue raised by
unobserved but correlated factors, fails to completely remove crosssection dependence although correlation is reduced
– which is not surprising because the years of education have followed very similar trends in most countries and thus
adding crosssection averages to the set of regressors generates multicollinearity of variables more than it helps wiping
out the incidence of unobserved correlated factors.
16
that the increase in the number of years spent in formal education by the working age population in OECD
countries has shifted up the steadystate of economic growth. Convergence to this steady state takes place
at different pace, however, depending on population growth and investments in physical capital,
technological and institutional change. The empirical specification of the model we retained, based on a
Pooled Mean Group Estimation, suggests that one additional year of schooling in the population is
estimated to raise output per capita by around 10% per annum, which is close to the estimate obtained by
Arnold et al. (2011) or Eberhardt and Teale (2010) but lower than the upper bound suggested by Canton
(2007). Overall, our estimation implies that the increase in years of education accounts for slightly more
than 60% of output per capita growth, of which 34 percentage points result from the increase in years of
education of women. These estimates, however, do not provide a fully satisfactory control of the
unobserved but correlated country characteristics that potentially alter the influence of human capital on
growth.
Although we found evidence that a more equal access to prolonged education raises growth rates, we are
not able to explicitly identify the causes. On possibility, as suggested by Knowles et al. (2002), is that male
and female human capital are both characterized by decreasing returns but are complementary, so that for
certain values of the average of human capital stock, years of education of women are more rewarding than
men. Another possibility is that women now perform better and are less likely to lack basic skill than boys
and are thus more valuable in the labour market, despite persistent discrimination and professional
segregation (OECD, 2012b; Hanushek and Woessmann, 2010). Other positive externalities of female
education on the quality of life and productivity might also be at play, above and beyond their greater
integration in the labour market. There is therefore much scope for further research on the subject.
17
REFERENCES
Arnold J., A. Bassanini and S. Scarpetta (2011), “Solow or Lucas? Testing the speed of convergence on a
panel of OECD countries”, Research in Economics, Vol. 65, pp. 11023.
Barro R.J. (1999), “Human capital and growth in crosscountry regressions”, Swedish Economic Policy
Review, Vol. 6(2), pp. 23777.
Barro, R.J. and J.W. Lee (1994), “Sources of economic growth”, CarnegieRochester Conference Series
on Public Policy, Vol. 40, pp. 146.
Barro, R.J. and J.W. Lee (2010), “A new data set of educational attainment in the world, 19502010”, WP
15902, NBER Working Paper Series.
Barro, R.J. and X. SalaiMartin (1995) Economic Growth, McGrawHill, New York.
Bazzi S., Clemens M. (2013), “Blunt instruments: avoiding common pitfalls in identifying the causes of
economic growth”, American Economic Journal: Macroeconomics, 5(2): 152186.
Bond S., Leblebicioglu A., and F. (2010), "Capital accumulation and growth: a new look at the empirical
evidence," Journal of Applied Econometrics, John Wiley & Sons, Ltd., vol. 25(7), pp. 10731099.
Caselli, F., G. Esquivel, and F. Lefort (1996), “Reopening the convergence debate: a new look at cross
country growth empirics”, Journal of Economic Growth, Vol. 1, pp. 36389.
Bassanini A. and S. Scarpetta (2002), “Does human capital matter for growth in OECD countries? A
pooled meangroup approach”, Economic Letters, Vol. 74, pp. 399405.
Bond S., Hoeffler A., Temple J. (2001), “GMM estimation of empirical growth models”, Discussion Paper
No. 3048, Center for for Economic Policy Research, London.
Brummet Q (2008), “The effect of gender inequality on growth: a crosscountry empirical study”, The
Park Place Economist: Vol. 16, pp. 1322.
Bun M., Winmeijer F. (2010), “The weak instrument problem of the System GMM estimator in Dynamic
Panel Data Models”, Econometrics Journal 13(1):95126.
Canton E. (2007), “Social returns to education: Macroevidence”, De Economist, Vol. 155(4), pp. 44966.
Cohen D., Soto M. (2007), “Growth and human capital: good data, good results”, Journal of Economic
Growth, Vol. 12, pp. 5176.
Crespo Cuaresma, J. (2005), “Convergence of educational attainment levels in the OECD: More data, more
problems?”, Economics of Education Review, Vol. 25, pp. 17378.
Delgado M., D.J. Henderson and C. Parmeter (2013), “Does Education Matter for Economic Growth?”
forthcoming, Oxford Bulletin of Economics and Statistics.
18
Di Vaio G. And K. Enflo (2009), "Did Globalization Lead to Segmentation? Identifying CrossCountry
Growth Regimes in the LongRun" Working Papers CELEG 0902, Dipartimento di Economia e
Finanza, LUISS Guido Carli.
Dollar, D., R. Gatti (1999), “Gender inequality, income, and growth: are good times good for women? ”,
Policy Research Report on Gender and Development Working Paper Series, No. 1, The World Bank,
Washington, DC.
Durlauf, S. N., P.A. Johnson and J.R.W. Temple (2005), “Growth econometrics” in P. Aghion and S. N.
Durlauf (eds.), Handbook of Economic Growth, Vol. 1A, NorthHolland: Amsterdam, 2005, pp.
555677.
Eberhardt M. and F. Teal (2011), “Econometrics for grumblers: a new look at the literature on cross
country growth empirics”, Journal of Economic Surveys, Vol. 25(1), pp. 10955.
Eberhardt M. and F. Teal (2010), “Aggregation versus heterogeneity in crosscountry growth empirics”,
CSAE Working Paper Series 201032, University of Oxford.
Eberhardt M., Ch. Helmers and H. Strauss (2012), “Do Spillovers Matter When Estimating Private Returns
to R&D?”, forthcoming Review of Economics and Statistics,
EsteveVolart B. (2000), “Sex discrimination and growth”, IMF Working Paper, 00/84.
Forbes, K.J. (2000), “A reassessment of the relationship between inequality and growth”, American
Economic Review, Vol. 90, pp. 86987.
Fuente A. and R. Domenech (2006), “Human capital in growth regressions: how much difference does data
quality make?”, Journal of the European Economic Association, Vol. 4(1), pp. 136.
Goldin C. (1995), “The Ushaped female labor force function in economic development and economic
history”, in T.P. Schultz (ed.), Investments in Women’s Human Capital, University of Chicago Press.
Hanushek E. and L. Woessman (2010), “How do educational outcomes matter in OECD countries?” IZA
Discussion paper n°5401.
Hill, M.A. and E.M. King (1995), 'Women's education and economic wellbeing', Feminist Economics, Vol.
1, pp. 2146.
Islam, N. (1995), “Growth Empirics: A Panel Data Approach,” Quarterly Journal of Economics, Vol. 110,
pp. 112770.
Kao, Chihwa (1999), “Spurious regression and residualbased tests for cointegration in panel data”,
Journal of Econometrics, Vol. 65 (1), pp. 915.
Klasen S. (2002), “Low Schooling for Girls, Slower Growth for All? CrossCountry Evidence on the
Effect of Gender Inequality in Education on Economic Development," World Bank Economic
Review, Vol. 16(3), pp. 34573.
Knowles S., P. Lorgelly and P.D. Owen (2002), “Are Educational Gender Gaps a Brake on Economic
Development? Some CrossCountry Empirical Evidence”, Oxford Economic Papers, Vol. 54(1), pp.
11849.
19
Kourtellos A. (2011), "Modeling parameter heterogeneity in crosscountry regression models," University
of Cyprus Working Papers in Economics, University of Cyprus Department of Economics.
Krueger A. and M. Lindahl (2001), “Education and Growth”, Journal of Economic Literature, XXXIX, pp.
10011136.
L’Angevin C., and N. Laib (2005), “Éducation et croissance en France et dans un panel de 21 pays de
l’OCDE”, Document de travail, Direction des Etudes et Synthèses Economiques G200508, INSEE,
Paris.
Lee, K., M.H. Pesaran and R. Smith (1997), “Growth and convergence in a multicountry empirical
stochastic Solow model”, Journal of Applied Econometrics, Vol 12, pp. 35792.
Lincove J.A. (2008), “Growth, girls’ education, and female labor: a longitudinal analysis”, The Journal of
Developing Areas, Vol 41(2), pp. 4568.
Lorgelly P. (2000), “Are there GenderSeparate Human Capital Effects on Growth? A Review of the
Recent Empirical Literature”, Credit Research Paper, No 00/13, Centre for Research in Economic
Development and International Trade, University of Nottingham.
Lorgelly P. and D. Owen (1999), “The effect of female and male schooling on economic growth in the
BarroLee model”, Empirical Economics, Vol. 24(3), pp. 53757.
Lutz W., A. Goujon and W. Sanderson (2007), “Reconstruction of population by age, sex and level of
educational attainment of 120 countries for 19702000”, Vienna Yearbook of Population Research,
Vol. 2007, pp. 193235.
Maasoumi, E., J. Racine and T. Stengos (2007), “Growth and Convergence: A profile of distribution
dynamics and mobility,” Journal of Econometrics, Vol. 136, pp. 483508.
Mammen K. and Ch. Paxson (2000), “Women’s Work and Economic Development”, Journal of Economic
Perspectives, Vol. 14(4), pp. 14164.
Morrisson C. and F. Murtin (2009), “The century of education”, CEP Discussion Paper n°934.
Morrisson C. and F. Murtin (2013), “The Kuznets curve of human capital inequality: 1870–2010”, The
Journal of Economic Inequality, Vol. 3, pp. 283301
OECD (2012a), OECD Economic Outlook, Vol. 2012/1, OECD Publishing.
OECD (2012b), Closing the Gender Gap: Act Now!, OECD Publishing, Paris.Pedroni P. (2007), “Social
Capital, Barriers to Production and Capital Shares: Implications for the Importance of Parameter
Heterogeneity from a Nonstationary Panel Approach”, Journal of Applied Econometrics, Vol. 22, pp.
42951.
Pesaran, H. M (2006), “Estimation and Inference in Large Heterogeneous Panels with a Multifactor Error
Structure”, Econometrica Vol. 74 (4), pp. 9671012.
Pesaran M.H (1997), “The Role of Economic Theory in Modelling the Long Run”, Economic Journal, Vol.
107(440), pp. 178191.
20
Pesaran M.H., Shin Y. (1999), “An autoregressive distributed lag modelling approach to cointegration
analysis”, Econometrics and Economic Theory in the 20th century: The Ragnar Frish Centennial
Symposium. ed. / Steinar Strom. Cambridge : Cambridge University Press, 1999. p. 371413.
Pesaran, M.H. and R. Smith (1995), “Estimating longrun relationships from dynamic heterogeneous
panels”, Journal of Econometrics, Vol. 68, pp. 79113.
Schultz, T.P. (2002), “Why governments should invest more to educate girls”, World Development, Vol.
30(2), pp. 20725.
Schultz, T.P. (1988), “Education investments and returns”, in H. Chenery and T.N. Srinivasan (eds),
Handbook of Development Economics, Vol. I, NorthHolland, Amsterdam.
Soto M. (2002), “Rediscovering Education in Growth Regressions”, Technical Paper 202, OECD
Development Centre, Paris.
Stokey N. (1994), “Sources of Economic Growth: Comments on Barro and Lee”, CarnegieRochester
Conference Series on Public Policy, Vol. 40, pp. 4757.
Tam H. (2011), “Ushaped Female Labor Participation with Economic Development: Some Panel Data
Evidence”, Economic Letters, Vol. 110(2), pp. 14042.
Thévenon O., Ali N., Adema W., Salvi del Pero A., (2012). "Effects of Reducing Gender Gaps in
Education and Labour Force Participation on Economic Growth in the OECD," OECD Social,
Employment and Migration Working Papers 138, OECD Publishing.
Topel, R. (1999), “Labor markets and economic growth”, in O. Ashenfelter and D. Card (eds), Handbook
of Labor Economics, Vol. 3C, NorthHolland, Amsterdam.
21
ANNEX
Figure A 1.1. Evolution of years in education – men and women aged 2564
22
Figure A 1.1. Evolution of years in education – men and women aged 2564 (cont.)
23
Figure A 1.1. Evolution of years in education – men and women aged 2564 (cont.)
24
Figure A 1.1. Evolution of years in education – men and women aged 2564 (cont.)
25
Table A 1. Comparison of specifications based on the Schwartz Bayesian Information Criterion (BIC)
Specification
Number of lags for the shortrun dynamics
Human cap
Physical cap
Population growth
BIC
a (baseline)
1
1
2
6656.247
B
1
2
2
6701.624
C
2
2
2
6753.304
D
1
1
1
6737.156
E
1
2
1
6699.823
F
2
2
1
6553.937
Table A 2 Growth equation – baseline estimation
Mean Group
(MG)
Pooled Mean Group (PMG)
Dependent variable:
∆log Y
(1)
No time
effects
(2)
with 5year
dummies
(3)
CCEMG
(4)
No time
effects
(5)
with period
effects
Convergence coefficient
logYt1
0.18***
(0.05)
0.38***
(0.06)
0.36
(0.22)
0.09***
(0.02)
0.28***
(0.04)
Longrun coefficients
log K
0.45***
(0.12)
0.41***
(0.09)
0.26***
(0.04)
0.47***
(0.02)
0.28***
(0.01)
log H
1.88**
(0.81)
2.28
(1.55)
0.74***
(0.26)
0.32***
(0.04)
1.07***
(0.06)
∆log N
4.82
(21.83)
0.13
(15.9)
1.52
(3.89)
5.00***
(1.15)
1.57***
(0.70)
Hausman test of longrun slope homogeneity (p
value)
0.28
0.45
Estimated returns to scale
1.60

0.54
1.06
Test for constant returns to scale (pvalue)
0.30

0.000
0.24
Diagnostics of residuals:
Test of cross section independence abs. ρ (p
value)1
0.89 (0.00)
0.89 (0.00)
0.83 (0.00)
0.77 (0.00)
0.86 (0.00)
Stationarity  unit root test2
0.06
0.19
0.00
0.00
0.00
Implied α3
0.31
0.29
0.21
0.32
0.22
Implied β3
1.30
1.61
0.59
0.22
0.83
Implied λ3
0.20
0.48
0.45
0.09
0.33
Speed of convergence test (pvalue) 4
(1)
0.002
0.000
(2)
0.14
0.000
N. of countries
30
30
30
30
30
N. of obs.
1150
1150
1150
1150
1150
N. of instruments





Log likelihood
3266
3512
3114
3184
Notes: Only longrun parameters are presented. Period effects are assumed to be linear or not and captured by 5years time
dummies. Standard errors in brackets. ***, **, *: significant at the 1%, 5% and 10% level, respectively.
1) Pesaran (2004) CD test, the null hypothesis assuming that all residuals are crosssection independent. Absolute
correlation and pvalue of the test are reported; a pvalue below 0.05 leads the rejection of crosssection independence.
2) Results of the Pesaran (2007) CIPS tests which assume crosssection dependence between panel units; a pvalue
below 0.05 does not reject the assumption that all residuals are stationary.
3) Implied α is equal to
; β=
; λ speed of convergence is given by : ln(1
)/s, with s taken at
1.
26
4) Test for estimated speed of convergence to be compatible with the value predicted by the Solow augmented model.
Predicted values are computed on the basis of plausible values for population growth rate, the depreciation rate and the
rate of technological progress. (1) assumes a standard value of 2% for depreciation rate; for technological progress, we
consider the average estimated time trend (0.3%) or the average annual shift implied by the fiveyear dummies (0.1%),
depending on the specification. For population growth, we take the average value of our sample (0.9%). (2) assumes a
much higher value of capital depreciation (10%) and allow technological progress to growth at higher 3%.
Table A .3. Growth equation estimated with pooled mean estimator
Baseline – balanced panel sample
Specification c
Convergence coefficient
logYt1
0.31***
(0.05)
0.32***
(0.04)
0.26***
(0.05)
0.29***
(0.04)
Longrun coefficients
log K
0.27***
(0.01)
0.29***
(0.01)
0.31***
(0.01)
0.36***
(0.02)
log H
1.09***
(0.06)
1.02***
(0.07)
0.87***
(0.7)
0.85***
(0.07)
log Rf/m

0.75***
(0.19)

0.83***
(0.17)
∆log N
1.78***
(0.67)
2.29***
(0.65)
2.26***
(0.74)
2.81***
(0.63)
Estimated returns to scale for reproductible
factors
1.07

Test for constant returns to scale (pvalue)
0.132

Diagnostics of residuals:
Test of cross section independence ρ (p
value)3
0.92
(0.00)
0.88
(0.00)
Pesaran (2007) unit root test
0.00
0.00
Publications des Documents de Travail
N° 202. – Olivier THÉVENON et Angelica SALVI DEL PERO, Gender Equality ( f)or Economic Growth? Effects of
Reducing the Gender Gap in Education on Economic Growth in OECD Countries, février, 26 p.
N° 201. – Olivier THÉVENON, Institutional drivers of female labour Force particpation in OECD countries,
janvier, 40 p.
N° 200. – Olivier THÉVENON, Politiques familiales, fécondité et emploi des femmes : apports et limites des
comparaisons au niveau national/Family policy, fertility and women’s employment: value and limitations of
national comparisons, janvier, 44 p
N° 199. – Olivier THÉVENON et Anne SOLAZ, Parental Leave and Labour Market Outcomes:Lessons from
40 Years of Policies in OECD countries, janvier, 38 p
N° 198. – Marie DIGOIX, Coming out et ordre normatif en Islande, octobre 2013, 134 p.
N° 197. – Actes de la Journée Doctorale de l’Ined du 30 juin 2011, septembre 2013, 220 p.
N° 196. – Patrick SIMON, Vincent TIBERJ, Sécularisation ou regain religieux : la religion des immigrés et de
leurs descendants, juillet 2013, 38 p.
N° 195. – Gilles PISON, Laetitia DOUILLOT, Géraldine DUTHÉ, Malick KANTE, Cheikh SOKHNA,
JeanFrançois TRAPE, Successes and Failures in the Fight against Child Mortality in SubSaharan Africa:
Lessons from Senegal, juin 2013, 34 p.
N° 194. – François HÉR AN, L’usage des langues dans la recherche publique en France, d’après l’enquête Elvire,
juin 2013, 90 p.
N° 193. – André WIELKi, Pratique de Sas Windows ...9.3, Volume 2, mars 2013, 255 p.
N° 192. – André WIELKI, Pratique de Sas Windows v 9.2…, Volume 1, mars 2013, 211 p.
N° 191. – Carole BONNET, Alice KEOGH, Benoît RAPOPORT, Quels facteurs pour expliquer les écarts de patri
moine entre hommes et femmes en France ?, mars 2013, 44 p.
– Carole BONNET, Alice KEOGH, Benoît RAPOPORT, How can we explain the gender wealth gap in
France?, mars 2013, 46 p.
N° 190. – Christelle HAMEL et Muriel MOISY, Immigrés et descendants d’immigrés face à la santé, Série Trajectoires
et Origines (TeO) : enquête sur la diversité des populations en france, janvier 2013, 50 p.
N° 189. – Christelle HAMEL, Bertrand LHOMMEAU, Ariane PAI LHÉ, Emmanuelle SANTELLI, Rencontrer son
conjoint dans un espace multiculturel et international, Série Trajectoires et Origines (TeO) : enquête sur la diver
sité des populations en france, janvier 2013, 32 p.
N° 188. – Cora Leonie MEZGER KVEDER, Temporary Migration:A Review of the literature, janvier 2013, 44 p.
N° 187. – Arnaud RÉGNIERLOILIER, Présentation, questionnaire et documentation de la troisième vague de l’Étude
des relations familiales et intergénérationnelles (ErﬁGGS 2011), décembre 2012, 328 p.
N° 186. – Zahia OUADAHBEDIDI et Jacques VA L LIN, Fécondité et politique de limitation des naissances en Algérie :
une histoire paradoxale, octobre 2012, 20 p.
N° 185. – Zahia OUADAHBEDIDi, Fécondité et nuptialité différentielles en Algérie : l’apport du recensement de
1998, octobre 2012, 42 p.
N° 184. – Maud LESNÉ et Patrick SIMON, La mesure des discriminations dans l’enquête « Trajectoires et Origines »,
septembre 2012, 32 p.
N° 183. – Yael BRINBAUM, Mirna SAFI, Patrick SIMON, Les discriminations en France : entre perception et expé
rience, septembre 2012, 34 p.
N° 182. – Dominique MEURS, Bertrand LHOMMEAU et Mahrez OKBA, Emplois, salaires et mobilité intergénéra
tionnelle, 2012, 48 p.
N° 181. – Christelle HAMEL et Ariane PAILHÉ , Former une famille en contexte migratoire, septembre 2012, 35 p.
N° 180. – MarieThérèse LETABLIER et Anne SALLES, Labour market uncertainties for the young workforce in
France and Germany: implications for family formation and fertility, juillet 2012, 79 p.
N° 179. – Aline F. DÉSESQUELLES, SelfRated Health of French Prison Inmates: Does Time Spent Behind Bars
Matter?, mai 2012, 14 p.
N° 178. – Aline F. DÉSESQUELLES, SelfRated Health of French Prison Inmates: Measurement and Comparison
with Other Health Indicators, mai 2012, 20 p.
N° 177. – Béatrice VALDES, Khalid ELJIM, Christophe BERGOUIGNAN, Patrick FESTY, JeanPaul SARDON (coord.),
Évaluer et valoriser la base européenne de microdonnées de recensement. Les résultats du projet ANR CENSUS,
janvier 2012, 355 p.
N° 176. – Patrick SIMON et Vincent TIBERJ, Les registres de l’identité. Les immigrés et leurs descendants face à
l’identité nationale, janvier 2012, 30 p.
N° 175. – Vincent TIBERJ et Patrick SIMON, La fabrique du citoyen : origines et rapport au politique en France,
janvier 2012, 32 p.
N° 174. – Angela LUCI et Olivier THÉVENON, The impact of family policy packages on fertility trends in developed
countries, janvier 2012, 40 p.
N° 173. – Arnaud RÉGNIERLOILIER, Leila SABONI, Béatrice VALDES, Presentation and Modiﬁcations to the
Generations and Gender Survey Questionnaire in France (Wave 2). L’Étude des relations familiales et intergéné
rationnelles (Érﬁ.), novembre 2011, 146 p.
N° 172. – Cris BEAUCHEMIN, Hugues LAGRANGE, Mirna SAFI, Transnationalism and immigrant assimilation in
France: between here and there?, Imiscoe Annual Conference, Workshop 6: Integration and transnationalism: how
are the two connected? Warsaw, Poland – 79 September 2011, 2011, 26 p.
N° 171. – JeanLouis PAN K É SHON, Claire SCODELLARO, Discrimination au logement et ségrégation ehno raciale
en France, 2011, 30 p.
N° 170. – Audrey SIEURIN, Emmanuelle Cambois, JeanMarie Robine, Les espérances de vie sans incapacité en
France : Une tendance récente moins favorable que dans le passé, 2011, 30 p.
N° 169. – Ariane PAIHLÉ et Anne SOLAZ, Does job insecurity cause missing births in a high fertility European
country. Evidence for France, 2011, 32 p.
N° 168. – Équipe TeO, coordonné par Patrick SIMON, Cris BEAUCHEMIN et Christelle HAMEL, Trajectoire et
Origines. Enquête sur la diversité des populations en France. Premiers résultats, 2010, 152 p.
N° 167. – Angela LUCI, Olivier THEVENON, Does economic development drive the fertility rebound in OECD
countries?, 2010, 45 p.
N° 166. – Cris BEAUCHEMIN , Lama KABBANJI, Bruno SCHOUMAKER, Sept communications présentées lors de
la table ronde sur les migrations entre l’Afrique et l’Europe, Dakar, Sénégal, 21 novembre 2009, 244 p.
N° 165. – Arnaud RÉGNIERLOILIER, Présentation, questionnaire et documentation de la seconde vague de l’étude
des relations familiales et intergénérationnelles (ErﬁGGS2), 211 p.
N° 164. – Carole BONNET, Anne SOLAZ, Elisabeth ALGAVA, La séparation conjugale affectetelle l’activité profes
sionnelle ? Une estimation basée sur les méthodes d’appariement, 2009, 36 p.
N° 163. – Olivia EKERT JAFFÉ, Le coût du temps consacré aux enfants : contraintes de temps et activité féminine/
The Real Time Cost of Children in France is Equally Shared by Mothers and Fathers, 2009, 48 p.
N° 162. – Laurent GOBILLON et FrançoisCharles WOLFF, Housing and location choices of retiring households :
Evidence from France, 2009, 28 p.
N° 161. – Matthieu SOLIGNAC, Les politiques de conciliation vie professionnelle /vie familiale menées par les em
ployeurs : élaboration d’une typologie des établissements de l’Enquête FamillesEmployeurs, 2009, 143 p.
N° 160. – Géraldine DUTHÉ, Raphaël LAURENT, Gilles PISON, Vivre et mourir après 60 ans en milieu rural afri
cain. Isolement, recours aux soins et mortalité des personnes âgées à Mlomp, 2009, 26 p.
N° 159. – Nathalie DONZEAU et JeanLouis PAN KÉ SON, La mobilité résidentielle depuis la ﬁn des Trente
Glorieuses, 2009, 34 p.
N° 158. – Olivier THÉVENON, The costs of raising children and the effectiveness of policies to support parenthood
in European countries: a Literature Review 2009, 612 p.
N° 157. – JeanLouis PAN KÉ SON, L’émergence du sentiment d’insécurité en quartiers défavorisés. Dépassement
du seuil de tolérance… aux étrangers ou à la misère ? 2009, 20 p.
N° 156. – Maryse Marpsat, The Ined Research on Homelessness, 19932008, 2008, 218 p.
N° 155. – Éva BEAUJOUAN, Anne SOLAZ, Childbearing after separation: Do second unions make up for earlier
missing births? Evidence from France, 2008, 24 p.
N° 154. – Carole BONNET, Laurent GOBILLON, Anne LAFERRÈRE, The effect of widowhood on housing and loca
tion choices, 2008, 40 p.
N° 153. – Louise MARIE DIOPMAES, La population ancienne de l’Afrique subsaharienne. Les éléments d’évalua
tion, 2008, 20 p.
N° 152. – Traduction en Russe du N° 121.
N° 151. – P. FESTY, J. ACCARDO, D. DEMAILLY, L. PROKOFIEVA, I. KORTCH AGINA, A. SZUKIELOJCBIENKUNS
KA, L. NIVOROZHKINA, L. OVTCHAROVA, M. SEBTI, A. PATERNO, S. STROZZA, I. ELISEEVA, A. SHEVYAKOV,
Mesures, formes et facteurs de la pauvreté. Approches comparative, 2008, 196 p.
N° 150. – Géraldine DUTHÉ, Serge H. D. FAYE, Emmanuelle GU YAVARCH, Pascal ARDUIN, Malick A. KANTE,
Aldiouma DIALLO, Raphaël LAURENT, Adama MARRA, Gilles PISON, La détermination des causes de décès par
autopsie verbale : étude de la mortalité palustre en zone rurale sénégalaise, 2008, 42 p.
N° 149. – Maryse MARPSAT, Services for the Homeless in France. Description, ofﬁcial statistics, client recording
of information. A report for the European Commision, 2007, 84 p.
N° 148. – Olivier THÉVENON, L’activité féminine après l’arrivée d’enfants : disparités et évolutions en Europe à
partir des enquêtes sur les Forces de travail, 19922005, 2007, 56 p.
N° 147. – Magali BARBIERI, Population en transition. Dix communications présentées au XXVe Congrès général de
la population, Tours, France, 1823 juillet 2005, 2007, 201 p.
N° 146. – François CHAPIREAU, La mortalité des malades mentaux hospitalisés en France pendant la deuxième
guerre mondiale, 2007, 36 p.
N° 145. – Maryse MARPSAT, Explorer les frontières. Recherches sur des catégories « en marge », Mémoire présenté
en vue de l’habilitation à diriger des recherches en sociologie, 2007, 274 p.
N° 144. – Arnaud RÉGNIERLOILIER et Pascal SEBILLE, Modiﬁcations to the Generations and Gender Surveys
questionnaire in France (wave 1), 192 p.
N° 143. – Ariane PAILHÉ et Anne SOLAZ, L’enquête Familles et employeurs. Protocole d’une double enquête et
bilan de collecte, 180 p.
N° 142. – Annie BACHELOT et Jacques de MOUZON, Données de l’enquête « Caractéristiques des couples deman
dant une fécondation in vitro en France », 2007, 44 p.
N° 141. – Olivia EKERTJAFFÉ, Shoshana GROSSBARD et Rémi MOUGIN, Economic Analysis of the Childbearing
Decision, 2007, 108 p.
N° 140. – Véronique HERTRICH and Marie LESCLINGAND, Transition to adulthood and gender: changes in rural Mali
N° 139. – Patrick SIMON et Martin CLÉMENT, Rapport de l’enquête « Mesure de la diversité ». Une enquête expéri
mentale pour caractériser l’origine, 2006, 86 p.
N° 138. – Magali BARBIERI, Alfred NIZARD et Laurent TOULEMON, Écart de température et mortalité en France,
2006, 80 p.
N° 137. – JeanLouis PAN KÉ SHON, Mobilités internes différentielles en quartiers sensibles et ségrégation, 2006, 42 p.
N° 136. – Francisco MUNOZPÉREZ, Sophie PENNEC, avec la collaboration de Geneviève Houriet Segard, Évolution
future de la population des magistrats et perspectives de carrière, 20012040, 2006, XXX + 114 p.
N° 135. – Alexandre DJIRIKIAN et Valérie LAFLAMME, sous la direction de Maryse MARPSAT, Les formes margi
nales de logement. Étude bibliographique et méthodologique de la prise en compte du logement non ordinaire,
2006, 240 p.
N° 134. – Catherine BONVALET et Éva LELIÈVRE, Publications choisies autour de l’enquête « Biographies et en
tourage », 2006, 134 p.
N° 133. – Arnaud RÉGNIERLOILIER, Présentation, questionnaire et documentation de l’« Étude des relations
familiales et intergénérationnelles » (Erﬁ). Version française de l’enquête « Generations and Gender Survey »
(GGS), 2006, 238 p.
N° 132. – Lucie BONNET et Louis BERTRAND (sous la direction de), Mobilités, habitat et identités, Actes de la jour
née d’étude « Jeunes chercheurs ». Le logement et l’habitat comme objet de recherche. Atelier 3, 2005, 92 p.
N° 131. – Isabelle FRECHON et Catherine VilleneuveGokalp, Étude sur l’adoption, 2005, 64 p.
N° 130. – Dominique MEURS, Ariane PAIHLÉ et Patrick SIMON, Mobilité intergénérationnelle et persistance des
inégalités. L’accès à l’emploi des immigrés et de leurs descendants en France, 2005, 36 p.
N° 129. – Magali MAZUY, Nicolas RAZAFINDRATSIMA, Élise de LA ROCHEBROCHARD, Déperdition dans l’en
quête « Intentions de fécondité », 2005, 36 p.
N° 128. – Laure MOGUEROU et Magali BARBIERI, Population et pauvreté en Afrique. Neuf communications pré
sentées à la IVe Conference africaine sur la population, Tunis, Tunisie, 812 décembre 2003, 2005, 184 p.
N° 127. – JeanLouis PAN KÉ SHON, Les sources de la mobilité résidentielle. Modiﬁcations intervenues sur les
grandes sources de données dans l’étude des migrations, 2005, 30 p.
N° 126. – Thierry DEBRAND et AnneGisèle PRIVAT, L’impact des réformes de 1993 et de 2003 sur les retraites.
Une analyse à l’aide du modèle de microsimulation Artémis, 2005, 28 p.
N° 125. – Kees WAALDIJK (ed), More or less together: levels of legal consequences of marriage, cohabitation and
registered partnership for differentsex and samesex partners: a comparative study of nine European countries,
2005, 192 p. (s’adresser à Marie DIGOIX)
N° 124. – Marie DIGOIX et Patrick FESTY (eds), Samesex couples, samesex partnerships, and homosexual mar
riages: A Focus on crossnational differentials, 2004, 304 p.
N° 123. – Marie DIGOIX et Patrick FESTY (sous la dir.), Séminaire « Comparaisons européennes », années 2001
2002, 2004, 220 p.
N° 122. – Emmanuelle GUYAVARCH et Gilles PISON, Les balbutiements de la contraception en Afrique au Sud du
Sahara, septembre 2004, 48 p.
N° 121. – Maryse JASPARD et Stéphanie CONDON, Genre, violences sexuelles et justice. Actes de la journée
séminaire du 20 juin 2003, 2004, 135p.
N° 120. – Laurent TOULEMON et Magali MAZUY, Comment prendre en compte l’âge à l’arrivée et la durée de séjour
en France dans la mesure de la fécondité des immigrants ? 2004, 34 p.
N° 119. – Céline CLÉMENT et Bénédicte GASTINEAU (coord.), Démographie et sociétés. Colloque international
« Jeunes Chercheurs », CerposUniversité Paris XNanterre, 1er et 2 octobre 2002, 2003, 350 p.
N° 118. – Monique BERTRAND, Véronique DUPONT et France GU ÉRIN PACE (sous la dir.), Espaces de vie. Une
revue des concepts et des applications, 2003, 188 p.
N° 117. – Stephanie CONDON et Armelle ANDRO, Questions de genre en démographie. Actes de la journée du
22 juin 2001, 2003, 128 p.
N° 116. – Maryse JASPARD et l’équipe Enveff, Le questionnaire de l’enquête Enveff. Enquête nationale sur les vio
lences envers les femmes en France, 2003, 10 + 88 p.
N° 115. – Zahia OUADAHBEDIDI et Jacques VALLIN, Disparités régionales de l’écart d’âge entre conjoints en
Algérie. Évolution depuis 1966, 2003, 32 p.
N° 114. – Magali MAZUY, Situations familiales et fécondité selon le milieu social. Résultats à partir de l’enquête
EHF de 1999, 2002, 60 p.
N° 113. – JeanPaul SARDON, Fécondité et transition en Europe centrale et orientale, 2002, 38 p.
N° 112. – Thérèse LOCOH, Deux études sur la fécondité en Afrique : 1) Structures familiales et évolutions de la
fécondité dans les pays à fécondité intermédiaire d’Afrique de l’Ouest ; 2) Baisse de la fécondité et mutations fa
miliales en Afrique subsaharienne, 2002, 24 p. et 30 p.
N° 111. – Thierry DEBRAND et AnneGisèle PRIVAT, Individual real wages over business cycle: The impact of
macroeconomic variations on individual careers and implications concerning retirement pensions, 2002, 38 p.
N° 110. – Recueil préparé par Amandine LEBUGLE et Jacques VALLIN, Sur le chemin de la transition. Onze commu
nications présentées au XXIVe Congrès général de la population à Salvador de Bahia, Brésil, août 2001, 2002, 234 p.
N° 109. – Éric BRIAN, JeanMarc ROHRBASSER, Christine THÉRÉ, Jacques VÉRON (intervenants et organisa
teurs), La durée de vie : histoire et calcul. Séminaire de la valorisation de la recherche, 7 février 2000, 2002, 70 p.
N° 108. – France MESLÉ et Jacques VALLIN, Montée de l’espérance de vie et concentration des âges au décès, 2002, 20 p.
N° 107. – Alexandre AVDEEV, La mortalité infantile en Russie et en URSS : éléments pour un état des recherches,
2002, 48 p.
N° 106. – Isabelle ATTANÉ (organisatrice), La Chine en transition : questions de population, questions de société.
Séminaire de la valorisation de la recherche, 31 janvier et 1er février 2001 (s’adresser à Céline PERREL), 2002, 46 p.
N° 105. – A. AVDEEV, J. BELLENGER, A. BLUM, P. FESTY, A. PAILHÉ, C. GOUSSEFF, C. LEFÈVRE, A. MONNIER,
J.C. SEBAG, J. VALLIN (intervenants et organisateurs), La société russe depuis la perestroïka : rupture, crise ou
continuité ? Séminaire de la valorisation de la recherche, 1er mars 2001 (s’adresser à Céline PERREL), 2001, 124 p.
N° 104. – Jacques VÉRON, Sophie PENNEC, Jacques LÉGARÉ, Marie DIGOIX (éds), Le contrat social à l’épreuve
des changements démographiques ~ The Social Contract in the Face of Demographic Change, Actes des 2e Ren
contres Sauvy, 2001, 386 p.
N° 103. – Gilles PISON, Alexis GABADINHO, Catherine ENEL, Mlomp (Sénégal). Niveaux et tendances démogra
phiques ; 19852000, 2001, 182 p.
N° 102. – La famille en AOF et la condition de la femme. Rapport présenté au Gouverneur général de l’AOF. par
Denise SAV I N E AU (1938). Introduction de Pascale Barthélémy, 2001, XXII222 p.
N° 101. – JeanPaul SARDON, La fécondité dans les Balkans, 2001, 88 p.
N° 100. – JeanPaul SARDON, L’évolution récente de la fécondité en Europe du Sud, 26 p.
N° 99. – S. JUSTEAU, J.H. KALTENBACH, D. LAPEYRONNIE, S. ROCHÉ, J.C. SEBAG, X. THIERRY et M. TRIBA
LAT (intervenants et organisateurs), L’immigration et ses amalgames. Séminaire de la valorisation de la recherche,
24 mai 2000, 2001, 94 p.
N° 98. – Juliette HALIFAX, L’insertion sociale des enfants adoptés. Résultats de l’enquête « Adoption internatio
nale et insertion sociale », 2000 (Ined – Les Amis des enfants du monde), 2001, 58 p.
N° 97. – Michèle TRIBALAT, Modéliser, pour quoi faire ? 2001, 10 p.
N° 96. – O. EKERTJAFFÉ, H. LERIDON, S. PENNEC, I. THÉRY, L. TOULEMON et J.C. SEBAG (intervenants et orga
nisateurs), Évolution de la structure familiale. Séminaire de la valorisation de la recherche, 28 juin 2000, 2001, 110 p.
N° 95. – A. ANDRO, A. LEBUGLE, M. LESCLINGAND, T. LOCOH, M. MOUVAGHASOW, Z. OUADAHBEDIDI,
J. VALLIN, C. VANDERMEERSCH, J. VÉRON, Genre et développement. Huit communications présentées à la
Chaire Quetelet 2000, 2001, 158 p.
N° 94. – C. BONVALET, C. CLÉMENT, D. MAISON, L. ORTALDA et T. VICHNEVSKAIA, Réseaux de sociabilité et
d’entraide au sein de la parenté : Six contributions, 2001, 110 p.
N° 93. – Magali MAZUY et Laurent TOULEMON, Étude de l’histoire familiale. Premiers résultats de l’enquête en
ménages, 2001, 100 p.
N° 92. – Politiques sociales en France et en Russie, INED/IPSEP, 2001, 246 p.
N° 91. – Françoise MOREAU, Commerce des données sur la population et libertés individuelles, 2001, 20 p. + Annexes.
N° 90. – Youssef COURBAGE, Sergio DELLAPERGOLA, Alain DIECKHOFF, Philippe FARGUES, Emile MALET,
Elias SANBAR et JeanClaude SEBAG (intervenants et organisateurs), L’arrièreplan démographique de l’explosion
de violence en IsraëlPalestine. Séminaire de la valorisation de la recherche, 30 novembre 2000, 2000, 106 p.
N° 89. – Bénédicte GASTINEAU et Elisabete de CARVALHO (coordonné par), Démographie : nouveaux champs,
nouvelles recherches, 2000, 380 p.
N° 88. – Gil BELLIS, JeanNoël BIRABEN, MarieHélène CAZES et Marc de BRAEKELEER (modérateur et interve
nants), Génétique et populations. Séminaire de la valorisation de la recherche, 26 janvier 2000, 2000, 96 p.
N° 87. – JeanMarie FIRDION, Maryse MARPSAT et Gérard MAUGER (intervenants), Étude des sansdomicile : le
cas de Paris et de l’IledeFrance. Séminaire de la valorisation de la recherche, 19 avril 2000, 2000, 90 p.
N° 86. – François HÉRAN et JeanClaude SEBAG (responsables modérateurs), L’utilisation des sources administra
tives en démographie, sociologie et statistique sociale. Séminaire de la valorisation de la recherche, 20 septembre
2000, 2000, 170 p.
N° 85. – Michel BOZON et Thérèse LOCOH (sous la dir.), Rapports de genre et questions de population. II. Genre,
population et développement, 2000, 200 p.
N° 84. – Michel BOZON et Thérèse LOCOH (sous la dir.), Rapports de genre et questions de population. I. Genre et
population, France 2000, 2000, 260 p.
N° 83. – Stéphanie CONDON, Michel BOZON et Thérèse LOCOH, Démographie, sexe et genre : bilan et perspec
tives, 2000, 100 p.
N° 82. – Olivia EKERTJAFFE et Anne SOLAZ, Unemployment and family formation in France, 2000, 26 p.
N° 81. – JeanMarie FIRDION, L’étude des jeunes sans domicile dans les pays occidentaux : état des lieux, 1999, 28 p.
N° 80. – Age, génération et activité : vers un nouveau contrat social? Age, cohort and activity: A new “social
contract”?, Actes des 1res rencontres Sauvy (s’adresser à Marie DIGOIX), 1999, 314 p.
N° 79. – Maryse MARPSAT, Les apports réciproques des méthodes quantitatives et qualitatives : le cas particulier
des enquêtes sur les personnes sans domicile, 1999, 24 p.
N° 78. – Les populations du monde, le monde des populations. La place de l’expert en sciences sociales dans le
débat public, Actes de la Table ronde pour l’inauguration de l’Ined, 1999, 54 p.
N° 77. – Isabelle SÉGUY, Fabienne LE SAGER, Enquête Louis Henry. Notice descriptive des données informatiques,
1999, 156 p.
N° 76. – I. SÉGUY, H. COLENÇON et C. MÉRIC, Enquête Louis Henry. Notice descriptive de la partie nominative,
1999, 120 p.
N° 75. – AnneClaude LE VOYER (s’adresser à H. LERIDON), Les processus menant au désir d’enfant en France,
1999, 200 p.
N° 74. – Jacques VALLIN et France MESLÉ, Le rôle des vaccinations dans la baisse de la mortalité, 1999, 20 p.
N° 73. – Bernard ZARCA, Comment passer d’un échantillon de ménages à un échantillon de fratries ? Les enquêtes
« Réseaux familiaux » de 1976, « Proches et parents » de 1990 et le calcul d’un coefﬁcient de pondération, 1999, 20 p.
N° 72. – Catherine BONVALET, Famillelogement. Identité statistique ou enjeu politique ? 1998, 262 p.
N° 71. – Denise ARBONVILLE, Normalisation de l’habitat et accès au logement. Une étude statistique de l’évolu
tion du parc « social de fait » de 1984 à 1992, 1998, 36 p.
N° 70. – Famille, activité, vieillissement : générations et solidarités. Bibliographie préparée par le Centre de
Documentation de l’Ined, 1998, 44 p.
N° 69. – XXIIIe Congrès général de la population, Beijing, Chine, 1117 octobre 1997 :
Contribution des chercheurs de l’Ined au Congrès, 1997, 178 p.
Participation of Ined Researchers in the Conference, 1997, 180 p.
N° 68. – France MESLÉ et Jacques VALLIN, Évolution de la mortalité aux âges élevés en France depuis 1950, 1998, 42 p.
N° 67. – Isabelle SEGUY, Enquête JeanNoël Biraben « La population de la France de 1500 à 1700 ». Répertoire des
sources numériques, 1998, 36 p.
N° 66. – Alain BLUM, I. Statistique, démographie et politique. II. Deux études sur l’histoire de la statistique et de
la statistique démographique en URSS (19201939), 1998, 92 p.
N° 65. – Annie LABOURIERACAPÉ et Thérèse LOCOH, Genre et démographie : nouvelles problématiques ou effet
de mod e ? 1998, 27 p.
N° 64. – C. BONVALET, A. GOTMAN et Y. GRAFMEYER (éds), et I. BertauxViame, D. Maison et L. Ortalda, Proches
et parents : l’aménagement des territoires, 1997.
N° 63. – Corinne BENVENISTE et Benoît RIANDEY, Les exclus du logement : connaître et agir, 1997, 20 p.
N° 62. – Sylvia T. WARGON, La démographie au Canada, 19451995, 1997, 40 p.
N° 61. – Claude RENARD, Enquête Louis Henry. Bibliographie de l’enquête, 1997, 82 p.
N° 60. – H. AGHA, J.C. CHASTELAND, Y. COURBAGE, M. LADIERFOULADI, A.H. MEHRYAR, Famille et fécon
dité à Shiraz (1996), 1997, 60 p.
N° 59. – Catherine BONVALET, Dominique MAISON et Laurent ORTALDA, Analyse textuelle des entretiens
« Proches et Parents », 1997, 32 p.
N° 58. – B. BACCAÏNI, M. BARBIERI, S. CONDON et M. DIGOIX (éds),
Questions de population. Actes du Colloque Jeunes Chercheurs :
I. Mesures démographiques dans des petites populations, 1997, 50 p.
II. Nuptialité – fécondité – reproduction, 1997, 120 p.
III. Histoire des populations, 1997, 90 p.
IV. Économie et emploi, 1997, 50 p.
V. Vieillissement – retraite, 1997, 66 p.
VI. Famille, 1997, 128 p.
VII. Santé – mortalité, 1997, 136 p.
VIII. Population et espace, 1997, 120 p.
IX. Migration – intégration, 1997, 96 p.
N° 57. – Isabelle SÉGUY et Corinne MÉRIC, Enquête Louis Henry. Notice descriptive non nominative, 1997, 106 p.
N° 56. – Máire Ní BHROLCHÁIN and Laurent TOULEMON, Exploratory analysis of demographic data using gra
phical methods, 1996, 50 p.
N° 55. – Laurent TOULEMON et Catherine de GUIBERTLANTOINE, Enquêtes sur la fécondité et la famille dans
les pays de l’Europe (régions ECE des Nations unies). Résultats de l’enquête française, 1996, 84 p.
N° 54. – G. BALLAND, G. BELLIS, M. DE BRAEKELEER, F. DEPOID, M. LEFEBVRE, I. SEGUY, Généalogies et re
constitutions de familles. Analyse des besoins, 1996, 44 p.
N° 53. – Jacques VALLIN et France MESLÉ, Comment suivre l’évolution de la mortalité par cause malgré les dis
continuités de la statistique ? Le cas de la France de 1925 à 1993, 1996, 46p.
N° 52. – Catherine BONVALET et Eva LELIÈVRE, La notion d’entourage, un outil pour l’analyse de l’évolution des
réseaux individuels, 1996, 18 p.
N° 51. – Alexandre AVDEEV, Alain BLUM et Serge ZAKHAROV, La mortalité atelle vraiment augmenté brutale
ment entre 1991 et 1995 ? 1996, 80 p.
N° 50. – France MESLÉ, Vladimir SHKOLNIKOV, Véronique HERTRICH et Jacques VALLIN, Tendances récentes de
la mortalité par cause en Russie, 19651993, 1995, 70 p. Avec, en supplément, 1 volume d’Annexes de 384 p.
N° 49. – Jacques VALLIN, Espérance de vie : quelle quantité pour quelle qualité de vie ? 1995, 24 p.
N° 48. – François HÉRAN, Figures et légendes de la parenté :
I. Variations sur les ﬁgures élémentaires, 1995, 114 p.
II. La modélisation de l’écart d’âge et la relation groupe/individu, 1995, 84 p.
III. Trois études de cas sur l’écart d’âge : Touaregs, Alyawara, Warlpiri, 1995, 102 p.
IV. Le roulement des alliances, 1995, 60 p.
V. Petite géométrie fractale de la parenté, 1995, 42 p.
VI. Arbor juris. Logique des ﬁgures de parenté au Moyen Age, 1996, 62 p.
VII. De Granet à LéviStrauss, 1996, 162 p.
VIII. Les vies parallèles. Une analyse de la coalliance chez les Etoro de NouvelleGuinée, 1996, 80 p.
IX. Ambrym ou l’énigme de la symétrie oblique : histoire d’une controverse, 1996, 136 p.
N° 47. – Olivia EKERTJAFFÉ, Denise ARBONVILLE et Jérôme WITTWER, Ce que coûtent les jeunes de 18 à 25 ans,
1995, 122 p.
N° 46. – Laurent TOULEMON, Régression logistique et régression sur les risques. Deux supports de cours, 1995, 56 p.
N° 45. – Graziella CASELLI, France MESLÉ et Jacques VALLIN, Le triomphe de la médecine. Évolution de la mor
talité en Europe depuis le début de siècle, 1995, 60 p.
N° 44. – Magali BARBIERI, Alain BLUM, Elena DOLGIKH, Amon ERGASHEV, La transition de fécondité en
Ouzbékistan, 1994, 76 p.
N° 43. – Marc De BRAEKELEER et Gil BELLIS, Généalogies et reconstitutions de familles en génétique humaine,
1994, 66 p.
N° 42. – Serge ADAMETS, Alain BLUM et Serge ZAKHAROV, Disparités et variabilités des catastrophes démogra
phiques en URSS, 1994, 100 p.
N° 41. – Alexandre AVDEEV, Alain BLUM et Irina TROITSKAJA, L’avortement et la contraception en Russie et dans
l’exURSS : histoire et présent, 1993, 74 p.
N° 40. – Gilles PISON et Annabel DESGREES DU LOU, Bandafassi (Sénégal) : niveaux et tendances démogra
phiques 19711991, 1993, 40 p.
N° 39. – Michel Louis LÉVY, La dynamique des populations humaines, 1993, 20 p.
N° 38. – Alain BLUM, Systèmes démographiques soviétiques, 1992, 14 + X p.
N° 37. – Emmanuel LAGARDE, Gilles PISON, Bernard LE GUENNO, Catherine ENEL et Cheikh SECK, Les facteurs
de risque de l’infection à VIH2 dans une région rurale du Sénégal, 1992, 72 p.
N° 36. – Annabel DESGREES DU LOU et Gilles PISON, Les obstacles à la vaccination universelle des enfants des
pays en développement. Une étude de cas en zone rurale au Sénégal, 1992, 26 p.
N° 35. – France MESLÉ, Vladimir SHKOLNIKOV et Jacques VALLIN, La mortalité par causes en URSS de 1970 à
1987 : reconstruction de séries statistiques cohérentes, 1992, 36 p.
N° 34. – France MESLÉ et Jacques VALLIN, Évolution de la mor talité par cancer et par maladies cardio vasculaires
en Europe depuis 1950, 1992, 48 p.
N° 33. – Didier BLA NCHET, Vieillissement et perspectives des retraites : analyses démoéconomiques, 1991, 120 p.
N° 32. – Noël BONNEUIL, Démographie de la nuptialité au XIXe siècle, 1990, 32 p.
N° 31. – JeanPaul SARDON, L’évolution de la fécondité en France depuis un demisiècle, 1990, 102 p.
N° 30. – Benoît RIANDEY, Répertoire des enquêtes démographiques : bilan pour la France métropolitaine, 1989, 24 p.
N° 29. – Thérèse LOCOH, Changement social et situations matrimoniales : les nouvelles formes d’union à Lomé,
1989, 44 p.
N° 28. – Catherine ENEL, Gilles PISON, et Monique LEFEBVRE, Migrations et évolution de la nuptialité. L’exemple
d’un village joola du sud du Sénégal, Mlomp, 1989, 26 p.
(Sénégal) depuis 50 ans, 1re édition : 1989, 36 p. ; 2e édition revue et augmentée : 1990, 48 p.
N° 27. – Nicolas BROUARD, L’extinction des noms de famille en France : une approche, 1989, 22 p.
N° 26. – Gilles PISON, Monique LEFEBVRE, Catherine ENEL et JeanFrançois TR APE, L’inﬂuence des changements
sanitaires sur l’évolution de la mortalité : le cas de Mlomp, 1989, 36 p.
N° 25. – Alain BLUM et Philippe FARGUES, Estimation de la mortalité maternelle dans les pays à données incom
plètes. Une application à Bamako (19741985) et à d’autres pays en développement, 1989, 36 p.
N° 24. – Jacques VALLIN et Graziella CASELLI, Mortalité et vieillissement de la population, 1989, 30 p.
N° 23. – Georges TAPINOS, Didier BLANCHET et Olivia EKERTJAFFÉ, Population et demande de changements
démographiques, demande et structure de consommation, 1989, 46 p.
N° 22. – Benoît RIANDEY, Un échantillon probabiliste de A à Z : l’exemple de l’enquête Peuplement et dépeuplement
de Paris. I NED (1986), 1989, 12 p.
N° 21. – Noël BONNEUIL et Philippe FARGUES, Prévoir les « caprices » de la mortalité. Chronique des causes de
décès à Bamako de 1964 à 1985, 1989, 44 p.
N° 20. – France MESLÉ, Morbidité et causes de décès chez les personnes âgées, 1988, 18 p.
N° 19. – Henri LERIDON, Analyse des biographies matrimoniales dans l’enquête sur les situations familiales,
1988, 64 p.
N° 18. – Jacques VALLIN, La mortalité en Europe de 1720 à 1914 : tendances à long terme et changements de struc
ture par âge et par sexe, 1988, 40 p.
N° 17. – Jacques VALLIN, Évolution sociale et baisse de la mortalité : conquête ou reconquête d’un avantage fémi
nin ? 1988, 36 p.
N° 16. – Gérard CALOT et Graziella CASELLI, La mortalité en Chine d’après le recensement de 1982 :
I. – Analyse selon le sexe et l’âge au niveau national et provincial, 1988, 72 p. II. – Tables de mortalité par pro
vince, 1988, 112 p.
N° 15. – Peter AABY (s’adresser à J. VALLIN), Le surpeuplement, un facteur déterminant de la mortalité par rou
geole en Afrique, 1987, 52 p.
N° 14. – Jacques VALLIN, Théorie(s) de la baisse de la mortalité et situation africaine, 1987, 44 p.
N° 13. – Kuakuvi GBENYON et Thérèse LOCOH, Différences de mortalité selon le sexe, dans l’enfance en Afrique
au Sud du Sahara, 1987, 30 p.
N° 12. – Philippe FARGUES, Les saisons et la mortalité urbaine en Afrique. Les décès à Bamako de 1974 à 1985,
1987, 38 p.
N° 11. – Gilles PISON, Les jumeaux en Afrique au Sud du Sahara : fréquence, statut social et mortalité, 1987, 48 p.
N° 10. – Philippe FARGUES, La migration obéitelle à la conjoncture pétrolière dans le Golfe ? L’exemple du Koweït,
1987, 30 p.
N° 9. – Didier BLANCHET, Deux études sur les relations entre démographie et systèmes de retraite, 1986, 26 p.
N° 8. – Didier BLANCHET, Équilibre malthusien et liaison entre croissances économique et démographique dans
les pays en développement : un modèle, 1986, 20 p.
N° 7. – Jacques VALLIN, France MESLÉ et Alfred NIZARD, Reclassement des rubriques de la 8e révision de la
Classiﬁcation internationale des maladies selon l’étiologie et l’anatomie, 1986, 56 p.
N° 6. – Philippe FARGUES, Un apport potentiel des formations sanitaires pour mesurer la mortalité dans l’en
fance en Afrique, 1986, 34 p.
N° 5. – Jacques VALLIN et France MESLÉ, Les causes de décès en France de 1925 à 1978, 1986, 36 p.
N° 4. – Graziella CASELLI, Jacques VALLIN, J. VAUPEL et A. YASHIN, L’évolution de la structure par âge de la
mortalité en Italie et en France depuis 1900, 1986, 28 p.
N° 3. – Paul PAILLAT, Le vécu du vieillissement en 1979, 1981, 114 p.
N° 2. – Claude LÉVY, Aspects sociopolitiques et démographiques de la planiﬁcation familiale en France, en Hon
grie et en Roumanie, 1977, 248 p.
N° 1. – Georges TAPINOS, Les méthodes d’analyse en démographie économique, 1976, 288 p.
Février 2014