It may seem like the line between male and female is
slowly getting blurry. The reality, though, shows that the inequality problem is
nothing but solved. Not only is the progression of women’s economic
participation slow, and the number of politically powered women not particularly
large, the gender pay gap is also still of great significance (Elsesser, 2016).
It is important to identify the factors, and the extent to which these drive the
issue, as it is the only way to tackle it.
The main aim of this research is
analyse the difference between the private sector and public sector in terms of
how gender affects earnings. Lucifora and Meurs (2006) concluded in their
research on the pay gap in the private and public sector, that women were
better off in the public sector as the pay gap, using data of the UK, France
and Italy, showed to be less, especially in the lower percentile. The target of
this research is therefore not prove but to confirm these results.
investigate the research topic, a sample of 3000 people will be used with data
from the national longitudinal survey of youth (Cooksey & Light, 1979). The group will be
divided into two groups of public and private sector workers, to which a
quantitative method comparing male and female earnings will be correlated
against each other, using the multiple regression method, to identify the size
of the gender pay gap in both sectors. Afterwards, the correlation coefficients
and regression models of both samples can be compared against each other to
identify the difference between both sectors. In order to make this research
viable, a number of controlled variables were used, which were regarded as
other important factors contributing to earnings as well such as education.
This report will, first off, cover the theoretical
framework providing the concepts and theories regarding this research. Secondly,
having assembled the guidance through the report, the full research method will
be described in greater detail and executed, gathering results to the theory
outlined in the theoretical framework. Following up, an interpretation of the
results, and relations found between the variables in the data profiles.
Concluding with the discussion on the implications of the research on the
difference in gender pay inequality between the private and public sector.
It is important to
consider various concepts as sources of information during a research. Therefore,
it is essential to study the different underlying economic theories before doing
the regression analysis. Lots of research has been done on what the effect is
of being a woman on your earnings, the gender wage gap. The fact that there is
a significant difference in earnings between men and women is widely accepted.
Therefore, the aim of this research is not essentially to prove the gender pay
gap, but to reconfirm the claim and identify which labour sector is more
vulnerable to gender inequality. First in this paragraph, the human capital
model will be discussed, and afterwards theories that cover a connection
between the gender pay gap and the private and public sector and its causes.
The basis of any
research on wage determination, what effect different factors have on earnings,
is found in the human capital theory. Adam Smith set up the basis for which
would later become the dominant theory behind economic performance of
individuals in his book ‘The Wealth of Nations’ (Smith,
There are various views on and explanations of the human capital model.
Acemoglu’s and Autor’s (2011) view explain that the way in which society values
an individual as employee, and hereby prefers one over the other, all comes
down to their set of skills, their human capital, which in turn will make them
more productive. Their view of human capital combines Becker, Gardener and Schultz/Nelson-Phelps.
People can distinguish themselves by means of their
qualities and their abilities, which makes them of value. Less skilled
individuals will be less productive and effective, and therefore less
profitable. The difference in qualities and abilities between individuals is
ultimately due to the ‘investment’ into that individual to improve or extend
their ability, years and quality of schooling, training and non-schooling.
However, this is only the basis of any wage differential. (Acemoglu
& Autor, 2011)
After considering the
human capital model, the research by Lucifora and Meurs (2006) is considered in
greater detail, who investigated the public sector pay gap looking at both
female and male workers in the UK, France and Italy. They used surveys from
several different investigative boards in all countries and analysed the data
using quantile regression methods and OLS-estimators and controlling the
standard human capital variables. Their research outlined that the wage gap is
particularly large in the UK private sector, where the standard deviation of
the log hourly wage showed a difference of 0.051 compared to 0.019 in the
public sector. This was partially blamed on the decentralised labour laws of
the UK creating smaller union power and “sticky floors” existing in the lower
percentile and “glass ceilings” in the higher percentile of the private sector.
(Lucifora & Meurs, 2006)
and Bryan (2007) continued this investigation on the glass ceiling effect in
their research using a sample from European Community Household Panel Survey,
which was conveyed annually between 1994 to 2001 across several countries in
Europe. They started by showing raw inequality data from both sectors that in
all sample countries males earned, on average, more than their female counter
parts. For example, in the Netherlands the mean difference between log(male)
and log(female) hourly wage was shown to be 0.481 in the public sector and
0.643 in the private sector. They continued this research with a quantile
regression method and formed an OLS-estimator and confirmed the raw data as in
all countries the private sector showed a greater wage inequality than the
private sector. Lack of opportunity, Maternity leave (less experience), Lack of
childcare and discrimination (glass ceilings) where all listed as contributors
to gender pay inequality. (Arulampalam, Booth, & Bryan, 2007)
In both previous
mentioned papers above, the focus was often nog exactly on the difference
between the private and public sector wage inequality but used it as an
explanation tool. Barón and Cobb-Clark (2008) focused specifically on the
private and public sector separately in their research of occupational
segregation and the gender wage gap using the Household, Income, and Labour
Dynamics in Australia Survey. Their results showed a relatively consistent
inequality in the public sector and a rising inequality in the private sector
with women being worse off again in the 90th percentile of the
private sector. The difference between the sectors is partially blamed for the
greater presence of anti-discrimination and inclusion laws in the public
sector. (Barón & Cobb-Clark, 2008)
The Human Capital
model will always remain one of the most important theories outlined in
economic theory, which partially explains the difference in wage between men
and women. For example, when women get pregnant they will eventually have to go
on maternity leave for several months and in turn lose the work experience they
could have gained in this time and hurt their earning potential as explained by
the human capital model. Being female and earning less than male counterparts is
therefore not necessarily a causal relationship but simply shows a correlation
due to the other moderators and mediators. However, in all researches mentioned
above, the common concept was mentioned of “glass ceilings” occurring for women
in higher occupations. In the investigation, we will introduce several
controlled variables in order to investigate what factors most contribute to
the difference in pay in both sectors. However, the underlying theory explains
that the private sector will show a greater inequality than the public sector.
With the use of a
sample, we will investigate the research question. This section will first
outline the data set used and the variables that are used in the research with
greater detail followed with an explanation of the multiple regression analysis
In this research,
data from the National Longitudinal Survey of Youth (Cooksey & Light, 1979) will be used, which
recorded data from a sample of 3000 people between 1979 and 2002. Examples of
the measured values include marital status, education, and earnings and were
recorded and updated on a yearly basis. Earnings was recorded in dollars earned
per hour of labour.
The sample will first be divided into two groups:
labourers in the private sector and labourers in the public sector. Following, a
variety of controlled variables have been chosen in order to greater explain
the relation between being female and earnings. The following variables have
been chosen: educational level, work experience, work experience squared, and
marital status. Work
experience is also used as a squared quantity as experience does not
necessarily show a linear relationship. The motivation for
the use of these variables come from previous researches and the underlying
theory. Furthermore, males will be used as a reference group.
In this research, we will use five mathematical
models. The variable female will be used as a dummy variable, meaning that a 1
will be given if female and 0 if male. Similarly, we will use dummy variables
for sector and marriage as well. The mathematical models will be estimated
using the Ordinary Least Squares (OLS)-method and with the use of a logarithmic
transformation of the regression models the results can be indicated as
percentages. For every model, a variable will be added to the previous model,
which will in turn create a clear distinction of the individual effect of a
variable on earnings. The models include:
The same models will be used for the private sector
instead of the public sector creating a set of results for both sectors. As the
mathematical models expands, we expect the coefficient of determination ( to rise as more of the uncertainty will be
explained. For every variable added, it is important to confirm there is no endogeneity,
which means the controlled variables correlate with the error term. This method
was chosen as other researches showed to find successful and reliable correlations
using these methods. (Bun & Van Ophem, 2017)
Acemoglu, D., & Autor, D. (2011). Lectures in Labor
Arulampalam, W., Booth, A. L.,
& Bryan, M. L. (2007, January). Is There a Glass Ceiling over Europe?
Exploring the Gender Pay Gap across the Wage Distribution . ILR Review,
Barón, Juan D.; Cobb-Clark, Deborah A. (2008): Occupational
segregation and the gender wage
gap in private- and public-sector employment: a distributional analysis, IZA
Discussion Papers, No. 3562,
Bun, M., & Van Ophem, H.
(2017). Reader Introduction Econometrics. Amsterdam: Faculty of
Economics and Business.
Cooksey, E., & Light, A.
(1979). The National Longitudinal Survey of Youth. Survey.
Elsesser, K. (2016, October 27). Important
facts about the global gender gap. Retrieved from Forbes:
Lucifora, C., & Meurs, D.
(2006). The public sector pay gap in France, Great Britian and Italy. The
Review of Income and Wealth, 43-59.
Smith, A. (1776). The Wealth of
Nations. New York: Modern Library.
Stewart, M. (2014). Why is the
Gender Pay Gap Higher in the Private Sector?
The Economist. (2017). Gary
Becker’s concept of human capital. Retrieved from The Economist: