Will gender parity break the glass ceiling?

Will Gender Parity Break the Glass
Ceiling? Evidence from a Randomized
Experiment
Preliminary
Manuel F. Bagüés & Berta Esteve-Volart
(Universidad Carlos III) (York University)
Motivation
• Gender parity or gender quotas imposed
or considered in many countries
– France: electoral party lists
– Norway: public enterprises’ boards
– Spain: cabinet, considering all public sector
recruitment committees (legislation project
approved by Government in March 8)
• No previous evidence of gender quota
effectiveness
• We use data on public exams in Spain
Why?
• Few women in top positions
– Politics: women occupy at least 30%
parliamentary seats in 12 out of 179 countries
– Boards of large private companies: women
are 2% in Spain, 3% in Italy, 4% in France
• Policy: from equal opportunities to gender
parity
– The failure of the pipeline theory
How?
•
•
-
Directly: women hire more women
Indirectly:
Role model transmission
Women in top positions can choose policy
more adequate for women,
- Private sector: flexible working hours
- Politics: public expenditure more useful to
women (Duflo and Chattopadhyay 2004)
Will it work?
Empirical evidence
• Data on individual productivity
– General: evidence of wage gap (Blau & Kahn 1994)
– Top management: Bertrand & Hallock (1999)
– Researchers: CSIC (2003), Veugelers (2006), Long
(1993), Mairesse & Turner (2002)
• Data on firm productivity (Wolfers 2006)
• Experimental data
– Blind Evaluation vs Non-Blind Evaluation
• Blank (1990), Goldin & Rouse (2000), Lavy (2005)
– Randomization
• Lab Experiments (Gneezy et al 2003)
Background Information
• We use data from public examinations in Spain
• They determine the access to public positions
(judiciary, diplomacy, notaries, economists, tax
inspectors, and many others)
• Every year 175,000 young university graduates
take public exams
• Only a small number of candidates pass exams
• Elite formation: many political figures had to
pass public exam (e.g. Aznar)
Characteristics of public exams
• Each committee examines 500 candidates
• Random allocation of candidates to
evaluating committees
• Evaluation
– Oral
– Two or three stages, all qualifying
– Voting by majority basis
– Multiple choice test introduced in 2003 for
some exams
Data
• All results are published in the state official bulletin
(BOE)
• We examine public exams to the judiciary, years
1995-2004 (new data: 1985-2005)
• Type of exams: judge, prosecutor, court secretary
• 150 committees
• 75,000 candidates involved
• About 1,700 judges, prosecutors and court
secretaries recruited
Data: what do we know?
• Characteristics of evaluators
– Gender, age, age of entry, rank
• Characteristics of successful candidates
for all years
– Gender, age, age of entry, rank
• Characteristics of all candidates for 2003
and 2004
• We do NOT know the individual vote of
each committee member
Empirical strategy
1) Committee-level information:
ycet  scet  X cet  et   cet
where y is an outcome variable, s is
female share in committee, X are
committee characteristics
Interpretation
1. Female evaluators are tougher with
female candidates
2. Male evaluators are more generous with
female candidates
•
Possible non-linearities?
2) Candidate-level information for years
2003 and 2004 (multiple choice test):
Quantitatively
• A female candidate’s chances to pass the
public exam are 5.5% greater if evaluated
by a committee with fewer women than the
median committee, than if evaluated by a
committee with more women that the
median committee
Caveats
•
1)
2)
-
-
What is the motivation of the evaluators?
Evaluators have ‘irrational taste’
Evaluators behave according to rational choice but:
Women think women are worse (lack of
confidence)
Since the men in committees discriminated in the
past, men in committees now are more generous
with female candidates (past discrimination)
Women want to increase their group’s average
quality (statistical discrimination)
Next step
• Evolution over time of the observed gender
bias
– What happened since the first committee with
a female member?
• Data before 1995