socioeconomic determinants for fertility - demografi · how to measure fertility? i fertility...

133
Socioeconomic Determinants for Fertility Mette Gerster PhD Defense, September 15th 2009

Upload: buiduong

Post on 29-May-2019

233 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Socioeconomic Determinants for Fertility

Mette Gerster

PhD Defense, September 15th 2009

Page 2: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

How to measure fertility?

I Fertility - actual births

I Fertility is a process that evolves over several years (inprinciple, from menarche to menopause)

I Several aspects are potentially of interest - several ways tomeasure it:

1. Number of children at a given age → static or2. the parity progressions → dynamic (magnifying glass)

Page 3: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

How to measure fertility?

I Fertility - actual births

I Fertility is a process that evolves over several years (inprinciple, from menarche to menopause)

I Several aspects are potentially of interest - several ways tomeasure it:

1. Number of children at a given age → static or2. the parity progressions → dynamic (magnifying glass)

Page 4: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

How to measure fertility?

I Fertility - actual births

I Fertility is a process that evolves over several years (inprinciple, from menarche to menopause)

I Several aspects are potentially of interest - several ways tomeasure it:

1. Number of children at a given age → static or2. the parity progressions → dynamic (magnifying glass)

Page 5: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

How to measure fertility?

I Fertility - actual births

I Fertility is a process that evolves over several years (inprinciple, from menarche to menopause)

I Several aspects are potentially of interest - several ways tomeasure it:

1. Number of children at a given age → static or2. the parity progressions → dynamic (magnifying glass)

Page 6: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

How to measure fertility?

I Fertility - actual births

I Fertility is a process that evolves over several years (inprinciple, from menarche to menopause)

I Several aspects are potentially of interest - several ways tomeasure it:

1. Number of children at a given age → static or2. the parity progressions → dynamic (magnifying glass)

Page 7: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Determinants for fertility

I Biological factors, health, fecundity

I Factors which potentially influence the (woman’s) choice(when) to have children

I I will give two examples of the latter: the socioeconomicfactors

1. education and2. labour market attachment

Page 8: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Determinants for fertility

I Biological factors, health, fecundity

I Factors which potentially influence the (woman’s) choice(when) to have children

I I will give two examples of the latter: the socioeconomicfactors

1. education and2. labour market attachment

Page 9: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Determinants for fertility

I Biological factors, health, fecundity

I Factors which potentially influence the (woman’s) choice(when) to have children

I I will give two examples of the latter: the socioeconomicfactors

1. education and2. labour market attachment

Page 10: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

t

Page 11: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

tt−

X(t−)

Page 12: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

t t ++ ∆∆tt−

X(t−)

Page 13: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

t t ++ ∆∆tt−

X(t−)

T

X((T)) ↔↔ N((T))

Page 14: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Education and labour market attachmentEffect on fertility?

I Subject of numerous studies in the demographic literature formany years

I Economic and sociological theory provide a theoreticalframework for the underlying mechanisms

I Gary Becker, Nobel Prize (economics) 1992 - A Treatise onthe Family [Becker, 1991]

Page 15: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Education and labour market attachmentEffect on fertility?

I Subject of numerous studies in the demographic literature formany years

I Economic and sociological theory provide a theoreticalframework for the underlying mechanisms

I Gary Becker, Nobel Prize (economics) 1992 - A Treatise onthe Family [Becker, 1991]

Page 16: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Education and labour market attachmentEffect on fertility?

I Subject of numerous studies in the demographic literature formany years

I Economic and sociological theory provide a theoreticalframework for the underlying mechanisms

I Gary Becker, Nobel Prize (economics) 1992 - A Treatise onthe Family [Becker, 1991]

Page 17: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Two examples

Parity transitions

I Labour market attachment

I Norway

I Simultaneuos EquationsModels

I Administrative registerdata:Statistisk Sentralbyra

Ultimate fertility

I Educational attainment

I Denmark

I Marginal StructuralModels

I Administrative registerdata:Danmarks Statistik

Page 18: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Two examples

Parity transitions

I Labour market attachment

I Norway

I Simultaneuos EquationsModels

I Administrative registerdata:Statistisk Sentralbyra

Ultimate fertility

I Educational attainment

I Denmark

I Marginal StructuralModels

I Administrative registerdata:Danmarks Statistik

Page 19: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Two examples

Parity transitions

I Labour market attachment

I Norway

I Simultaneuos EquationsModels

I Administrative registerdata:Statistisk Sentralbyra

Ultimate fertility

I Educational attainment

I Denmark

I Marginal StructuralModels

I Administrative registerdata:Danmarks Statistik

Page 20: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Two examples

Parity transitions

I Labour market attachment

I Norway

I Simultaneuos EquationsModels

I Administrative registerdata:Statistisk Sentralbyra

Ultimate fertility

I Educational attainment

I Denmark

I Marginal StructuralModels

I Administrative registerdata:Danmarks Statistik

Page 21: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Two examples

Parity transitions

I Labour market attachment

I Norway

I Simultaneuos EquationsModels

I Administrative registerdata:Statistisk Sentralbyra

Ultimate fertility

I Educational attainment

I Denmark

I Marginal StructuralModels

I Administrative registerdata:Danmarks Statistik

Page 22: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Overview

Parity transitions in Norway

Ultimate fertility in Denmark

Page 23: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Background

I Transition from being a one-child mother to a two-childmother and from two-child to three-child mother

I How does it depend on her current labour market attachment(employed vs non-employed)?

I Is this relationship possibly different across the parities?

I Do unobserved characteristics of the women play a role?

PhD Defense, September 2009

Page 24: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Background

I Transition from being a one-child mother to a two-childmother and from two-child to three-child mother

I How does it depend on her current labour market attachment(employed vs non-employed)?

I Is this relationship possibly different across the parities?

I Do unobserved characteristics of the women play a role?

PhD Defense, September 2009

Page 25: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Background

I Transition from being a one-child mother to a two-childmother and from two-child to three-child mother

I How does it depend on her current labour market attachment(employed vs non-employed)?

I Is this relationship possibly different across the parities?

I Do unobserved characteristics of the women play a role?

PhD Defense, September 2009

Page 26: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Background

I Transition from being a one-child mother to a two-childmother and from two-child to three-child mother

I How does it depend on her current labour market attachment(employed vs non-employed)?

I Is this relationship possibly different across the parities?

I Do unobserved characteristics of the women play a role?

PhD Defense, September 2009

Page 27: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Study population

I All women of NO-origin

I whose first child reaches age 15 mths April 1994-Oct 2002

I 19-40 years old at first birth

I registered with a partner at first birth

I no students

I → 126608 women

PhD Defense, September 2009

Page 28: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Study population

I All women of NO-origin

I whose first child reaches age 15 mths April 1994-Oct 2002

I 19-40 years old at first birth

I registered with a partner at first birth

I no students

I → 126608 women

PhD Defense, September 2009

Page 29: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Study population

I All women of NO-origin

I whose first child reaches age 15 mths April 1994-Oct 2002

I 19-40 years old at first birth

I registered with a partner at first birth

I no students

I → 126608 women

PhD Defense, September 2009

Page 30: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Why effect of employment?

I Employment status might influence the decision to have thenext child via several channels:

I Periods away from the labour market are potentially morecostly for women who are currently in a job

1. loss of skills (human capital)2. forgone income

I the right to paid maternity leave

I Can better afford to have a child?

PhD Defense, September 2009

Page 31: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Why effect of employment?

I Employment status might influence the decision to have thenext child via several channels:

I Periods away from the labour market are potentially morecostly for women who are currently in a job

1. loss of skills (human capital)2. forgone income

I the right to paid maternity leave

I Can better afford to have a child?

PhD Defense, September 2009

Page 32: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Why effect of employment?

I Employment status might influence the decision to have thenext child via several channels:

I Periods away from the labour market are potentially morecostly for women who are currently in a job

1. loss of skills (human capital)2. forgone income

I the right to paid maternity leave

I Can better afford to have a child?

PhD Defense, September 2009

Page 33: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Why effect of employment?

I Employment status might influence the decision to have thenext child via several channels:

I Periods away from the labour market are potentially morecostly for women who are currently in a job

1. loss of skills (human capital)2. forgone income

I the right to paid maternity leave

I Can better afford to have a child?

PhD Defense, September 2009

Page 34: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Why effect of employment?

I Employment status might influence the decision to have thenext child via several channels:

I Periods away from the labour market are potentially morecostly for women who are currently in a job

1. loss of skills (human capital)2. forgone income

I the right to paid maternity leave

I Can better afford to have a child?

PhD Defense, September 2009

Page 35: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Why effect of employment?

I Employment status might influence the decision to have thenext child via several channels:

I Periods away from the labour market are potentially morecostly for women who are currently in a job

1. loss of skills (human capital)2. forgone income

I the right to paid maternity leave

I Can better afford to have a child?

PhD Defense, September 2009

Page 36: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Unobserved heterogeneityIntuition

Why...

I Possibly influencing thebirth intensities (e.g. morefamily-orientation)

I Possibly influencing theemployment process (e.g.career-orientation)

I Might give rise to aspurious relationship

I Potentially correlated

I The employment status isendogenous (as opposedto exogenous)

How...I Set up model equations for

the births with randomeffect

I Set up model equations forthe employment processwith random effect(s)

I allow these random effectsto be correlated byestimating these equationssimultaneously

Page 37: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Unobserved heterogeneityIntuition

Why...

I Possibly influencing thebirth intensities (e.g. morefamily-orientation)

I Possibly influencing theemployment process (e.g.career-orientation)

I Might give rise to aspurious relationship

I Potentially correlated

I The employment status isendogenous (as opposedto exogenous)

How...I Set up model equations for

the births with randomeffect

I Set up model equations forthe employment processwith random effect(s)

I allow these random effectsto be correlated byestimating these equationssimultaneously

Page 38: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Unobserved heterogeneityIntuition

Why...

I Possibly influencing thebirth intensities (e.g. morefamily-orientation)

I Possibly influencing theemployment process (e.g.career-orientation)

I Might give rise to aspurious relationship

I Potentially correlated

I The employment status isendogenous (as opposedto exogenous)

How...I Set up model equations for

the births with randomeffect

I Set up model equations forthe employment processwith random effect(s)

I allow these random effectsto be correlated byestimating these equationssimultaneously

Page 39: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Unobserved heterogeneityIntuition

Why...

I Possibly influencing thebirth intensities (e.g. morefamily-orientation)

I Possibly influencing theemployment process (e.g.career-orientation)

I Might give rise to aspurious relationship

I Potentially correlated

I The employment status isendogenous (as opposedto exogenous)

How...I Set up model equations for

the births with randomeffect

I Set up model equations forthe employment processwith random effect(s)

I allow these random effectsto be correlated byestimating these equationssimultaneously

Page 40: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Unobserved heterogeneityIntuition

Why...

I Possibly influencing thebirth intensities (e.g. morefamily-orientation)

I Possibly influencing theemployment process (e.g.career-orientation)

I Might give rise to aspurious relationship

I Potentially correlated

I The employment status isendogenous (as opposedto exogenous)

How...I Set up model equations for

the births with randomeffect

I Set up model equations forthe employment processwith random effect(s)

I allow these random effectsto be correlated byestimating these equationssimultaneously

Page 41: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Unobserved heterogeneityIntuition

Why...

I Possibly influencing thebirth intensities (e.g. morefamily-orientation)

I Possibly influencing theemployment process (e.g.career-orientation)

I Might give rise to aspurious relationship

I Potentially correlated

I The employment status isendogenous (as opposedto exogenous)

How...I Set up model equations for

the births with randomeffect

I Set up model equations forthe employment processwith random effect(s)

I allow these random effectsto be correlated byestimating these equationssimultaneously

Page 42: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Unobserved heterogeneityIntuition

Why...

I Possibly influencing thebirth intensities (e.g. morefamily-orientation)

I Possibly influencing theemployment process (e.g.career-orientation)

I Might give rise to aspurious relationship

I Potentially correlated

I The employment status isendogenous (as opposedto exogenous)

How...I Set up model equations for

the births with randomeffect

I Set up model equations forthe employment processwith random effect(s)

I allow these random effectsto be correlated byestimating these equationssimultaneously

Page 43: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Unobserved heterogeneityIntuition

Why...

I Possibly influencing thebirth intensities (e.g. morefamily-orientation)

I Possibly influencing theemployment process (e.g.career-orientation)

I Might give rise to aspurious relationship

I Potentially correlated

I The employment status isendogenous (as opposedto exogenous)

How...I Set up model equations for

the births with randomeffect

I Set up model equations forthe employment processwith random effect(s)

I allow these random effectsto be correlated byestimating these equationssimultaneously

Page 44: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Unobserved heterogeneityIntuition

Why...

I Possibly influencing thebirth intensities (e.g. morefamily-orientation)

I Possibly influencing theemployment process (e.g.career-orientation)

I Might give rise to aspurious relationship

I Potentially correlated

I The employment status isendogenous (as opposedto exogenous)

How...I Set up model equations for

the births with randomeffect

I Set up model equations forthe employment processwith random effect(s)

I allow these random effectsto be correlated byestimating these equationssimultaneously

Page 45: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Unobserved heterogeneityIntuition

Why...

I Possibly influencing thebirth intensities (e.g. morefamily-orientation)

I Possibly influencing theemployment process (e.g.career-orientation)

I Might give rise to aspurious relationship

I Potentially correlated

I The employment status isendogenous (as opposedto exogenous)

How...I Set up model equations for

the births with randomeffect

I Set up model equations forthe employment processwith random effect(s)

I allow these random effectsto be correlated byestimating these equationssimultaneously

Page 46: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Model

Birth intensities

log λ2(t) = log λ(2)0 (t) + β′2 · X (2)(t−)

log λ3(t) = log λ(3)0 (t) + β′3 · X (3)(t−)

where t denotes age of previous child (-15 months)

Employment and non-employment intensities

log λe(s) = log λ(e)0 (s) + β′e · X (e)(s)

log λne(s) = log λ(ne)0 (s) + β′ne · X (ne)(s)

where s denotes time since beginning of each spell

→ Simultaneous Equations Model (SEM)

Page 47: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Model

Birth intensities

log λ2(t) = log λ(2)0 (t) + β′2 · X (2)(t−)

log λ3(t) = log λ(3)0 (t) + β′3 · X (3)(t−)

where t denotes age of previous child (-15 months)

Employment and non-employment intensities

log λe(s) = log λ(e)0 (s) + β′e · X (e)(s)

log λne(s) = log λ(ne)0 (s) + β′ne · X (ne)(s)

where s denotes time since beginning of each spell

→ Simple Model (SM)

Page 48: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Model

Birth intensities

log λ2(t) = log λ(2)0 (t) + β′2 · X (2)(t−) + εb

log λ3(t) = log λ(3)0 (t) + β′3 · X (3)(t−) + εb

where t denotes age of previous child (-15 months)

Employment and non-employment intensities

log λe(s) = log λ(e)0 (s) + β′e · X (e)(s) + εe

log λne(s) = log λ(ne)0 (s) + β′ne · X (ne)(s) + εne

where s denotes time since beginning of each spell

→ Simultaneous Equations Model (SEM)

Page 49: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Not quite done...

Assume that

I (εb εe εne)T ∼ N3(0,Ωεb,εe ,εne )

I conditional on (εb εe εne)T , the separate birth spells foreach woman are independent

I and so are the employment and non-employment spells

Page 50: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Not quite done...

Assume that

I (εb εe εne)T ∼ N3(0,Ωεb,εe ,εne )

I conditional on (εb εe εne)T , the separate birth spells foreach woman are independent

I and so are the employment and non-employment spells

Page 51: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Not quite done...

Assume that

I (εb εe εne)T ∼ N3(0,Ωεb,εe ,εne )

I conditional on (εb εe εne)T , the separate birth spells foreach woman are independent

I and so are the employment and non-employment spells

Page 52: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Not quite done...

Assume that

I (εb εe εne)T ∼ N3(0,Ωεb,εe ,εne )

I conditional on (εb εe εne)T , the separate birth spells foreach woman are independent

I and so are the employment and non-employment spells

Page 53: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Results

2nd child:

Model SEM Model SM

RR p RR p

Employed (ref) 1 - 1 -Non-employed 0.929 < 0.01 0.956 < 0.01

Controlled for...

mother’s age, calendar year, and education.

PhD Defense, September 2009

Page 54: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Results

3rd child:

Model SEM Model SM

RR p RR p

Employed (ref) 1 < 0.01 1 < 0.01Non-employed 1.097 < 0.01 1.132 < 0.01

Controlled for...

mother’s age, calendar year, and education

PhD Defense, September 2009

Page 55: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

ResultsUnobserved heterogeneity

sd (εb) 0.379 corr (εb, εe) −0.245sd (εe) 1.436 corr (εb, εne) −0.318sd (εne) 0.782 corr (εe , εne) 0.551

PhD Defense, September 2009

Page 56: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

ConclusionParity transitions in Norway

I The second birth intensity is smaller for non-employed women(RR = 0.93)

I The third birth intensity is larger for non-employed women(RR = 1.097)

I Child 2: when??

I Child 3: if??

PhD Defense, September 2009

Page 57: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

ConclusionParity transitions in Norway

I The second birth intensity is smaller for non-employed women(RR = 0.93)

I The third birth intensity is larger for non-employed women(RR = 1.097)

I Child 2: when??

I Child 3: if??

PhD Defense, September 2009

Page 58: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

ConclusionParity transitions in Norway

I The second birth intensity is smaller for non-employed women(RR = 0.93)

I The third birth intensity is larger for non-employed women(RR = 1.097)

I Child 2: when??

I Child 3: if??

PhD Defense, September 2009

Page 59: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

ConclusionParity transitions in Norway

I The second birth intensity is smaller for non-employed women(RR = 0.93)

I The third birth intensity is larger for non-employed women(RR = 1.097)

I Child 2: when??

I Child 3: if??

PhD Defense, September 2009

Page 60: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Overview

Parity transitions in Norway

Ultimate fertility in Denmark

Page 61: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Illustration

t t ++ ∆∆tt−

X(t−)

T

X((T)) ↔↔ N((T))

Page 62: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Ultimate fertility

I Number of children at age 41

I How does it depend on educational attainment?

I Is this relationship static?

I Feedback...

PhD Defense, September 2009

Page 63: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Ultimate fertility

I Number of children at age 41

I How does it depend on educational attainment?

I Is this relationship static?

I Feedback...

PhD Defense, September 2009

Page 64: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Ultimate fertility

I Number of children at age 41

I How does it depend on educational attainment?

I Is this relationship static?

I Feedback...

PhD Defense, September 2009

Page 65: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Ultimate fertility

I Number of children at age 41

I How does it depend on educational attainment?

I Is this relationship static?

I Feedback...

PhD Defense, September 2009

Page 66: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Why an effect of education on fertility?

I Women with a higher education might have higheropportunity costs - more likely to pursue a career

I Their labour market situation might be more flexible - easierto combine

I Other factors? More resources?

PhD Defense, September 2009

Page 67: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Why an effect of education on fertility?

I Women with a higher education might have higheropportunity costs - more likely to pursue a career

I Their labour market situation might be more flexible - easierto combine

I Other factors? More resources?

PhD Defense, September 2009

Page 68: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Why an effect of education on fertility?

I Women with a higher education might have higheropportunity costs - more likely to pursue a career

I Their labour market situation might be more flexible - easierto combine

I Other factors? More resources?

PhD Defense, September 2009

Page 69: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

The study populationAll women who...

I born in 1963

I living in Denmark Jan 1st 1981 (and each year 1982-2005)

I of Danish origin

I who have completed a preparatory upper secondary education(PUSE, da: Studentereksamen) no later than October 1983

PhD Defense, September 2009

Page 70: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

DescriptivesEducation and fertility (2005)

Education % chless Avg. children Frequency Per cent

PUSE 17.8 1.71 1169 14.6Vocational 13.2 1.84 1317 16.4Short tertiary 14.7 1.80 672 8.4Medium tertiary 11.4 1.97 3544 44.1Long tertiary 16.7 1.81 1330 16.6

Total 13.8 1.87 8032 100.1

PhD Defense, September 2009

Page 71: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

DescriptivesEducation and fertility (2005)

Education % chless Avg. children Frequency Per cent

PUSE 17.8 1.71 1169 14.6Vocational 13.2 1.84 1317 16.4Short tertiary 14.7 1.80 672 8.4Medium tertiary 11.4 1.97 3544 44.1Long tertiary 16.7 1.81 1330 16.6

Total 13.8 1.87 8032 100.1

PhD Defense, September 2009

Page 72: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

DescriptivesEducation and fertility (2005)

Education % chless Avg. children Frequency Per cent

PUSE 17.8 1.71 1169 14.6Vocational 13.2 1.84 1317 16.4Short tertiary 14.7 1.80 672 8.4Medium tertiary 11.4 1.97 3544 44.1Long tertiary 16.7 1.81 1330 16.6

Total 13.8 1.87 8032 100.1

PhD Defense, September 2009

Page 73: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

DescriptivesEducation and fertility (2005)

Education % chless Avg. children Frequency Per cent

PUSE 17.8 1.71 1169 14.6Vocational 13.2 1.84 1317 16.4Short tertiary 14.7 1.80 672 8.4Medium tertiary 11.4 1.97 3544 44.1Long tertiary 16.7 1.81 1330 16.6

Total 13.8 1.87 8032 100.1

PhD Defense, September 2009

Page 74: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

DescriptivesEducation and fertility (2005)

Education % chless Avg. children Frequency Per cent

PUSE 17.8 1.71 1169 14.6Vocational 13.2 1.84 1317 16.4Short tertiary 14.7 1.80 672 8.4Medium tertiary 11.4 1.97 3544 44.1Long tertiary 16.7 1.81 1330 16.6

Total 13.8 1.87 8032 100.1

PhD Defense, September 2009

Page 75: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

FeedbackIntuition

I We wish to assess to which extent educational differences inultimate fertility are attributable to feedback patterns

I Example:I Assume women who become mothers while enrolled in

university are more inclined to interrupt/change to a shorterone (deviate from their original strategy as a result of theirfertility)

I → fewer children among highly educated women

I The birth process itself acts as a time-dependent confounderfor the effect of education on fertility

PhD Defense, September 2009

Page 76: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

FeedbackIntuition

I We wish to assess to which extent educational differences inultimate fertility are attributable to feedback patterns

I Example:I Assume women who become mothers while enrolled in

university are more inclined to interrupt/change to a shorterone (deviate from their original strategy as a result of theirfertility)

I → fewer children among highly educated women

I The birth process itself acts as a time-dependent confounderfor the effect of education on fertility

PhD Defense, September 2009

Page 77: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

FeedbackIntuition

I We wish to assess to which extent educational differences inultimate fertility are attributable to feedback patterns

I Example:I Assume women who become mothers while enrolled in

university are more inclined to interrupt/change to a shorterone (deviate from their original strategy as a result of theirfertility)

I → fewer children among highly educated women

I The birth process itself acts as a time-dependent confounderfor the effect of education on fertility

PhD Defense, September 2009

Page 78: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

FeedbackIntuition

I We wish to assess to which extent educational differences inultimate fertility are attributable to feedback patterns

I Example:I Assume women who become mothers while enrolled in

university are more inclined to interrupt/change to a shorterone (deviate from their original strategy as a result of theirfertility)

I → fewer children among highly educated women

I The birth process itself acts as a time-dependent confounderfor the effect of education on fertility

PhD Defense, September 2009

Page 79: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

FeedbackIntuition

I We wish to assess to which extent educational differences inultimate fertility are attributable to feedback patterns

I Example:I Assume women who become mothers while enrolled in

university are more inclined to interrupt/change to a shorterone (deviate from their original strategy as a result of theirfertility)

I → fewer children among highly educated women

I The birth process itself acts as a time-dependent confounderfor the effect of education on fertility

PhD Defense, September 2009

Page 80: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

FeedbackIs it in the data? Can we remove it?

1. Is feedback present in the study population at hand?

2. If so, to which extent are the educational differences inultimate fertility attributable to the feedback?

3. Educational differences in ultimate fertility if there were nofeedback?

4. One particular aspect of ultimate fertility: What is theprobability of having 3 children at age 41 for differenteducational attainments?

PhD Defense, September 2009

Page 81: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Feedback (Endogeneity)Definition [Hernan et al., 2001]

I Assume study population followed throughout the time-period0, 1, . . . ,T

I Let B(t) be the fertility process and E (t) the educationprocess

I Feedback: If there exists t ∈ 0, 1, . . . ,T s.t. the condition

E (t)∐

B(t) | (E (t − 1),Z)

is not met.

I Endogeneity (vs. exogeneity)

PhD Defense, September 2009

Page 82: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Feedback (Endogeneity)Definition [Hernan et al., 2001]

I Assume study population followed throughout the time-period0, 1, . . . ,T

I Let B(t) be the fertility process and E (t) the educationprocess

I Feedback: If there exists t ∈ 0, 1, . . . ,T s.t. the condition

E (t)∐

B(t) | (E (t − 1),Z)

is not met.

I Endogeneity (vs. exogeneity)

PhD Defense, September 2009

Page 83: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Feedback (Endogeneity)Definition [Hernan et al., 2001]

I Assume study population followed throughout the time-period0, 1, . . . ,T

I Let B(t) be the fertility process and E (t) the educationprocess

I Feedback: If there exists t ∈ 0, 1, . . . ,T s.t. the condition

E (t)∐

B(t) | (E (t − 1),Z)

is not met.

I Endogeneity (vs. exogeneity)

PhD Defense, September 2009

Page 84: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Feedback (Endogeneity)Definition [Hernan et al., 2001]

I Assume study population followed throughout the time-period0, 1, . . . ,T

I Let B(t) be the fertility process and E (t) the educationprocess

I Feedback: If there exists t ∈ 0, 1, . . . ,T s.t. the condition

E (t)∐

B(t) | (E (t − 1),Z)

is not met.

I Endogeneity (vs. exogeneity)

PhD Defense, September 2009

Page 85: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Is feedback present in the study population?

Young mothers and drop-outs

I Mother before 1986 - education in 2005? Table

I Leaving education after birth - education in 2005? Table

Model probability of dropping out of education

I Yit : indicator for woman i interrupting educational enrolmentin year t (cond. on being enrolled)

I logit [Pr(Yit | Xit ,Zi , enrolled)] = α + β′ · Xit + γ′ · Zi

I Interaction between giving birth and education in which she isenrolled Illustration

PhD Defense, September 2009

Page 86: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Is feedback present in the study population?

Young mothers and drop-outs

I Mother before 1986 - education in 2005? Table

I Leaving education after birth - education in 2005? Table

Model probability of dropping out of education

I Yit : indicator for woman i interrupting educational enrolmentin year t (cond. on being enrolled)

I logit [Pr(Yit | Xit ,Zi , enrolled)] = α + β′ · Xit + γ′ · Zi

I Interaction between giving birth and education in which she isenrolled Illustration

PhD Defense, September 2009

Page 87: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Educational differences if there were no feedback?Marginal Structural Models (MSM) [Hernan et al., 2001]

Potential Outcomes:

Y e : the indicator of being a mother of 3 children (as opposed toless than 3) by age 41 if educational strategy e were followed

Marginal Structural Model (MSM):

logit[Pr(Y e = 1 | Zi )

]= δ1 + ε1 · Zi + φ1 · f (e)

How to assess information on potential outcomes?

Use observed data - along with a suitable set of assumptions

Page 88: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Educational differences if there were no feedback?Marginal Structural Models (MSM) [Hernan et al., 2001]

Potential Outcomes:

Y e : the indicator of being a mother of 3 children (as opposed toless than 3) by age 41 if educational strategy e were followed

Marginal Structural Model (MSM):

logit[Pr(Y e = 1 | Zi )

]= δ1 + ε1 · Zi + φ1 · f (e)

How to assess information on potential outcomes?

Use observed data - along with a suitable set of assumptions

Page 89: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Educational differences if there were no feedback?Marginal Structural Models (MSM) [Hernan et al., 2001]

Potential Outcomes:

Y e : the indicator of being a mother of 3 children (as opposed toless than 3) by age 41 if educational strategy e were followed

Marginal Structural Model (MSM):

logit[Pr(Y e = 1 | Zi )

]= δ1 + ε1 · Zi + φ1 · f (e)

How to assess information on potential outcomes?

Use observed data - along with a suitable set of assumptions

Page 90: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Marginal Structural Models (contd)Inverse Probability of Treatment Weights

I Idea: re-weight the original population to construct ahypothetical (pseudo-) population which is free of feedback

I

SW i (T + 1) =∏s≤T

Pr(Ei (s) = eis | E i (s − 1),Zi

)Pr(Ei (s) = eis | E i (s − 1),B(s),Zi

)I Recall the definition of feedback:

If there exists t ∈ 0, 1, . . . ,T s.t. the condition

E (t)∐

B(t) | (E (t − 1),Z)

is not met

I The weights need to be estimated - need models

Page 91: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Marginal Structural Models (contd)Inverse Probability of Treatment Weights

I Idea: re-weight the original population to construct ahypothetical (pseudo-) population which is free of feedback

I

SW i (T + 1) =∏s≤T

Pr(Ei (s) = eis | E i (s − 1),Zi

)Pr(Ei (s) = eis | E i (s − 1),B(s),Zi

)I Recall the definition of feedback:

If there exists t ∈ 0, 1, . . . ,T s.t. the condition

E (t)∐

B(t) | (E (t − 1),Z)

is not met

I The weights need to be estimated - need models

Page 92: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Marginal Structural Models (contd)Inverse Probability of Treatment Weights

I Idea: re-weight the original population to construct ahypothetical (pseudo-) population which is free of feedback

I

SW i (T + 1) =∏s≤T

Pr(Ei (s) = eis | E i (s − 1),Zi

)Pr(Ei (s) = eis | E i (s − 1),B(s),Zi

)I Recall the definition of feedback:

If there exists t ∈ 0, 1, . . . ,T s.t. the condition

E (t)∐

B(t) | (E (t − 1),Z)

is not met

I The weights need to be estimated - need models

Page 93: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Marginal Structural Models (contd)Inverse Probability of Treatment Weights

I Idea: re-weight the original population to construct ahypothetical (pseudo-) population which is free of feedback

I

SW i (T + 1) =∏s≤T

Pr(Ei (s) = eis | E i (s − 1),Zi

)Pr(Ei (s) = eis | E i (s − 1),B(s),Zi

)I Recall the definition of feedback:

If there exists t ∈ 0, 1, . . . ,T s.t. the condition

E (t)∐

B(t) | (E (t − 1),Z)

is not met

I The weights need to be estimated - need models

Page 94: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

The hypothetical populationThe pseudo-population

I By employing the weighting technique we get a hypotheticalpopulation in which some women are ”weighted up” and someare ”weighted down” -and by construction the educational attainment in 2005 is notaffected by previous fertility

I Hence, by using this population we can answer the questionWhat would be the educational differences in the odds ofbeing a mother of 3 - if there were no feedback in the data?

PhD Defense, September 2009

Page 95: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

The hypothetical populationThe pseudo-population

I By employing the weighting technique we get a hypotheticalpopulation in which some women are ”weighted up” and someare ”weighted down” -and by construction the educational attainment in 2005 is notaffected by previous fertility

I Hence, by using this population we can answer the questionWhat would be the educational differences in the odds ofbeing a mother of 3 - if there were no feedback in the data?

PhD Defense, September 2009

Page 96: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Example

Imagine a woman who...

I Takes her Studentereksamen in 1983, enrols in university

I Becomes a mother 1984

I Leaves university, start nursing school 1985

I Graduates 1991, 2 more children at ages 30 and 32

Weights:

I Pr(Ei (1985) = nurse | E i (1984),Zi

)= 1

20

I Pr(Ei (1985) = nurse | E i (1984),B(1984),Zi

)= 1

10

I SW i (2005) = 1 · 1 · (10)/20 · 1 · · · 1 = 0.5

Page 97: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Example

Imagine a woman who...

I Takes her Studentereksamen in 1983, enrols in university

I Becomes a mother 1984

I Leaves university, start nursing school 1985

I Graduates 1991, 2 more children at ages 30 and 32

Weights:

I Pr(Ei (1985) = nurse | E i (1984),Zi

)= 1

20

I Pr(Ei (1985) = nurse | E i (1984),B(1984),Zi

)= 1

10

I SW i (2005) = 1 · 1 · (10)/20 · 1 · · · 1 = 0.5

Page 98: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Example

Imagine a woman who...

I Takes her Studentereksamen in 1983, enrols in university

I Becomes a mother 1984

I Leaves university, start nursing school 1985

I Graduates 1991, 2 more children at ages 30 and 32

Weights:

I Pr(Ei (1985) = nurse | E i (1984),Zi

)= 1

20

I Pr(Ei (1985) = nurse | E i (1984),B(1984),Zi

)= 1

10

I SW i (2005) = 1 · 1 · (10)/20 · 1 · · · 1 = 0.5

Page 99: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Example

Imagine a woman who...

I Takes her Studentereksamen in 1983, enrols in university

I Becomes a mother 1984

I Leaves university, start nursing school 1985

I Graduates 1991, 2 more children at ages 30 and 32

Weights:

I Pr(Ei (1985) = nurse | E i (1984),Zi

)= 1

20

I Pr(Ei (1985) = nurse | E i (1984),B(1984),Zi

)= 1

10

I SW i (2005) = 1 · 1 · (10)/20 · 1 · · · 1 = 0.5

Page 100: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

ResultsActual vs. hypothetical population

Actual: Hypothetical:Education (2005) OR p-value OR p-value

Never enrolled 0.84 0.15 0.36 < .0001Prev enrolled 0.81 0.12 0.46 < .0001Tert(s)/voc 0.92 0.40 0.51 < .0001Tert(m) 1.32 0.001 0.65 < .0001Tert(l) (REF) 1 − 1 −Enrolled: T(s)/voc 2.54 0.02 1.13 0.80Enrolled: T(m) 1.47 0.07 1.22 0.24Enrolled: T(l) 0.97 0.91 1.52 0.01

(controlled for baseline variables)

Page 101: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

ResultsActual vs. hypothetical population

Actual: Hypothetical:Education (2005) OR p-value OR p-value

Never enrolled 0.84 0.15 0.36 < .0001Prev enrolled 0.81 0.12 0.46 < .0001Tert(s)/voc 0.92 0.40 0.51 < .0001Tert(m) 1.32 0.001 0.65 < .0001Tert(l) (REF) 1 − 1 −Enrolled: T(s)/voc 2.54 0.02 1.13 0.80Enrolled: T(m) 1.47 0.07 1.22 0.24Enrolled: T(l) 0.97 0.91 1.52 0.01

(controlled for baseline variables)

Page 102: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

ResultsActual vs. hypothetical population

Actual: Hypothetical:Education (2005) OR p-value OR p-value

Never enrolled 0.84 0.15 0.36 < .0001Prev enrolled 0.81 0.12 0.46 < .0001Tert(s)/voc 0.92 0.40 0.51 < .0001Tert(m) 1.32 0.001 0.65 < .0001Tert(l) (REF) 1 − 1 −Enrolled: T(s)/voc 2.54 0.02 1.13 0.80Enrolled: T(m) 1.47 0.07 1.22 0.24Enrolled: T(l) 0.97 0.91 1.52 0.01

(controlled for baseline variables)

Page 103: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

ResultsActual vs. hypothetical population

Actual: Hypothetical:Education (2005) OR p-value OR p-value

Never enrolled 0.84 0.15 0.36 < .0001Prev enrolled 0.81 0.12 0.46 < .0001Tert(s)/voc 0.92 0.40 0.51 < .0001Tert(m) 1.32 0.001 0.65 < .0001Tert(l) (REF) 1 − 1 −Enrolled: T(s)/voc 2.54 0.02 1.13 0.80Enrolled: T(m) 1.47 0.07 1.22 0.24Enrolled: T(l) 0.97 0.91 1.52 0.01

(controlled for baseline variables)

Page 104: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

ResultsActual vs. hypothetical population

Actual: Hypothetical:Education (2005) OR p-value OR p-value

Never enrolled 0.84 0.15 0.36 < .0001Prev enrolled 0.81 0.12 0.46 < .0001Tert(s)/voc 0.92 0.40 0.51 < .0001Tert(m) 1.32 0.001 0.65 < .0001Tert(l) (REF) 1 − 1 −Enrolled: T(s)/voc 2.54 0.02 1.13 0.80Enrolled: T(m) 1.47 0.07 1.22 0.24Enrolled: T(l) 0.97 0.91 1.52 0.01

(controlled for baseline variables)

Page 105: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

ResultsActual vs. hypothetical population

Actual: Hypothetical:Education (2005) OR p-value OR p-value

Never enrolled 0.84 0.15 0.36 < .0001Prev enrolled 0.81 0.12 0.46 < .0001Tert(s)/voc 0.92 0.40 0.51 < .0001Tert(m) 1.32 0.001 0.65 < .0001Tert(l) (REF) 1 − 1 −Enrolled: T(s)/voc 2.54 0.02 1.13 0.80Enrolled: T(m) 1.47 0.07 1.22 0.24Enrolled: T(l) 0.97 0.91 1.52 0.01

(controlled for baseline variables)

Page 106: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

ResultsActual vs. hypothetical population

Actual: Hypothetical:Education (2005) OR p-value OR p-value

Never enrolled 0.84 0.15 0.36 < .0001Prev enrolled 0.81 0.12 0.46 < .0001Tert(s)/voc 0.92 0.40 0.51 < .0001Tert(m) 1.32 0.001 0.65 < .0001Tert(l) (REF) 1 − 1 −Enrolled: T(s)/voc 2.54 0.02 1.13 0.80Enrolled: T(m) 1.47 0.07 1.22 0.24Enrolled: T(l) 0.97 0.91 1.52 0.01

(controlled for baseline variables)

Page 107: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

ResultsActual vs. hypothetical population

Actual: Hypothetical:Education (2005) OR p-value OR p-value

Never enrolled 0.84 0.15 0.36 < .0001Prev enrolled 0.81 0.12 0.46 < .0001Tert(s)/voc 0.92 0.40 0.51 < .0001Tert(m) 1.32 0.001 0.65 < .0001Tert(l) (REF) 1 − 1 −Enrolled: T(s)/voc 2.54 0.02 1.13 0.80Enrolled: T(m) 1.47 0.07 1.22 0.24Enrolled: T(l) 0.97 0.91 1.52 0.01

(controlled for baseline variables)

Page 108: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

ResultsActual vs. hypothetical population

Actual: Hypothetical:Education (2005) OR p-value OR p-value

Never enrolled 0.84 0.15 0.36 < .0001Prev enrolled 0.81 0.12 0.46 < .0001Tert(s)/voc 0.92 0.40 0.51 < .0001Tert(m) 1.32 0.001 0.65 < .0001Tert(l) (REF) 1 − 1 −Enrolled: T(s)/voc 2.54 0.02 1.13 0.80Enrolled: T(m) 1.47 0.07 1.22 0.24Enrolled: T(l) 0.97 0.91 1.52 0.01

(controlled for baseline variables)

Page 109: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Assumptionsare they possibly violated?

I 4 main assumptions:

1. Exchangeability (no unmeasured confounders) Ye

∐E | Z

baseline cov

2. No mis-specification of the models for the weights Weight Model

3. Consistency4. Positivity

I All of these are important → all subject to possible violations

I Not testable...

PhD Defense, September 2009

Page 110: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Assumptionsare they possibly violated?

I 4 main assumptions:

1. Exchangeability (no unmeasured confounders) Ye

∐E | Z

baseline cov

2. No mis-specification of the models for the weights Weight Model

3. Consistency4. Positivity

I All of these are important → all subject to possible violations

I Not testable...

PhD Defense, September 2009

Page 111: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Assumptionsare they possibly violated?

I 4 main assumptions:

1. Exchangeability (no unmeasured confounders) Ye

∐E | Z

baseline cov

2. No mis-specification of the models for the weights Weight Model

3. Consistency4. Positivity

I All of these are important → all subject to possible violations

I Not testable...

PhD Defense, September 2009

Page 112: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Assumptionsare they possibly violated?

I 4 main assumptions:

1. Exchangeability (no unmeasured confounders) Ye

∐E | Z

baseline cov

2. No mis-specification of the models for the weights Weight Model

3. Consistency4. Positivity

I All of these are important → all subject to possible violations

I Not testable...

PhD Defense, September 2009

Page 113: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Assumptionsare they possibly violated?

I 4 main assumptions:

1. Exchangeability (no unmeasured confounders) Ye

∐E | Z

baseline cov

2. No mis-specification of the models for the weights Weight Model

3. Consistency4. Positivity

I All of these are important → all subject to possible violations

I Not testable...

PhD Defense, September 2009

Page 114: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Assumptionsare they possibly violated?

I 4 main assumptions:

1. Exchangeability (no unmeasured confounders) Ye

∐E | Z

baseline cov

2. No mis-specification of the models for the weights Weight Model

3. Consistency4. Positivity

I All of these are important → all subject to possible violations

I Not testable...

PhD Defense, September 2009

Page 115: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Assumptionsare they possibly violated?

I 4 main assumptions:

1. Exchangeability (no unmeasured confounders) Ye

∐E | Z

baseline cov

2. No mis-specification of the models for the weights Weight Model

3. Consistency4. Positivity

I All of these are important → all subject to possible violations

I Not testable...

PhD Defense, September 2009

Page 116: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

ConclusionUltimate fertility in Denmark

I There are educational differences in ultimate fertility

I There is some tendency that women who become motherswhile enrolled in education are more likely to drop out andwomen who become mothers early are less likely to have auniversity degree, more likely to have no further degree or aT(m)

I This might play a role in the relationship between ultimatefertility and educational attainment

PhD Defense, September 2009

Page 117: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

ConclusionUltimate fertility in Denmark

I There are educational differences in ultimate fertility

I There is some tendency that women who become motherswhile enrolled in education are more likely to drop out andwomen who become mothers early are less likely to have auniversity degree, more likely to have no further degree or aT(m)

I This might play a role in the relationship between ultimatefertility and educational attainment

PhD Defense, September 2009

Page 118: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

ConclusionUltimate fertility in Denmark

I There are educational differences in ultimate fertility

I There is some tendency that women who become motherswhile enrolled in education are more likely to drop out andwomen who become mothers early are less likely to have auniversity degree, more likely to have no further degree or aT(m)

I This might play a role in the relationship between ultimatefertility and educational attainment

PhD Defense, September 2009

Page 119: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Thank you very much!

PhD Defense, September 2009

Page 120: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

References

Becker, G. S. (1991).A Treatise on the Family.Harvard University Press, Cambridge, Massachussetts.

Hernan, M. A., Brumback, B., and Robins, J. M. (2001).Marginal Structural Models to Estimate the Joint CausalEffect of Nonrandomized Treatment.Journal of the American Statistical Association, 96:440–448.

PhD Defense, September 2009

Page 121: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Assumptions for MSMsExchangeability

Need to include enough baseline covariates, Z, s.t. within eachsubgroup defined by these, the women are exchangeable:

In this study, Z includes:

1. mark attained at the PUSE

2. number of mother-siblings

3. grandmother’s age at first birth

4. grandmother’s and grandfather’s educational attainment

What’s missing? Health, men, others?

exchange

PhD Defense, September 2009

Page 122: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Assumptions for MSMsExchangeability

Need to include enough baseline covariates, Z, s.t. within eachsubgroup defined by these, the women are exchangeable:

In this study, Z includes:

1. mark attained at the PUSE

2. number of mother-siblings

3. grandmother’s age at first birth

4. grandmother’s and grandfather’s educational attainment

What’s missing? Health, men, others?

exchange

PhD Defense, September 2009

Page 123: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Parity transitions in Norway Ultimate fertility in Denmark

Assumptions for MSMsExchangeability

Need to include enough baseline covariates, Z, s.t. within eachsubgroup defined by these, the women are exchangeable:

In this study, Z includes:

1. mark attained at the PUSE

2. number of mother-siblings

3. grandmother’s age at first birth

4. grandmother’s and grandfather’s educational attainment

What’s missing? Health, men, others?

exchange

PhD Defense, September 2009

Page 124: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Model for the weights

Models

log

[πikt

πi1t

]= αk + βk · Ait + γk · Zi , k = 2, . . . , 8

log

[πikt

πi1t

]= αk + βk · Ait + γk · Zi + δk1 · Bi ,t−1 + δk2 · Bi ,t−2

where πikt = Pr(Ei (t) = k)

Categories (k):

1. not enrolled, several subcategories

2. enrolled, several subcategories

exchange

Page 125: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Final educational attainments

Ever enrolled in university

Edu(2005) PUSE Voc/T(s) T(m) T(l) Total

Everyone 14.6% 24.8% 44.1% 16.6% 8032Int/Change 17.6% 10.1% 44.1% 28.2% 1007

Ever enrolled in university→ child(yes/no)

Edu(2005) PUSE Voc/T(s) T(m) T(l) Total

Child 24.8% 5.7% 51.4% 18.1% 105No child 16.7% 10.6% 43.2% 29.4% 902

Int/Change 17.6% 10.1% 44.1% 28.2% 1007

Go Back!

Page 126: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Final educational attainments

Ever enrolled in university

Edu(2005) PUSE Voc/T(s) T(m) T(l) Total

Everyone 14.6% 24.8% 44.1% 16.6% 8032Int/Change 17.6% 10.1% 44.1% 28.2% 1007

Ever enrolled in university→ child(yes/no)

Edu(2005) PUSE Voc/T(s) T(m) T(l) Total

Child 24.8% 5.7% 51.4% 18.1% 105No child 16.7% 10.6% 43.2% 29.4% 902

Int/Change 17.6% 10.1% 44.1% 28.2% 1007

Go Back!

Page 127: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Final educational attainments

Ever enrolled in university

Edu(2005) PUSE Voc/T(s) T(m) T(l) Total

Everyone 14.6% 24.8% 44.1% 16.6% 8032Int/Change 17.6% 10.1% 44.1% 28.2% 1007

Ever enrolled in university→ child(yes/no)

Edu(2005) PUSE Voc/T(s) T(m) T(l) Total

Child 24.8% 5.7% 51.4% 18.1% 105No child 16.7% 10.6% 43.2% 29.4% 902

Int/Change 17.6% 10.1% 44.1% 28.2% 1007

Go Back!

Page 128: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Final educational attainments

Ever enrolled in university

Edu(2005) PUSE Voc/T(s) T(m) T(l) Total

Everyone 14.6% 24.8% 44.1% 16.6% 8032Int/Change 17.6% 10.1% 44.1% 28.2% 1007

Ever enrolled in university→ child(yes/no)

Edu(2005) PUSE Voc/T(s) T(m) T(l) Total

Child 24.8% 5.7% 51.4% 18.1% 105No child 16.7% 10.6% 43.2% 29.4% 902

Int/Change 17.6% 10.1% 44.1% 28.2% 1007

Go Back!

Page 129: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Final educational attainments

Ever enrolled in university

Edu(2005) PUSE Voc/T(s) T(m) T(l) Total

Everyone 14.6% 24.8% 44.1% 16.6% 8032Int/Change 17.6% 10.1% 44.1% 28.2% 1007

Ever enrolled in university→ child(yes/no)

Edu(2005) PUSE Voc/T(s) T(m) T(l) Total

Child 24.8% 5.7% 51.4% 18.1% 105No child 16.7% 10.6% 43.2% 29.4% 902

Int/Change 17.6% 10.1% 44.1% 28.2% 1007

Go Back!

Page 130: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Mothers by Jan 1st 1986 by education (2005)

Mother PUSE Voc T(s) T(m) T(l) Total

No 14.4% 16.2% 8.5% 44.0% 17.0% 7677Yes 18.0% 21.4% 6.5% 47.6% 6.5% 355

Total 14.6% 16.4% 8.4% 44.1% 16.6% 8032

Go Back!

Page 131: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Mothers by Jan 1st 1986 by education (2005)

Mother PUSE Voc T(s) T(m) T(l) Total

No 14.4% 16.2% 8.5% 44.0% 17.0% 7677Yes 18.0% 21.4% 6.5% 47.6% 6.5% 355

Total 14.6% 16.4% 8.4% 44.1% 16.6% 8032

Go Back!

Page 132: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Mothers by Jan 1st 1986 by education (2005)

Mother PUSE Voc T(s) T(m) T(l) Total

No 14.4% 16.2% 8.5% 44.0% 17.0% 7677Yes 18.0% 21.4% 6.5% 47.6% 6.5% 355

Total 14.6% 16.4% 8.4% 44.1% 16.6% 8032

Go Back!

Page 133: Socioeconomic Determinants for Fertility - DEMOGRAFI · How to measure fertility? I Fertility -actualbirths I Fertility is a process that evolves over several years (in principle,

Is feedback present in the study population?Go Back!

Birth this year

log

odds

Up2 (voc) Tert(s) Tert(m) Tert(l)

Inte

rcep

t−4

−3−2

BirthNo birth

β1 = 1.27

β2 = 0.34

β3 = 0.39

β4 = 0.3