The data

Time Bomb or Damp Squib?  Fertility in Contemporary Northern Ireland

Patrick McGregor Patricia McKee

“Predicting Short Run Changes in Fertility in Northern Ireland”

a project funded by

The authors are grateful to the NILS core team and NILS RSU at NISRA for their assistance

NUI Galway, 29 October, 2010

2

Time Bomb or Damp Squib?  Fertility in Contemporary Northern Ireland

• Overview• Religion and fertility• The NILS• The data• The statistical model• Results• Conclusion

Overview 3

"Don't worry, Gerry, your numbers are getting better all the time!“

President Bill Clinton to Gerry Adams, 1998

The data 4

Age Specific Fertility Rates

0

25

50

75

100

125

150

175

200

1977 1982 1987 1992 1997 2002 2007

Age

-spe

cific

ferti

lity

rate

15-19 20-24 25-29 30-34 35-39 40-44

Total Period Fertility Rates

1.51.71.92.12.32.52.72.93.13.33.5

1970 1975 1980 1985 1990 1995 2000 2005 2010

TP

FR

Ireland United Kingdom United States

OECD-30 N Ireland

Religion and Fertility 5

Religion and Fertility

Static economic analysis

Max ST

Children assumed to be normal ‘goods’ so: but possibly

The home production framework allows the cost of children to be expressed as a function of the parents’ wages and their respective shares in the costs of producing child quality

Influence of religion then is through preferences

sqnUU ,,sc snqI

0 qn 0n

Religion and Fertility 6

Demographic approach Given that Catholic doctrine is pro-natalist, a Catholic would be anticipated

to rank quantity more highly than quality thus leading to higher fertility relative to a member of the mainstream Protestant religions where the fertility is considered a matter of individual choice (Lehrer (1996)).

This has been characterised as the ‘particularized theology’ approach by Goldscheider (1971). Change then is essentially driven by doctrine and doctrinal change has been extremely limited in the Catholic Church.

Three elements required to establish a credible role for religion in determining fertility (McQuillan (2004)):

• it must articulate norms that have linkages to fertility outcomes• the religious group must be capable of communicating its teachings

and to enforce compliance• members must feel a strong sense of attachment to the religious

community

An inevitable conclusion from this is that if religion in contemporary Northern Ireland is to have a role it is as a social category rather than a doctrinal one.

Religion and Fertility 7

Identity

Akerlof and Kranton (2000), “Economics and Identity”, QJE

A person’s identity depends:

• on the social categories they are assigned

• the extent to which their characteristics match the ideal of these

• the extent their actions correspond to the associated prescribed behaviour

Developed by Manning and Roy (2010) "Culture Clash or Culture Club? National Identity in Britain", EJ

Religion and Fertility 8

Developed by Manning and Roy (2010) "Culture Clash or Culture Club? National Identity in Britain", EJ

Let: f* be a woman’s desired fertility based on a purely private economic assessment f-C be the previous cohort’s actual fertility when at a similar age x be a local characteristic that is linked to prescribed behaviour

f~

, that minimizes the loss function: The woman chooses the fertility,

22

2

1*

2

1CffxgfffL

xg

fxgff C

1

*~

21

*~

g

ffg

x

f C

The NILS 9

The NILS• The NILS potential mothers:

those women with health card registrations, aged 16-44 years and whose DOB is one of the 104 in the systematic sample

• Registrations downloaded biannually and constitute potential panel members

• Details of any birth to a NILS mother are forwarded by the GRO to the NILS

• 2001 Census: An attempt is made to link the Census details of all NILS mothers

The result is a panel of 124,107 women for 1997 – 2007 in which there are 988,194 observations

The data 10

Average Protestant Parity by Age & Cohort

0.0

0.5

1.0

1.5

2.0

2.5

3.0

15 20 25 30 35 40 45

Age

Pa

rity

P53-57

P58-62

P63-67

P68-72

P73-77

P78-82

P83-87

P88-92

Average Catholic Parity by Age & Cohort

0.0

0.5

1.0

1.5

2.0

2.5

3.0

15 20 25 30 35 40 45

Age

Pa

rity

C53-57

C58-62

C63-67

C68-72

C73-77

C78-82

C83-87

C88-92

The data 11

Average Parity of Women aged 16-44

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

year

aver

age

parit

y

Protestant

Catholic

Difference in Average Parity by Age and Cohort

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

15 20 25 30 35 40 45

Age

Dif

fere

nce

in

Par

ity

(C -

P) 53-57

58-62

63-67

68-72

73-77

78-82

83-87

88-92

The data 12

POPNOQUAL % population with no/low levels of qualificationsPOPDEG % population with educational level 4 or 5

SOCDEPER % persons aged >15 in social grades DESOCDEHH % households in social grades DERENTED % households rentedMEDAGE median age of population in the areaFLOOKFAM % females 16-74 economically inactive and looking after home/family

POPSINGLE % persons aged 16 and over: single (never married)HHMARDEP % married households with dependent childrenHHCOHABDEP % households cohabiting with dependent childrenHHLONEDEP % households lone parent with dependent children

Census variables at Super Output Area

The data 13

RPC1 RPC2 RPC3 RPC4

POPNOQUAL 0.514 0.024 -0.066 -0.044

POPDEG -0.561 0.099 0.128 -0.014

SOCDEPER 0.278 0.309 0.041 0.029

SOCDEHH 0.295 0.302 0.027 0.018

RENTED 0.088 0.422 0.150 -0.023

MEDAGE 0.000 0.182 -0.708 -0.117

FLOOKFAM 0.406 -0.142 0.288 -0.159

POPSINGLE -0.216 0.377 0.499 -0.209

HHMARDEP 0.088 -0.636 0.240 -0.051

HHCOHABDEP -0.016 0.004 0.020 0.922

HHLONEDEP 0.171 0.157 0.252 0.247

The loadings of the census variables on the rotated components

The data 14

0

100

200

300

400

500

600

1997 1999 2001 2003 2005 2007

Year

Gro

ss

We

ek

ly p

ay

(£

)

63.0

64.0

65.0

66.0

67.0

68.0

Act

ivit

y R

ate

Q1 median Q3 Activity Rate

Women in the Labour Market

Source: DETINI

The Statistical Model 15

The Statistical Model

• Raftery, AE, Lewis, SM and Aghajanian, A (1995).• Demand or Ideation? Evidence from the Iranian

Marital Fertility Decline, Demography, vol. 32.• Data: 1977 Iran Fertility Survey• “ each woman-year of exposure is treated as a

separate case” • Five clocks:

– Age Period Cohort Parity Duration

The Statistical Model 16

The Logit Model

iiii

iiii

XLOCALITY

ECONOMICPERIODCOHORTCDEMOGRAPHIB

/

5

4321*

0*0

0*1

ii

ii

BifB

BifB

Let B*I be the propensity for the ith woman to have a baby

Estimation results are on the handout

Results 17

Logit Regression Tests

• CoefficientsProtestant = CoefficientsCatholic LRT 673.54 [0.000]

• CATHOLIC*DEMOGRAPHIC = 0 Wald 247.14 [0.000]

• CATHOLIC*PERIOD = CATHOLIC*ECONOMIC = CATHOLIC*LOCALITY = 0 Wald 24.53 [0.177]

112

442

192

Results 18

Marginal Effects

0,1ˆ1,1ˆ, CATHOLICXBPCATHOLICXBPME iiiiiCATHOLIC

Where is the vector of values of all the explanatory variables except for CATHOLIC

iX

For ith woman:

The marginal effect is taken as the mean of individual marginal effects over intervals of particular variables

PERCATH

XBPME

ii

iPERCATH

1ˆ

,

Results 19

MECATHOLIC and Locality

0.00%

1.00%

2.00%

3.00%

4.00%

5.00%

6.00%

7.00%

8.00%

9.00%

-10 -5 0 5 10RPC!

%

cath%

prod%MEcatholic by RPC1

-0.0040

-0.0020

0.0000

0.0020

0.0040

0.0060

0.0080

0.0100

0.0120

0.0140

-8 -6 -4 -2 0 2 4 6 8

RPC 1

cath

prod

Results 20

MECATHOLIC and Locality

MEcatholic by RPC2

-0.0050

-0.0030

-0.0010

0.0010

0.0030

0.0050

0.0070

0.0090

0.0110

0.0130

0.0150

-8 -6 -4 -2 0 2 4 6 8

RPC2

cath

prod

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

-5 -3 -1 1 3 5

RPC2

cath%

prod%

Results 21

MECATHOLIC and Locality

MEcatholic by PERCATH

0.002

0.003

0.004

0.005

0.006

0.007

0 0.2 0.4 0.6 0.8 1

PERCATH

cath

prod

Cumulative ogive

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 1

PERCATH

cu

m %

cumcath%

cumprod%

Results 22

MECATHOLIC and Locality

MEcatholic

0

0.001

0.002

0.003

0.004

0.005

0.006

0.007

0.008

1996 1998 2000 2002 2004 2006 2008

Results 23

MECATHOLIC and Locality

MEcatholic by Age and Years

-0.0150

-0.0100

-0.0050

0.0000

0.0050

0.0100

0.0150

0.0200

0.0250

15 20 25 30 35 40 45

age

ME

cath

oli

c

97-99

overall

05- 07

Results 24

MECATHOLIC and Locality

MEpercath by PERCATH

0.0029

0.0030

0.0031

0.0032

0.0033

0.0034

0.0035

0 0.2 0.4 0.6 0.8 1PERCATH

cath

prod

Total

Conclusion 25

Conclusions• Protestant fertility is considerably more stable than Catholic in the

period 1997 – 2007. • Both communities responded in the same fashion to economic and

period effects• There is an identity dimension to Catholic adjustment but this is not

large • There is a small social rather than economic effect to the adjustment

of the Catholic fertility• Catholic adjustment seems substantially complete – two

communities with similar but distinct demographic profiles?