Skip to main content

Table 2 Non-linear decomposition of the difference in facility-based delivery between wealth groups

From: Utilisation of health services and the poor: deconstructing wealth-based differences in facility-based delivery in the Philippines

\( {\overline{Y}}_{NP} \)

0.746

       

\( {\overline{Y}}_P \)

0.411

       

Wealth gap \( \left({\overline{Y}}_{NP}-{\overline{Y}}_P\right) \)

0.336

       

Benchmark Coefficients

Non-Poor

Poor

 

Decomp.

 

S.E.

Cont. (%)

Decomp.

 

S.E.

Cont. (%)

Supply-side

        

 Distance to facility

0.014

**

(0.006)

4.06

0.011

***

(0.004)

3.42

Socioeconomic factors

        

 Mother’s education

0.017

 

(0.015)

4.93

0.026

***

(0.007)

7.79

 Partner’s education

0.025

*

(0.015)

7.49

0.031

***

(0.009)

9.14

 Mother watches TV

0.011

 

(0.011)

3.20

0.028

***

(0.006)

8.48

 Mother’s employment

0.028

***

(0.009)

8.22

0.020

***

(0.006)

5.87

 Partner’s employment

0.065

***

(0.013)

19.29

0.037

***

(0.010)

11.09

Maternal factors

        

 Multiple birth

−0.0005

 

(0.0005)

−0.14

−0.001

 

(0.000)

−0.16

 Previously term. preg.

0.0003

 

(0.001)

0.10

0.001

 

(0.001)

0.30

 Child’s birth order

0.058

***

(0.008)

17.26

0.042

***

(0.006)

12.59

 Mother’s age at birth

−0.003

 

(0.001)

−0.82

−0.001

*

(0.001)

−0.38

Sociocultural factors

        

 Religion

0.007

 

(0.005)

2.19

0.014

***

(0.004)

4.18

 Marital status

−0.007

**

(0.003)

−2.00

0.003

 

(0.004)

0.79

Geography

        

 Regions

0.020

**

(0.010)

6.01

0.013

 

(0.017)

3.95

Total explained

0.234

  

69.78

0.225

  

67.12

Total unexplained

0.101

  

30.22

0.110

  

32.88

  1. Notes: Total number of observations is 7,121. Specifications include a dummy for missing partner information. The contribution of each variable to the regional gap is given under the column “Decomp.” computed as the mean values of the decomposition using 100 random non-poor income sub-samples with the order of the variables in each replication randomised. Standard errors are given in parentheses and (*), (**) and (***) represent statistical significance at 10 %, 5 % and 1 % levels, respectively. Decomp., decomposition; S.E., standard error; Cont., contribution; term. preg., terminated pregnancy