Evidence from Multivariate Analyses

5.1 Methodology

5.1.1 The Generic Model

The estimation results discussed in the subsequent part of the paper employ a generic model of the form

(1) y = α₁n + Xα₂ + ε

The dependent variable of interest y is a function of the number of children n and a host of other individual, household and often times also community variables X. The parameters to be estimated are α and β and ε the error terms assumed to have the usual convenient properties. The implied subscripts are omitted for clarity. The essential characteristic of this generic model is that n is endogenous and explained by a function, so

(2) n = β₁z + Xβ₃ + μ
(3) μ = ρε + η

The basic motivations for an endogenous n are the quantity-quality hypothesis (Becker and Lewis 1973), and that children are a form of old-age security (Neher 1971). The quantity-quality hypothesis argues that there is a trade-off between the number and the quality (usually expressed in terms of human capital investments) of children, i.e., the number of children is chosen with a given quality in the parent’s mind. The variable z is often called the instruments to identify n in the y equation. The error term µ is then correlated with ε as in (3). Given (2), if y is estimated by OLS or some LDV estimation techniques if the dependent variable of interest is discrete the estimate would be biased. One needs to use instrumental variable (IV) estimation of two-stage LDV estimation techniques to generate consistent estimates. The problem is that it is not easy to find an appropriate instrument z for n that is not included in X. This is problem we turn to in the next section.

To provide estimates for the responses of the different socioeconomic classes, the number of children variable was interacted with per capita income quintiles.

5.1.2 Balanced Sex-Mix as an Instrument

There are not too many instruments that one can find for the number children in household models. Most of the likely candidates such the household income, education of the parents or age of marriage are also related to the dependent variable of interest such as labor force participation of parents, and savings or education of children, rendering these inappropriate as instruments. Recent research using US data such as Angrist and Evans (1998) has used the hypothesis that families prefer to have a balanced sex mix of children as an instrument for the number of children. The Philippines is one of the countries in Asia where a balanced sex-mix is found to have prevailed in contrast to countries in South and Eastern Asia where indications for son preference is often found (Wongboonsin and Ruffolo, 1995). Early literature that confirms a preference for a balanced sex-mix in the Philippines is found in Stinner and Mader (1975). The other instruments that are available are limited by their applicability only in very specific circumstances. The occurrence of twins have been also been used as instruments again using US data first in Rosenzweig and Wolpin (1980) and in subsequent studies such as Angrist and Evan (1998). A much more recent application was for Romania (Glick, Marini and Sahn, 2005). Son-preference in the Republic of Korea was also used as an instrument for the fertility for instance in Lee (2004). Finally, another instrument would be an exogenous policy change that could affect child bearing. Quian (2004), for instance, used the relaxation of the one-child policy in the People’s Republic of China that allows rural households to have another child if the first child is a girl. Viitanen (2003), on the other hand, used the large-scale giving out of vouchers for privately provided childcare in Finland.

In the case of the balanced sex-mix hypothesis, the fact that families do not have control over the sex of their children makes same sex for the first two children virtually a random assignment. As argued in Angrist and Evans (1998) using same sex as an instrument will allow a causal interpretation. It should be noted, however, that the downside of this instrument is that it will render families that has less than two children unusable for analysis. While this maybe a serious problem in low fertility areas, this may not be in the case of the Philippines where the average number of children exceeds four.

To check on the validity of this instrument, Table 7 [ PDF 49.1KB | 1 page ] provides a cross tabulation of the average proportion of families that have additional children and the average number of number of children by sex of their first two children for 24,000 families that have two or more children using the APIS 2002 dataset. The table shows that 67.4% of families that had one male and one female for their first two children had another child, while 71.8% had another child when they have the same sex for their first two children, or a difference of more than 4%. In terms of average number of children, this is 3.49 as against 3.61 or an average difference of a little over 0.12 children. These average differences are statistically significant under conventional levels. Comparing this with Table 3 and Table 5 in Angrist and Evans (1998) one can observe several differences. The difference in the proportion of families having a third child for the two groups of families is smaller and the standard error is larger. In the case of the difference in the average number of children, the difference is larger but so is the standard error. This is not unexpected given the larger family size in the Philippines compared to the US and the expected larger dispersion of the distribution. Consequently, the implied t statistics in Table 7 [ PDF 49.1KB | 1 page ] are not as large as those in Angrist and Evans (1998) indicating that discrimination generated from the same-sex instrument may not be as strong as that obtained using US data.

5.1.3 Data Sources

The data on most individual and household characteristics and location characteristics were taken from the 2002 Annual Poverty Indicator Survey (APIS). The APIS is a rider survey to the July round of the quarterly Labor Force Survey conducted by the National Statistics Office (NSO). The 2002 APIS is the third of the series conducted by the NSO. The other two were conducted in 1998 and 1999. It provides basic demographic information on all members of the household as well as household amenities. Income and expenditure data for the past 6 months are also gathered.

All monetary values such as income and savings are deflated using provincial consumer price indices compiled by the Price Division of the NSO. This is done to control for interprovincial price variability.

Barangay and municipal-level data from the 2000 Census of Population and Housing are also used to provide measures of investment opportunities, availability of financial institutions and school facilities. It is therefore assumed that there is not much difference in the structure of distribution of the facilities in 2000 and in 2002 or that whatever changes happened did not upset the distribution of the availability of facilities. These barangay and municipal data set were aggregated at the domain level of the APIS and attached to the APIS data set using domain identification variables.

5.2 Number of Children and their Education

The impact of additional children on their education was estimated by using the proportion of school-age children 6 to 24 years old to the number of children in the household. Estimates for the different age groups corresponding to the three education levels, elementary (6-12), secondary (13-16) and tertiary (17-24), were also done to provide indications of the differential impacts.

The estimate given in Orbeta (2005a) shows that each additional child reduces the proportion of school-age children attending school. The estimated impact of each additional child for the total school-age population of 6-24 is -19% of current attendance rates (Table 8 [ PDF 60.1KB | 1 page ]). The impact for the elementary age group is not significant. The estimated impact for the secondary and tertiary levels are, respectively, -26% and -57%. By socioeconomic class, the impact exhibits a regressive effect with a larger impact for poorer households. For instance, for the 6-24 age group, the impact is -24% for the poorest quintile and this is -16% for the quintile. In the secondary age group, this is – 29% for the poorest quintile and -17% for the richest quintile. Finally, for the tertiary age group, this is -77% for the poorest age group and -22% for the richest quintile.

The preceding discussion highlights several important conclusions. One, the impact of additional children on school attendance is negative. Two, the impact is regressive with bigger negative impacts on poorer households relative to richer households. Three, the regressiveness intensifies as one goes up the levels of the education ladder.

5.3 Children and the Labor Supply and Wage Income of Parents

The impact of the additional children on the labor supply of parents and their wage income is estimated. A distinction is drawn between all types of work and paid work for the mothers. For fathers this distinction is not made.

The estimates, given in Orbeta (2005b), show that labor force participation of the mother declines by –1.68% per additional child (Table 9 [ PDF 66.7KB | 1 page ]). This effect rises to –2.13% when one considers only paid work. Another noteworthy result is that the presence of children below the normal school age of 6 years results in a -7.2% decline for all types of work and -5.7% for paid work of mothers. The estimates for fathers show insignificant results.

The estimates using the interaction between the number of children and the per capita income quintile show that the impact for mothers in the bottom quintile is higher than the average; –2.12% of all types of work and -5.68% for paid work. The impact for mothers in the higher income groups interestingly becomes smaller negative for the lower middle and middle-income quintiles and turns positive for the top two classes. This positive effect for higher income groups may mean that mothers are not affected by the presence of children. This may mean that for richer income groups the families are perhaps able to pay for child care and still contribute to household income.

In the case of the fathers, while the average effect is not significant, the not significant effect is only found in the poorest income class. From the lower middle up to the richest income class, the impact is positive although not as large as the one obtained for mothers. This may be explained by the already high labor force participation rate. It would have been interesting to see the impact on labor hours, but unfortunately, the data does not contain information on labor hours.

Turning to the impact on wage income, each additional child is estimated to reduce mothers’ average earnings by 1,010 pesos (deflated with 1994=100) (Table 10 [ PDF 60.4KB | 1 page ]). This represents about a 5% decline from a six-month8 average wage income of 20,200. The impact on the wage income of fathers is 233 (deflated with 1994 =100). This is about 1.1% of the six-month average wage income of 21,900.

The impact across income class shows that the negative impact on the earnings of mothers is for the bottom two quintiles. This is -13% for the poorest quintile and -7% for the lower middle quintile. The impact of the higher income quintile is positive at 2%, 15% and 33% for the middle, upper middle and the richest quintile, respectively. In the case of fathers, the positive impact is only for the top four quintiles as the impact of the poorest quintile is still negative (-6%). The impact for the higher income classes are 5%, 12%, 19% and 35% for the lower middle, middle, upper middle and richest quintile, respectively.

The foregoing discussions can be summarized in the following conclusions. One, the impact of additional children on the labor force participation of mothers is negative, on the average, with a higher impact for wage employment compared to all types of work while for the fathers this is insignificant. Two, the impact on labor force participation of additional children is regressive with a negative impact on the poorer households and a positive one for the top two quintiles. Three, the impact of additional children on wage incomes substantially echoes the impact on labor force participation with the addition that the impact on fathers, although much more subdued than that for mothers, is positive and significant. It should also be pointed out that while the average impact on the wage income of fathers is small and positive, for the poorest quintile it is still negative.

5.4 Children and Household Savings

The impact of children on saving were estimated using two measures of savings, namely: (a) the average savings rates – the ratio of savings to disposable income; and (b) savings levels. There are, in turn, two savings definitions of savings used: (i) income minus expenditures (definition 1) and, (ii) (i) with expenditure on durable furniture, education and health which have benefits over the longer term added back (definition 2).

The estimates in Orbeta (2005c) showed that each additional child will cause an average reduction in savings rates of about -0.36% for definition 1 and insignificant for definition 2 (Table 11 [ PDF 66.2KB | 1 page ]). While this number may look small in absolute value, it is substantial when measured relative to the recorded average savings rates. Given the average savings rate in the sample of 0.028 (definition 1) this estimate represents a reduction of about - 13%.

The impact across income classes shows that the negative impact is only for the bottom per capita income quintile. In addition, the negative impact is larger for the poorest quintile at about –3% for both definitions; in terms of proportion to the recorded savings rate it is -14% for definition 1 and –18% for definition 2. For the rest of the income classes the impact is positive indicating that children increase the household savings rates and increasing as one goes up the income classes. The pattern of the percentage change is declining, because the rates of savings rise faster with the income classes.

Turning to the total household savings, each additional child is expected to cause a reduction of -254 (definition 1) or -309 (definition 2) in 1994⁹ pesos. This would mean a - 3.3% and -2.7% reduction, respectively, with a recorded average savings levels of 7,742 and 10,854 under the two savings definitions.

The impact across income classes shows that it is negative for all except for the poorest quintile where it is not significant. The impact for the lower middle quintile is –594 and this rises to –9,114 for the richest quintile per additional child for definition 1. A similar pattern is seen for the impact using definition two although at slightly smaller magnitudes. Again in percentage terms the negative impact declines because of the higher levels of savings as one goes up the income classes.

These results highlight the regressive impacts additional children have on the savings rates and savings levels of households that can be summarized in two statements. One, the impact on the savings rates of the bottom quintile is negative. Two, the impact on savings levels is negative and in percentage terms is bigger among lower income households.

Download this Discussion Paper [ PDF 187.8KB| 24 pages ].

Comment(s)

There are [5] comment(s) for this entry. Post a comment.

1. Edgar
   (posted 01 August 2007 / 03:06:59 AM)
   
   I haven't really exhaustively read the discussion paper, just breezed through, so I might sound unfair, but I am now old enough to be highly distrustful of kneejerk reactions and conventional wisdom that suggest population is largely causative of family poverty. Has there not been enough research that large scale asset reform in highly skewed income distributive countries such as the Philippines is needed to industrialize a population and as a side effect put more females in the smokestack workforce? And this industrialization will lead to more families involuntarily spacing births? Come on, address land ownership first and produce indigenous jobs, rather than depending on migrant employment to create a surplus. Be more critical of theories that population is a major contributory factor to poverty.
2. krisjoy
   (posted 22 March 2007 / 03:27:52 PM)
   
   It's not enough to just be aware that we are poor. Since the government posted ambitious poverty reduction targets, then the average 'juan' will continue to think rice is special for more years. I think what we need is to work as a nation, achieving realistic marks for developments in agriculture, land reform, population planning, security issues, among others. I am young and I am earning; however, I feel very poor seeing that many of my fellowmen do not eat rice and that the tax I religiously pay does not get to projects.
3. MARIONNE
   (posted 06 March 2007 / 10:38:21 AM)
   
   poverty is a very hard opponent. if we want to have even a wee bit of development in our poor country, the government must see to it that poverty is lessened.
   tnx...
4. Junior
   (posted 01 March 2007 / 09:05:08 PM)
   
   In our country corruption is the number one cause of poverty thanks.
5. john
   (posted 11 July 2006 / 05:43:07 PM)
   
   Yes! It is evident here in the philippines
   because, as we can see now it is our problem ....thats all.