Sei sulla pagina 1di 23

The Journal of Development Studies, 2016

Vol. 52, No. 6, 854–875, http://dx.doi.org/10.1080/00220388.2015.1113261

Heterogeneity in the Intrahousehold Allocation of


International Remittances: Evidence from
Philippine Households
MARJORIE PAJARON
School of Economics, University of the Philippines, Quezon City, Philippines

(Final version received September 2015; final version accepted October 2015)

ABSTRACT This article shows that heterogeneity in the intrahousehold allocation of remittances sent by
temporary international migrant workers exists among households in the Philippines, and may be attributable
to the absence of the spouse, suggesting moral hazard, and to the gender and educational attainment of the
household head, indicating differences in preferences. GLM estimates and post-estimation tests reveal that male
heads spend more on alcohol and tobacco, that the presence of a female decision-maker in a household increases
allocations to goods that improve the welfare of the children, and that household heads with less formal education
allocate more to education.

1. Introduction
Recent studies on resource allocation within households have shown that individuals’ control and
management of resources affect intrahousehold allocation outcomes. In particular, the higher the
relative resources controlled by women, the higher the expenditures on food and children’s clothing
and education, and the lower the expenditures on alcohol and cigarettes (Hoddinott & Haddad, 1995;
Quisumbing & de la Brière, 2000; Quisumbing & Maluccio, 2003; Rubalcava, Teruel, & Thomas,
2004). In addition, resources in the hands of women improve the health status of children and have
greater effects on the family’s health (Duflo, 2003; Thomas, 1990).1
These research findings have important policy implications, and they have affected the design of
some public transfer programmes. For example, based on such research, the government of Mexico
has provided cash and in-kind benefits directly to mothers of poor households in rural areas through
the Programa Nacional de Educación, Salud y Alimentación (PROGRESA) in an attempt to alleviate
future levels of poverty by encouraging investment in education, health and nutrition (Adato et al.,
2000; Skoufias & McClafferty, 2001).
Heterogeneity in the intrahousehold allocation of resources, however, may be more complex than
implied by a simple male–female dichotomy, because subgroups of male and female household heads
differ in their access to and allocation of resources. For example, in Jamaica, single female heads
allocate more to adult clothing and less to alcohol and tobacco compared to female heads whose
husbands are present (Handa, 1996). In rural Ecuador, children of widowed heads have lower school
enrolment than children of married female heads (DeGraff & Bilsborrow, 1993). In addition, children

Correspondence Address: Marjorie Pajaron, School of Economics, Univeristy of the Philippines, Diliman Quezon City,
Philippines. Email: Mpajaron@econ.upd.edu.ph
An Online Appendix is available for this article which can be accessed via the online version of this journal available at http://dx.
doi.org/10.1080/00220388.2015.1113261

© 2016 Informa UK Limited, trading as Taylor & Francis Group


Heterogeneity in the intrahousehold allocation of international remittances 855

of widowed heads in Bangladesh are more likely to work outside the home (Joshi, 2004). Single male
heads in Jamaica have children with a higher degree of risk (with significantly lower school enrolment
and health status, and higher labour force participation rates) and allocate more to alcohol and tobacco
and less to food than those with partners (Handa, 1994, 1996).
The goal of this article is to determine whether heterogeneity exists in the intrahousehold allocation
of international remittances sent to households in the Philippines by temporary international migrant
workers, or Overseas Filipino Workers (OFWs). Migration may affect household structures and alter
the dynamics of control over household resources, which then could impact intrahousehold allocation;
the household member working abroad could have more influence over the allocation of household
resources due to an increase in earned income, or less influence because of the limited ability of an
absent member to monitor the use of remittances – a classic example of a moral hazard problem.2 In
China, for example, Chen (2006) found that imperfect monitoring by migrant fathers affected how
mothers allocated resources. Mothers’ consumption of some goods that are difficult to monitor
increased, and mothers’ household labour hours decreased while children’s household labour hours
– along with children’s nutritional intake – increased.
In the Philippines, OFWs have been called ‘new heroes’ because they often serve as the primary
income providers for their families in the Philippines, and their transfers are a huge source of foreign
reserves. Approximately two million OFWs (about 2% of the total population of the country) were
working abroad in 2009, and about 86 per cent of them had spouses still in the Philippines; about 51 per
cent of the OFWs were male and 49 per cent female.3 One out of three worked as labourers or unskilled
workers, including domestic helpers, cleaners and manufacturing labourers. Most OFWs – 20 per cent –
worked in Saudi Arabia; 14 per cent worked in the United Arab Emirates, Singapore, Hong Kong, Japan,
Qatar and Taiwan; 9 per cent in Europe; and 8 per cent in North and South America. These numbers only
refer to OFWs and not Filipinos overseas (FOs) in general, the latter of which include permanent residents
and citizens of North American countries, Australia and other developed countries. The focus of this
article is on OFWs whose migration is temporary and who have spouses or parents in the Philippines who
receive the remittances that they send. In such families, the reasons for remitting and the possible uses of
remittances may be different than in the families of other Filipino migrants.
On average, remittances of OFWs are about 50 per cent of the total household income of remittance-
receiving households.4 The inflow of international remittances to the Philippines in 2009 amounted to
approximately 15 billion US dollars, making the Philippines one of the top four remittance-receiving
countries (International Monetary Fund, 2010). Remittances also constituted about 10 per cent of the GDP
in 2013 (World Bank, 2015). Given the importance of remittances, how they are spent and allocated not
only affects the welfare of the individual households but also impacts the Philippine economy.
This study identifies four types of households based on the gender of the household head and the
presence/absence of a spouse to account for moral hazard or imperfect monitoring and variations in
allocation decisions assumed to be attributable to asymmetric preferences of men and women. This
article examines how these four household types differ in the allocation of the international remittances
they receive and whether the allocations improve the welfare of the households, and of the children in
particular. Educational attainment of the household head is incorporated into the analysis to test
whether those who have more education behave differently than those with less education in their
allocation of remittance income.
The results of generalised linear model (GLM) estimations and post-estimation tests reveal differ-
ences in the allocation of remittances related to the household heads’ educational levels and the
household structures. First, the presence of a female decision-maker or female authority rather than the
existence of a female head per se affects the allocation to children. However, for alcohol and tobacco,
regardless of the presence of a wife or the educational attainment of the head, male heads allocate
more to these goods than female heads whose husbands work as OFWs.
Second, heterogeneity in the allocation of resources exists within subgroups of female heads.
In particular, college-educated single female heads (divorced, widowed or separated) allocate
more to adult clothing than college-educated female heads with OFW spouses. However, single
856 M. Pajaron

female heads with only a high school education tend to allocate more to children’s clothing, and
alcohol and tobacco, than female heads (of similar educational backgrounds) with OFW spouses.
Third, the less-educated heads allocate more to education than the more-educated heads do. Fourth,
heterogeneity in resource allocation also exists within the same household type conditional on the
educational attainment of the household head: among male heads whose wives are present, those with
less formal education allocate less to alcohol and tobacco than those with more formal education.
The results provide a nuanced understanding of heterogeneity in intrahousehold allocation of
resources (remittances). The findings reveal differences that go beyond a simple male–female dichot-
omy, showing that the educational attainment of household heads and the presence of a female
authority play important roles in allocation decisions as well.
The remainder of this article is structured as follows. Section 2 briefly discusses the existing
literature on intrahousehold allocation and remittances in general, and intrahousehold allocation in
the Philippine context specifically. Section 3 presents the data and methodology used in the study.
Section 4 reports the main results of the article. Section 5 is the conclusion.

2. Intrahousehold Allocation and Remittances


Some authors have explored how remittances are used depending on the gender of the household head.
Guzman et al. (2008), using Ghana data, showed that while international remittances decreased the
expenditure share for food and increased the expenditure shares for consumer and durable goods,
housing, health, utilities and transport in female-headed households, the share spent on education was
unaffected. After controlling for the remitter’s gender and relation to the household head, the study
found that female heads with a spouse working abroad allocated more to education than male heads
with a spouse working abroad. Malone (2007) analysed how the impact of remittances on children’s
education depended on the revealed preference of mothers when they were the default household
heads. The study showed that asymmetric preferences exist and that the allocation of remittances
differs depending on the gender of the receiver.
This article expands on these gender-based studies of the allocation of remittances by incorporating
the presence of a spouse and educational attainment of the household head into its analyses to reveal
other factors that influence remittance allocation in the Philippine context.

2.1. Gender Differentials and Intrahousehold Allocation in the Philippines


Households in the Philippines are interesting to analyse because wives and husbands both have some
measure of control over resources (Illo, 1995; Israel-Sobritchea, 1994; Jefremovas, 2000). Eder (2006)
contended, however, that even if Filipino households are relatively egalitarian compared to those in other
societies, Filipino women are still disadvantaged at various levels when compared to men. Although the
custom is for husbands to hand their wages over to their wives, which may suggest that women have
control over household resources, caveats exist. For example, in some cases women’s access to economic
assets may be indirect, especially if they are unemployed, whether by choice or by circumstance, which
limits their role in the allocation of resources (Eviota, 1986). In addition, in poor households, women have
small amounts of money to allocate, which limits their ability to make economic plans and decisions. Not
only do women have a limited role in allocating resources, they may also be unable to refuse requests
from husbands for money to drink or gamble (Chant & McIlwaine, 1995).
Husbands and wives appear to allocate resources differently, as well. For example, Senauer, Garcia,
and Jacinto (1988) studied how the opportunity costs of husbands and wives can influence intrahouse-
hold allocation of food in the rural Philippines, and found that the estimated wage rate of the mother
and wife is positively correlated with the relative calories allocated to both herself and her children and
negatively correlated with those allocated to the husband. Inversely, an increase in the wage rate of the
husband and father increases his own and his wife’s allocation but decreases the children’s.
Heterogeneity in the intrahousehold allocation of international remittances 857

3. Data and Methodology


3.1. Data Description
This article uses the 2003 merged Philippine dataset from the Labour Force Survey (LFS), the Family
Income and Expenditure Survey (FIES) and the Survey on Overseas Filipinos (SOF). This is the only
officially merged dataset that contains both household and migrant information in the Philippines. The
SOF contains data on the amount of remittances and on the socio-economic characteristics of workers
who were working or had worked abroad. The LFS reports employment status, age, educational
attainment and income of each household member. The FIES is a nationally representative survey
conducted every three years, which provides socio-economic information on Philippine households.5
Included in the analysis are 2,055 households with OFW members, which are categorised into four
types: female-headed with OFW spouses; de jure female-headed; male-headed with OFW spouses; and
male-headed with spouses present. About 55 per cent of these households are headed by males and
about 45 per cent are headed by females (Table 1). Female-headed households are classified based on
the marriage status of the head, with no husband present in either case: female household heads who
are married women and whose husbands are absent due to international labour migration (that is, their
husbands are OFWs) and de jure female household heads who are legal or customary heads, either
widowed, single or divorced, and who have full control of the household income and expenditures. De
jure female-headed households account for only about 14 per cent of the households in the analysis.

Table 1. Mean (standard deviation) of the dependent variables

Female-headed De jure female- Male-headed Male-headed


All households with headed households with households with
households OFW spouses households OFW spouses wives present

Dependent
variables (1) (2) (3) (4) (5)

Food 0.407 0.371 0.408 0.427 0.428


(0.128) (0.117) (0.134) (0.124) (0.129)
Education 0.054 0.067 0.049 0.059 0.043
(0.076) (0.075) (0.078) (0.083) (0.069)
Clothing and 0.074 0.077 0.069 0.077 0.072
personal (0.039) (0.040) (0.041) (0.042) (0.035)
effects
Children’s 0.008 0.010 0.006 0.008 0.006
clothing and (0.012) (0.012) (0.012) (0.013) (0.009)
personal
effects
Adults’ clothing 0.066 0.067 0.062 0.069 0.066
and personal (0.037) (0.037) (0.039) (0.037) (0.034)
effects
Health 0.024 0.021 0.035 0.015 0.028
(0.055) (0.044) (0.079) (0.037) (0.058)
Household 0.022 0.028 0.021 0.022 0.018
operations (0.024) (0.028) (0.025) (0.023) (0.017)
Alcohol and 0.015 0.004 0.008 0.028 0.020
tobacco (0.022) (0.009) (0.015) (0.030) (0.022)
Durable goods 0.032 0.033 0.025 0.032 0.033
(0.079) (0.082) (0.139) (0.074) (0.080)
Non-durable 0.002 0.003 0.002 0.002 0.002
goods (0.006) (0.005) (0.006) (0.006) (0.006)
Other goods 0.368 0.396 0.381 0.339 0.355
(0.120) (0.106) (0.133) (0.118) (0.120)
Number of 2,055 644 287 394 730
observations
858 M. Pajaron

Male-headed households, on the other hand, are classified based on the presence or absence of the wife
in order to account for the differences in allocation that may be attributed to the influence of a female
authority: male heads with OFW spouses (that is, the wives are absent), and male heads with wives
present.
Table 1 provides information regarding budget shares for each type of good by each household type.
Between the two subgroups of female heads, female heads with OFW spouses allocate more to
education, on average, while de jure female heads allocate more to health (Table 1, columns 2 and
3). The higher budget share for health among de jure female heads may be attributable to their average
age (58 years old) and the higher percentage of women aged 60 and above in their households
(Table 2, columns 2 and 3), factors which could mean that they seek more medical attention and have a
higher demand for health care.
Comparing subgroups of male-headed households, those with OFW wives allocate more on average
to education, clothing and personal items (both adults’ and children’s), household operations, and
alcohol and tobacco, and less to health than male-headed households whose wives are present
(Table 1, columns 4 and 5).
Among the households with migrant spouses, male-headed households, on average, allocate more to
food, and alcohol and tobacco, and less to education and health than female-headed households
(Table 1, columns 2 and 4). In addition, male-headed households with OFW spouses are the poorest
in terms of net income among the four household types (Table 2, column 4). The economic
disadvantage of this group may be explained by the low educational attainment of the heads and
their spouses (high school level at most) and the fact that they live in the agriculture-dominated
northern region. They also receive fewer remittances than female-headed households with OFW
spouses, which receive the most remittances among all household types (Table 2, columns 2 and 4).
This may be attributable to the earning capacity of the migrant spouse, which can be inferred from his/
her educational attainment and type of job. On average, about 70 per cent of the migrant spouses of
female heads are college-educated, compared to only about 40 per cent of male heads’ spouses.
Consequently, most of the OFW wives of male heads work as labourers and unskilled workers (72 per
cent), compared to only about 10 per cent of the OFW husbands of female heads. It can be inferred
that OFW wives earn less and therefore remit less than OFW husbands.

3.2. Empirical Model


This article uses the 10 categories of goods specified in the FIES as household expenditure types:
food; education; adults’ clothing and personal items; children’s clothing and personal items; health;
household operations; alcohol and tobacco; durable goods; nondurable goods; and other goods.
To determine the differential impacts of remittances on household expenditures depending on
household structures, this article uses the following household level expenditure share function,
which is an extension of the Working–Leser expenditure function; variations of this have been used
in other papers as well (Guzmán et al., 2008; Hoddinott & Haddad, 1995; Quisumbing & Maluccio,
2003):

cih ¼ β0i þ β1i rh þ β2i eh þ β3i FHHh þ β4i DJ FHHh þ β5i MHHh þ β6i FHHh  rh
(1)
þ β7i DJ HHh  rh þ β8i MHHh  rh þ β9i Xh þ uih

where cih is the expenditure share on the ith good of household h; rh are remittances received by a
household in log form; eh is a discrete variable for education (1 if the household head is college-
educated); FHHh, DJ_FHHh and MHHh are discrete variables for female heads with OFW husbands,
de jure female heads and male heads with OFW wives respectively; FHHh*rh, DJ_FHHh*rh and
MHHh*rh are the interaction terms that represent the differences in the allocation of remittances of the
three different household structures relative to the omitted household structure – male heads whose
wives are present (MHH_Wh); Xh is a vector of household characteristics that affect the allocation of
Table 2. Mean (standard deviation) of the independent variables

All Female-headed households with De jure female-headed Male-headed households with Male-headed households with
households OFW spouses households OFW spouses wives present

Independent variables (1) (2) (3) (4) (5)

Income 137,187 121,249 137,378 107,545 167,170


(140,852) (133,734) (132,983) (124,193) (152,547)
Remittances 133,378 229,039 109,516 87,884 82,922
(157,749) (211,027) (110,644) (86,220) (100,243)
Household expenditures 208,163 262,233 196,362 154,014 194,329
(155,150) (187,647) (136,747) (115,653) (133,434)
Household size 4.90 5.38 5.82 4.92 6.63
(2.22) (1.81) (2.38) (1.75) (2.39)
Age of head 48.45 39.56 57.87 41.73 56.21
(13.58) (7.96) (16.44) (9.48) (10.94)
Educational attainment of household head
High school or less 0.630 0.382 0.753 0.685 0.77
College 0.369 0.616 0.247 0.312 0.230
Educational attainment of spouse of household head
High school or less 0.474 0.294 0.599 0.752
College 0.385 0.707 0.396 0.248
Household composition shares
Male less than 1 year old 0.006 0.008 0.008 0.003 0.006
Male 1–6 years old 0.048 0.063 0.034 0.048 0.040
Male 7–14 years old 0.078 0.095 0.075 0.100 0.052
Male 15–24 years old 0.093 0.081 0.087 0.099 0.103
Male 25–59 years old 0.214 0.225 0.160 0.242 0.212
Male 60 years old and higher 0.032 0.007 0.006 0.015 0.072
Female less than 1 year old 0.005 0.006 0.004 0.002 0.007
Female 1–6 years old 0.047 0.065 0.039 0.045 0.035
Female 7–14 years old 0.069 0.090 0.056 0.084 0.047
Female 15–24 years old 0.110 0.110 0.111 0.102 0.113
Female 25–59 years old 0.251 0.232 0.298 0.240 0.256
Female 60 years old and 0.047 0.017 0.121 0.021 0.058
higher

(continued )
Heterogeneity in the intrahousehold allocation of international remittances 859
860 M. Pajaron

Table 2. (Continued)

All Female-headed households with De jure female-headed Male-headed households with Male-headed households with
households OFW spouses households OFW spouses wives present

Independent variables (1) (2) (3) (4) (5)

Regions (Arranged from north to south)


Ilocos Region 0.101 0.071 0.087 0.165 0.099
Cagayan Valley 0.084 0.022 0.084 0.170 0.093
Central Luzon 0.106 0.127 0.084 0.096 0.100
Bicol Region 0.034 0.040 0.024 0.025 0.036
Western Visayas 0.078 0.068 0.118 0.061 0.081
Central Visayas 0.041 0.050 0.042 0.015 0.047
Eastern Visayas 0.021 0.025 0.028 0.008 0.023
Western Mindanao 0.023 0.012 0.014 0.036 0.029
Northern Mindanao 0.027 0.033 0.024 0.015 0.029
Southern Mindanao 0.032 0.017 0.028 0.053 0.036
Central Mindanao 0.033 0.016 0.038 0.041 0.041
National Capital Region 0.145 0.227 0.129 0.096 0.107
Cordillera Administrative 0.051 0.030 0.063 0.076 0.051
Region
Autonomous Region of 0.028 0.005 0.014 0.023 0.056
Muslim Mindanao
Caraga 0.018 0.012 0.031 0.008 0.022
CALABARZON 0.156 0.230 0.167 0.089 0.123
MIMAROPA 0.023 0.016 0.024 0.023 0.029
Number of observations 2,055 644 287 394 730
Heterogeneity in the intrahousehold allocation of international remittances 861

resources, which includes log of total expenditure per capita and its square, log of household size, age
of the household head, educational attainment of the spouse, the proportion of demographic groups in
the household, and location dummies; and uh is the error term.6 Square of per capita expenditure
ensures that any observed differences in the effects of household type cannot be attributed to
nonlinearities in the Engle Curve (Hoddinott & Haddad, 1995; Thomas & Chen, 1994).
The first hypothesis tests (Equations 2–5) simply aim to determine the impact of each household
type on the allocation of remittances.

@cih
¼ β1i þ β6i ¼ 0 (2)
@rh FHH¼1

@cih
¼ β1i þ β7i ¼ 0 (3)
@rh DJ FHH¼1


@cih
¼ β1i þ β8i ¼ 0 (4)
@rh MHH¼1

@cih
¼ β1i ¼ 0 (5)
@rh MHH W ¼1

Equations (6), (7) and (8) test for differences in the allocation of remittances within subgroups of
female heads; between the two household types with OFW spouses; and between de jure female heads
and male heads with OFW spouses respectively.

@cih @cih
 ¼ β6i  β7i ¼ 0 (6)
@rh FHH¼1 @rh DJ FHH¼1


@cih @cih
 ¼ β6i  β8i ¼ 0 (7)
@rh FHH¼1 @rh MHH¼1

@cih @cih
 ¼ β7i  β8i ¼ 0 (8)
@rh DJ FHH¼1 @rh MHH¼1

To compare the allocation behaviour of each of the three household structures in which spouses are
absent with that of households with male heads whose wives are present, the coefficients of the
interaction terms in Equation (1) are examined: β6i ¼ 0; β7i ¼ 0; β8i ¼ 0. The results of these 10
hypothesis tests are displayed in Tables 3 and 4, which will be discussed in detail in later sections.
To add to the analysis of heterogeneity in the allocation of remittances, educational attainment of the
household head is included in Equation (9):

cih ¼ γ0i þ γ1i rh þ γ2i eh þ γ3i FHHh þ γ4i DJ FHHh þ γ5i MHHh þ γ6i FHHh  rh
þ γ7i DJ FHHh  rh þ γ8i MHHh  rh þ γ9i rh  eh þ γ10i FHHh  eh þ γ11i DJ FHHh  eh
(9)
þ γ12i MHHh  eh þ γ13i FHHh  rh  eh þ γ14i DJ FHHh  rh  eh þ γ15i MHHh  rh  eh
þ γ16i Xh þ uih

The variables here are similar to those in Equation (1) with additional interaction terms that capture:
the effect of education of household head on how remittances are used (rh  eh ); the impact of
education of household head and household type on intrahousehold allocation
(FHHh  eh ; DJ FHHh  eh ; MHHh  eh ); and the importance of the household type and education
Table 3. Marginal effects of remittances on household allocation by household type (GLM)

Food Education Child clothing Adult clothing Health HH oper. Alcohol/tobacco Durable Non-durable Other goods

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

Household type (male head with wife present omitted)


862 M. Pajaron

Female head with −0.12* 0.10 0.57* 0.35*** −0.20 0.15 −0.16 −0.14 −0.64 0.06
OFW spouse (0.11) (0.44) (0.34) (0.12) (0.53) (0.34) (0.35) (0.73) (0.72) (0.11)
De jure female head −0.08 −0.48 0.28 0.03 −0.43 0.31 −0.34 −0.25 −1.01 0.14*
(0.12) (0.84) (0.37) (0.10) (0.52) (0.28) (0.26) (0.68) (0.72) (0.13)
Male head with −0.04 0.52 0.46 0.11 −0.05 0.24 0.18 −0.49 −0.39 −0.05
OFW spouse (0.08) (0.38) (0.30) (0.11) (0.58) (0.20) (0.21) (0.58) (0.65) (0.10)
Per capita −0.20 4.78*** 4.70*** 0.83* 3.81** 1.39* 2.42** −2.67 3.95** 1.08***
expenditure (0.33) (1.13) (1.10) (0.47) (1.66) (0.77) (1.20) (1.68) (1.81) (0.59)
Per capita −0.01 −0.21*** −0.23*** −0.04** −0.15** −0.05 −0.14** 0.18** −0.18** −0.04**
expenditure squared (0.02) (0.05) (0.05) (0.02) (0.08) (0.04) (0.06) (0.08) (0.09) (0.03)
Household size −0.06*** 0.42*** 0.37*** 0.01 0.10 0.22*** 0.00 0.08 −0.06 −0.02
(0.03) (0.11) (0.12) (0.04) (0.17) (0.09) (0.10) (0.18) (0.20) (0.04)
Age of head 0.00 0.01** −0.01** 0.00 0.01 0.00 −0.01** −0.01 −0.01** 0.01
(0.00) (0.00) (0.00) (0.00) (0.01) (0.00) (0.00) (0.01) (0.01) (0.00)
Educational attainment of household head (high school or less omitted)
College 0.00 0.11 0.06 0.04 0.03 0.17*** −0.42*** −0.64*** −0.14 0.04***
(0.02) (0.07) (0.07) (0.03) (0.11) (0.06) (0.08) (0.15) (0.15) (0.03)
Educational attainment of spouse of household head (high school or less omitted)
College −0.02 0.14* −0.02 0.01 −0.30** 0.20*** −0.08 −0.06 −0.17 0.01
(0.02) (0.08) (0.08) (0.03) (0.12) (0.06) (0.08) (0.14) (0.14) (0.03)
Remittances −0.01*** 0.06*** 0.04*** 0.01** 0.03 0.01 −0.02 0.03 0.00 0.00
(0.00) (0.02) (0.01) (0.01) (0.02) (0.01) (0.01) (0.03) (0.02) (0.00)
Interaction terms
Female head with 0.01** 0.01 −0.07** −0.03*** 0.00 0.00 −0.10*** −0.02 0.07 −0.01
OFW spouse × Remit (0.01) (0.04) (0.03) (0.01) (0.05) (0.03) (0.03) (0.06) (0.06) (0.01)
De jure female 0.01 0.05 −0.03 0.00 0.05 −0.01 −0.03 −0.01 0.09 −0.01
head × Remit (0.01) (0.07) (0.03) (0.01) (0.05) (0.03) (0.02) (0.06) (0.07) (0.01)
Male head with 0.00 −0.02 −0.04 −0.01 −0.03 0.00 0.00 0.03 0.02 0.00
OFW spouse × Remit (0.01) (0.03) (0.03) (0.01) (0.05) (0.02) (0.02) (0.05) (0.06) (0.01)
Number of obs. 2,055 2,055 2,055 2,055 2,055 2,055 2,055 2,055 2,055 2,055

Notes: ***, **, * indicate 1 per cent, 5 per cent and 10 per cent statistical significance respectively. Robust standard errors (in parentheses) are reported.
Dummy variables for the 17 regions in the Philippines and for 11 demographic groups are also included but are not shown.
Table 4. Wald tests on the differential impact of household structure on remittance allocation (Chi-square, p-values, marginal effects)

Food Education Child clothing Adult clothing Health HH oper. Alcohol/tobacco Durable Non-durable Other goods

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

A. Wald tests of the joint significance of coefficients for remittances and interaction terms
Female head with OFW spouse (coefficients for remittances + female head with OFW spouse × remittances = 0)
Chi-square 0.63 5.22 0.73 2.95 0.49 0.16 16.19 0.01 1.47 2.48
p-value 0.43 0.02 0.39 0.09 0.49 0.69 0.00 0.91 0.23 0.12
Marginal effects 0.00 0.06** −0.02 −0.01* 0.03 0.01 −0.11*** 0.00 0.07 0.00
De jure female head (coefficients for remittances + de jure female head × remittances = 0)
Chi-square 0.11 2.62 0.13 0.74 2.94 0.04 4.03 0.16 2.37 2.90
p-value 0.75 0.11 0.72 0.39 0.09 0.84 0.04 0.69 0.12 0.09
Marginal effects 0.00 0.10 0.01 0.00 0.07* 0.00 −0.04** 0.02 0.09 −0.01*
Male head with OFW spouse (coefficients for remittances + male head with OFW spouse × remittances = 0)
Chi-square 1.42 1.97 0.01 0.06 0.00 0.28 0.42 1.45 0.11 0.04
p-value 0.23 0.16 0.93 0.81 1.00 0.60 0.52 0.23 0.74 0.85
Marginal effects 0.00 0.03 0.00 0.00 0.00 0.00 −0.01 0.05 0.01 0.00
B. Wald tests of equality of coefficients between household structures
Coefficients for female head with OFW spouse × remittances = de jure female head × remittances
Chi-square 0.58 0.30 0.72 3.36 0.53 0.20 4.03 0.04 0.09 0.13
p-value 0.45 0.59 0.40 0.07 0.46 0.66 0.04 0.84 0.77 0.72
Marginal effects 0.00 −0.03 −0.03 −0.02* −0.04 0.01 −0.07** −0.01 −0.02 0.00
Coefficients for female head with OFW spouse × remittances = male head with OFW spouse × remittances
Chi-square 1.89 0.65 0.50 2.02 0.24 0.01 10.05 0.51 0.42 0.94
p-value 0.17 0.42 0.48 0.16 0.62 0.94 0.00 0.48 0.52 0.33
Marginal effects 0.00 0.03 −0.02 −0.01 0.03 0.00 −0.10*** −0.04 0.05 0.05
Coefficients for de jure female head × remittances = male head with OFW spouse × remittances
Chi-square 0.15 0.98 0.05 0.18 1.38 0.22 1.48 0.23 0.83 1.44
p-value 0.70 0.32 0.70 0.82 0.67 0.64 0.22 0.63 0.36 0.23
Marginal effects −0.00 0.07 −0.00 −0.00 0.07 −0.01 −0.03 −0.03 0.07 −0.01
Heterogeneity in the intrahousehold allocation of international remittances 863
864 M. Pajaron

of the household head in allocating remittances (FHHh  rh  eh ; DJ FHHh  rh  eh ;


MHHh  rh  eh ). As in Equation (1), a high school education is the base category for educational
attainment (eh ¼ 0) and male heads with wives present (MHH Wh ) is the omitted variable for
household type.
In all, 24 hypothesis tests were conducted to determine whether educational attainment of the
household head has differential impacts on the allocation of remittances. These tests are presented and
discussed in the Online Appendix.

3.3. Identification Issues


A biased and inconsistent ordinary least square (OLS) estimate may arise in Equations (1) and (9) if
remittances are endogenous. Endogeneity may be attributed to an unobservable process, such as
household preferences in allocating resources or a migration decision that impacts both remittances
and household allocation. To address this issue, this article considers the exchange rate as an excluded
instrumental variable. If the currency of the country where the migrant worked appreciated against the
Philippine peso during the period covered by the data, it is considered a positive income shock for the
migrant’s origin household in the Philippines (Yang, 2008).
Household structure, which depends on the gender and the presence or absence of a spouse in the
household, may also be endogenous. For example, unobserved variables such as migration preferences
of the household or factors that lead to separation or the death of a spouse may affect both the
allocation of resources and the structure of the household. Exchange rate, age and regional dummies
were tested as instruments.
These instrumental variables, however, were tested and shown to be weak instruments, which
indicates that their inclusion could lead to a larger inconsistency in the estimated coefficient than
the OLS estimator; therefore, this article proceeds without instrumenting (Wooldridge, 2006).7
Equations (1) and (9) are then estimated using the generalised linear model (GLM) with logit link
function and binomial distribution given that cih is a proportion bounded by 0 and 1 (Papke &
Wooldridge, 1996).8

4. Results
The regression analyses are divided into two: the first focuses on how the allocation of remittances
differs across household structures; the second tests for heterogeneity in the allocations of
remittances conditional on the household type and the educational attainment of the household
head.

4.1. Impact of Remittances on Household Allocation by Household Type


To determine the impact of remittances on expenditures in different household types, Equation (1) is
estimated using GLM, and the results are presented in Table 3.
It can be gleaned from the results that the marginal effects of the female head with OFW spouse
variable are statistically significant and negative in the food regression but positive in the children’s
clothing and adults’ clothing regressions (Table 3, columns 1, 3 and 4). These results suggest that
while female heads with OFW spouses allocate 57 per cent more to children’s clothing and 35 per cent
more to adults’ clothing, they allocate 12 per cent less to food relative to the omitted household type
(male heads with wives present).
The sign and statistical significance of the coefficients for per capita expenditure and its square
indicate that as per capita expenditure increases, heads allocate more to almost all types of goods
except food, household operations and durable goods, but that this positive effect is decreasing. This
implies that the relationship of per capita expenditure and the budget shares for these goods is
nonlinear (concave).
Heterogeneity in the intrahousehold allocation of international remittances 865

The results also show that as the household size increases, the allocations to food decrease, while the
allocations to education, children’s clothing and household operations increase. In particular, as household
size increases by 20 per cent (an increase of one individual, on average), the allocations to food decrease
by 1.2 per cent (about 14,750 pesos, on average) while there are increases in allocations to education by
8.4 per cent (about 2,700 pesos, on average), to children’s clothing by 7.4 per cent (1,900 pesos, on
average), and to household operations by 4.5 per cent (6,480 pesos, on average). It is as if the household
head decides to reallocate the budget shares for food to these three other categories of goods. Age of the
household head, on the other hand, is positively correlated with allocations to education but inversely
correlated with allocations to children’s clothing, alcohol and tobacco, and non-durable goods.
As for the impact of the educational attainment of household heads, keeping everything else
constant, college-educated household heads tend to allocate less than high-school-educated household
heads to alcohol and tobacco and durable goods, but more to household operations. Household heads
whose spouses are college educated tend to allocate less to health but more to education and household
operations than those whose spouses have less education.
The next subsections analyse how remittances are allocated depending on the household type.

4.1.1. Impact of remittances by household type. After the GLM regressions were run, post-estimation
Wald tests were conducted on the joint significance of the coefficients of remittances and the
interaction terms to test Equations (2) to (4) and to determine the remittance allocation of female
heads with OFW spouses, de jure female heads and male heads whose wives are OFWs. Panel A of
Table 4 displays the results (Chi-square statistics, p-values and marginal effects) of these tests. No
post-estimation test is needed for male heads whose wives are present, given that it is the omitted
household type; however, as indicated in Equation (5), their remittance allocation is reflected in the
marginal effects of the remittances variable shown in Table 3.
In female-headed households with OFW spouses, the welfare of the members is improved with an
increased allocation for education (6%) and a decreased budget for alcohol and tobacco (11%), given a
100 per cent increase in remittances (Table 4 panel A, columns 2 and 7). Similarly, although most of
the de jure female heads have only a high school education (75%), their households are no worse off in
terms of income and remittance receipt (Table 2, column 3). In addition, their household members are
better off, given that allocations for health increase by 7 per cent and the budget allotted to alcohol and
tobacco decreases by 4 per cent as remittances received increase (Table 4 panel A, columns 5 and 7).
In male-headed households, it appears that the presence of the wife results in welfare-improving
allocations of remittances. In particular, although male household heads whose wives are present are
the least educated and receive the least amount of remittances among the four household types, as seen
in Table 2 (column 5), they still allocate their remittances to education (6%), children’s clothing (4%),
and adults’ clothing (1%) (Table 3, columns 2, 3, and 4). On the other hand, the remittances received
by male-headed households with OFW wives are not allocated to improving the welfare of the
children. This pattern might contribute to perpetuating and extending poverty to the next generation
given that these households are also the poorest among the four household types.
The results thus far are somewhat contrary to the findings of Appleton (1996) for Uganda and Joshi
(2004) for Bangladesh, who found that while female-headed households are poorer and own less
agricultural land, this is offset by high remittance receipts and improved welfare of the children in
terms of short-term health status, caloric food intake, school enrolment rates and labour force
participation. In the Philippines, male heads are the ones at an economic disadvantage, but the
presence of a female authority in such households, even when she is not the household head herself,
increases the allocations for education, clothing and personal items of children.9 Before conjecturing
that poverty-alleviation policies in the Philippines should consider not only female-headed households
but also male-headed households with wives present, this article formally tests heterogeneity in
allocation across households to further reveal the links between remittances, household structure,
poverty and household welfare.
866 M. Pajaron

4.1.2. Heterogeneity across household structures. Post GLM tests were again conducted, this time
to test Equations (6) to (8) and to determine heterogeneity in allocating remittances across house-
hold structures. The results of these joint significance tests are shown in Table 4, panel B; any
comparisons with male heads whose wives are present, as the omitted variable, are shown in
Table 3 through the marginal effects of the interaction of remittances and the three household types
with spouses absent.
The Chi-square statistics and p-values show that the null hypothesis that the interaction of
remittances and female heads with OFW spouses is equal to the interaction of remittances and de
jure female heads can be rejected for adult clothing and alcohol and tobacco regressions. The signs and
statistical significance of the marginal effects imply that heterogeneity among the female-headed
households exists: female heads with OFW spouses allocate less to adults’ clothing (2%) and alcohol
and tobacco (7%) than de jure female heads (columns 4 and 7). As discussed in the previous
subsection, the two types of female-headed households both allocate less to alcohol and tobacco as
remittances increase, but the decrease is greater in female-headed households with OFW spouses.
The marginal effects of the interaction of female heads with OFW spouses and remittances (Table 3)
and tests on the equality of female heads with OFW spouses and male heads with OFW spouses
(Table 4, panel B) show that relative to female heads with OFW spouses, both types of male heads
allocate 10 per cent more to alcohol and tobacco. Consistent with the existing literature, these findings
suggest that male household heads have a strong preference for spending money on alcohol and
tobacco regardless of whether their wives are present or absent. It appears that while the presence of a
female authority in a household matters in spending on the education of children, it is irrelevant in
spending on alcohol and tobacco.

4.2. Impact of Remittances on Household Allocation by Household Type and Educational Attainment
of the Household Head
Given the differences in how remittances are allocated across the different household structures, this
article explores the possibility of attributing such heterogeneity to the educational attainment of the
household head. To accomplish this, Equation (9) is estimated using GLM, and the results are
displayed in Table 5.
The results show that female heads with OFW spouses allocate less to food but more to clothing
(children’s and adults’) relative to the omitted household type – male heads with wives present. In
addition, the relationship of per capita expenditure and budget share on almost all expenditure types
(except food, household operations and durable goods) is nonlinear. With household size, as it
increases, allocations for education, children’s clothing and household operations increase, while the
budget share for food decreases. Age of the head is positively correlated with the allocation to
education but inversely related to allocations for children’s clothing, alcohol and tobacco, and non-
durable goods. The allocation to alcohol and tobacco is higher among high-school-educated heads than
college-educated heads, while heads whose spouses are college-educated allocate more to education
and household operations, and less to food and health.
As to the allocation of remittances, four separate analyses were conducted and are discussed in the
following subsections. The results are displayed in Table 6 in four panels corresponding to the four
different sets of tests.

4.2.1. Impact of remittances by household type and educational attainment of household head. After
the GLM regression, Equations (A1) to (A7), presented in the Online Appendix, were tested in order
to determine the impact of household type and educational attainment of the household head on the
allocation of remittances. The results are displayed in panel A of Table 6.
For female heads with OFW spouses and with college education, the null hypothesis that remit-
tances do not affect allocations can only be rejected in the alcohol and tobacco regression: allocations
to these goods decrease by 9 per cent as remittances increase. For female heads with OFW spouses and
with high school education, allocations to food and education increase by 1 per cent and 17 per cent
Table 5. Marginal effects of remittances on household allocation by household type and educational attainment of the head (GLM)

Food Education Child clothing Adult clothing Health HH oper. Alcohol/tobacco Durable Non-durable Other goods

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

Household type (male head with wife present omitted)


Female head with −0.27*** −1.06 1.24*** 0.49*** −0.12 0.07 −0.12 0.38 −0.57 0.18***
OFW spouse (0.10) (0.77) (0.29) (0.17) (0.59) (0.29) (0.43) (0.81) (0.98) (0.11)
De jure female head −0.06 −0.88 −0.04 0.15 −1.32 0.03 −0.50 −0.09 −0.75 0.12
(0.16) (0.91) (0.51) (0.13) (0.89) (0.34) (0.33) (0.90) (0.85) (0.17)
Male head with −0.06 0.60 0.44 0.14 −0.16 0.23 0.17 −1.43* −0.28 −0.02
OFW spouse (0.09) (0.45) (0.34) (0.12) (0.67) (0.19) (0.21) (0.81) (0.63) (0.12)
Per capita expenditure −0.19 4.57*** 4.96*** 0.85** 3.73** 1.39* 2.46** −2.34 4.30** 1.01***
(0.34) (1.17) (1.09) (0.46) (1.64) (0.79) (1.21) (1.69) (1.92) (0.6)
Per capita expenditure −0.01 −0.20*** −0.24*** −0.04** −0.15** −0.05 −0.14** 0.17** −0.20** −0.04**
squared (0.02) (0.06) (0.05) (0.02) (0.08) (0.04) (0.06) (0.08) (0.09) (0.03)
Household size −0.06*** 0.40*** 0.38*** 0.01 0.08 0.23*** 0.00 0.06 −0.04 −0.02
(0.03) (0.11) (0.12) (0.04) (0.17) (0.09) (0.1) (0.19) (0.21) (0.04)
Age of head 0.00 0.01*** −0.01** 0.00 0.01 0.00 −0.01** −0.01 −0.01** 0.00
(0.00) (0.00) (0.01) (0.00) (0.01) (0.00) (0.00) (0.01) (0.01) (0.00)
Educational attainment of household head (high school or less omitted)
College −0.05 0.21 −0.16 0.15 −0.46 −0.02 −0.81*** −0.74 −0.68 0.18***
(0.08) (0.49) (0.27) (0.11) (0.47) (0.31) (0.23) (0.53) (0.45) (0.1)
Educational attainment of household head (high school or less omitted)
College −0.02* 0.17** −0.01 0.01 −0.31*** 0.20*** −0.09 −0.03 −0.19 0.01
(0.02) (0.08) (0.09) (0.03) (0.12) (0.06) (0.08) (0.14) (0.14) (0.03)
Remittances −0.01*** 0.07*** 0.04** 0.01** 0.02 0.00 −0.02* 0.03 −0.01 0.00
(0.00) (0.02) (0.02) (0.01) (0.02) (0.01) (0.01) (0.03) (0.02) (0.01)
Interaction terms
Female head with 0.02*** 0.10* −0.13*** −0.04*** 0.01 0.01 −0.10** −0.08 0.06 −0.01**
OFW spouse × Remit (0.01) (0.07) (0.03) (0.01) (0.05) (0.02) (0.04) (0.07) (0.09) (0.01)
De jure female 0.00 0.07 −0.00 −0.02 0.13 0.01 −0.01 −0.02 0.08 −0.01
head × Remit (0.01) (0.08) (0.05) (0.01) (0.08) (0.03) (0.03) (0.08) (0.08) (0.02)
Male head with 0.00 −0.02 −0.04 −0.01 −0.03 0.00 0.01 0.10 0.01 0.00
OFW spouse × Remit (0.01) (0.04) (0.03) (0.01) (0.06) (0.02) (0.02) (0.07) (0.06) (0.01)
College × Remit 0.00 −0.01 0.01 −0.01 0.05 0.02 0.05** −0.01 0.06 −0.01
(0.01) (0.04) (0.03) (0.01) (0.05) (0.03) (0.02) (0.05) (0.05) (0.01)
(continued )
Heterogeneity in the intrahousehold allocation of international remittances 867
868 M. Pajaron

Table 5. (Continued)

Food Education Child clothing Adult clothing Health HH oper. Alcohol/tobacco Durable Non-durable Other goods

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

Female head with OFW 0.34*** 1.56* −1.45** −0.37* −0.08 0.25 −0.04 −0.91 0.28 −0.31***
spouse × College (0.17) (0.93) (0.73) (0.23) (1.22) (0.59) (0.59) (1.38) (1.38) (0.19)
De jure female head × −0.05 0.74 0.76 −0.37** 1.78* 0.62 0.74 −0.47 −0.92 −0.02
College (0.25) (1.39) (0.72) (0.19) (1.05) (0.61) (0.49) (1.16) (1.48) (0.24)
Male head with 0.06 −0.53 −0.05 −0.06 0.52 −0.12 −0.25 2.42** −3.03 −0.13
OFW spouse × College (0.19) (0.80) (0.49) (0.26) (0.87) (0.62) (0.60) (1.16) (2.27) (0.18)
Female head with OFW −0.03*** −0.12 0.14** 0.03 −0.01 −0.02 −0.02 0.10 −0.02 0.02**
spouse × College × Remit (0.01) (0.08) (0.06) (0.02) (0.11) (0.05) (0.05) (0.12) (0.12) (0.02)
De jure female head × 0.00 −0.02 −0.08 0.04** −0.17* −0.06 −0.09* 0.05 0.06 0.00
College × Remit (0.02) (0.12) (0.07) (0.02) (0.10) (0.06) (0.05) (0.11) (0.14) (0.02)
Male head with OFW −0.01 0.03 0.01 0.01 −0.02 0.00 0.01 −0.19* 0.26 0.01
spouse × College × (0.02) (0.07) (0.05) (0.02) (0.08) (0.06) (0.06) (0.10) (0.22) (0.02)
Remit

Notes: ***, **, * indicate 1 per cent, 5 per cent and 10 per cent statistical significance respectively. Robust standard errors (in parentheses) are reported.
Dummy variables for the 17 regions in the Philippines and for 11 demographic groups are also included but are not shown.
Table 6. Wald tests on the differential impact of household structure and education of head on remittance allocation

Food Education Child clothing Adult clothing Health HH oper. Alcohol/tobacco Durable Non-durable Other goods

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

A. Wald tests of the joint significance of coefficients for remittances, education of head and interaction terms
(1) College female head with OFW spouse (coeffs. for remittances + female head with OFW spouse × remittances + remittances × college + Female head with OFW spouse
× remittances × college = 0)
Chi-square 1.74 1.49 1.22 0.15 0.48 0.05 9.68 0.29 1.08 0.04
p-value 0.19 0.22 0.27 0.70 0.49 0.82 0.00 0.59 0.30 0.85
Marginal effects 0.00 0.03 0.05 0.00 0.05 0.00 −0.09*** 0.04 0.08 0.00
(2) High school female head with OFW spouse (coeffs. for remittances + female head with OFW spouse × remittances = 0)
Chi-square 10.78 8.33 19.53 3.33 0.24 0.15 10.62 0.52 0.35 6.18
p-value 0.00 0.00 0.00 0.07 0.63 0.70 0.00 0.47 0.55 0.01
Marginal effects 0.01*** 0.17*** −0.09*** −0.02* 0.02 0.00 −0.11*** −0.04 0.05 −0.01**
(3) College de jure female head (coeffs. for remittances + de jure female head × remittances + remittances × college + de jure female head × remittances × college = 0)
Chi-square 0.02 1.50 0.31 5.99 0.31 0.65 5.21 0.69 2.70 5.82
p-value 0.89 0.22 0.58 0.01 0.58 0.42 0.02 0.41 0.10 0.02
Marginal effects 0.00 0.09 −0.02 0.02** 0.02 0.02 −0.07*** 0.05 0.17 −0.01**
(4) High school de jure female head (coeffs. for remittances + de jure female head × remittances = 0)
Chi-square 0.25 3.39 0.90 0.10 3.26 0.28 1.45 0.01 0.77 0.42
p-value 0.62 0.07 0.34 0.75 0.07 0.60 0.23 0.92 0.38 0.52
Marginal effects 0.00 0.13* 0.04 0.00 0.14* 0.01 −0.03 0.00 0.06 0.00
(5) College male head with OFW spouse (coeffs. for remittances + male head with OFW spouse × remittances + remittances × college + male head with OFW spouse ×
remittances × college = 0)
Chi-square 0.55 2.01 0.47 0.01 0.15 0.34 0.89 1.08 2.25 0.00
p-value 0.46 0.16 0.49 0.94 0.70 0.56 0.35 0.30 0.13 0.96
Marginal effects 0.00 0.05 0.01 0.00 0.01 0.02 0.04 −0.06 0.31 0.00
(6) High school male head with OFW spouse (coeffs. for remittances + male head with OFW spouse × remittances = 0)
Chi-square 1.09 1.74 0.00 0.07 0.09 0.15 0.76 3.87 0.00 0.06
p-value 0.30 0.19 0.99 0.80 0.77 0.70 0.38 0.05 0.97 0.81
Marginal effects 0.00 0.04 0.00 0.00 −0.01 0.00 −0.01 0.13** 0.00 0.00
(7) College male head with wife present (coeffs. for remittances + remittances × college = 0)
Chi-square 1.46 2.21 6.64 0.08 2.82 0.58 1.69 0.37 1.29 3.20
p-value 0.23 0.14 0.01 0.77 0.09 0.44 0.19 0.54 0.26 0.07
Marginal effects 0.00 0.05 0.05*** 0.00 0.06* 0.02 0.02 0.02 0.04 0.00*
(continued )
Heterogeneity in the intrahousehold allocation of international remittances 869
Table 6. (Continued)

Food Education Child clothing Adult clothing Health HH oper. Alcohol/tobacco Durable Non-durable Other goods

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

B. Wald tests of equality of coefficients within household structures conditional on education of head
870 M. Pajaron

(8) Coefficients for college female head with OFW spouse = high school female head with OFW spouse
Chi-square 9.62 4.45 7.28 1.44 0.11 0.00 0.33 0.76 0.08 1.93
p-value 0.00 0.03 0.01 0.23 0.74 0.99 0.56 0.38 0.78 0.17
Marginal effects −0.02*** −0.13** 0.15*** 0.01 0.03 0.00 0.02 0.09 0.03 0.01
(9) Coefficients for college de jure female head = high school de jure female head
Chi-square 0.19 0.13 1.13 4.16 1.69 0.90 0.86 0.24 0.76 1.18
p-value 0.67 0.72 0.29 0.04 0.19 0.34 0.35 0.63 0.38 0.28
Marginal effects 0.00 −0.03 −0.06 0.03** −0.11 −0.04 −0.03 0.04 0.11 −0.01
(10) Coefficients for college male head with OFW spouse = high school male head with OFW spouse
Chi-square 0.01 0.10 0.23 0.04 0.22 0.18 1.42 4.79 2.20 0.04
p-value 0.91 0.75 0.63 0.85 0.64 0.67 0.23 0.03 0.14 0.83
Marginal effects 0.00 0.01 0.01 0.00 0.03 0.01 0.05 −0.19** 0.31 0.00
C. Wald tests of equality of coefficients between household structures with college-educated heads
(11) Coefficients for college female head with OFW spouse = college de jure female head
Chi-square 0.66 0.58 1.49 4.14 0.11 0.54 0.19 0.00 0.57 2.83
p-value 0.42 0.45 0.22 0.04 0.74 0.46 0.66 0.95 0.45 0.09
Marginal effects 0.00 −0.06 0.08 −0.03** 0.02 0.03 −0.01 0.00 −0.09 0.01*
(12) Coefficients for College male head with OFW spouse = college male head with wife present
Chi-square 0.03 0.02 1.00 0.04 0.85 0.01 0.14 1.44 1.60 1.40
p-value 0.87 0.90 0.32 0.85 0.36 0.91 0.71 0.23 0.21 0.24
Marginal effects −0.00 0.00 −0.03 −0.00 −0.04 0.00 0.01 −0.08 0.26 0.00
(13) Coefficients for college female head with OFW spouse = college male head with OFW spouse
Chi-square 0.03 0.27 0.45 0.01 0.20 0.10 6.11 1.12 1.19 0.03
p-value 0.86 0.60 0.50 0.90 0.65 0.75 0.01 0.29 0.27 0.86
Marginal effects 0.00 −0.02 0.04 0.00 0.04 −0.01 −0.13** 0.10 −0.23 −0.00
(14) Coefficients for college de jure female head = college male head with OFW spouse
Chi-square 0.36 0.17 0.71 1.92 0.04 0.94 4.25 1.83 0.33 4.04
p-value 0.55 0.68 0.40 0.17 0.85 0.33 0.04 0.18 0.57 0.04
Marginal effects 0.00 −0.01 0.00 0.00 0.00 −0.01 −0.11*** 0.02 0.03 0.00

(continued )
Table 6. (Continued)

Food Education Child clothing Adult clothing Health HH oper. Alcohol/tobacco Durable Non-durable Other goods

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

(15) Coefficients for college female head with OFW spouse = college male head with wife present
Chi-square 0.22 0.19 0.01 0.23 0.01 0.07 11.1 0.07 0.18 0.81
p-value 0.64 0.66 0.93 0.63 0.92 0.78 0.00 0.80 0.67 0.37
Marginal effects 0.00 −0.01 0.00 0.00 0.00 −0.01 −0.11*** 0.02 0.03 0.00
(16) Coefficients for college de jure female head = college male head with wife present
Chi-square 0.33 0.25 2.64 3.17 0.42 1.21 6.89 0.18 1.36 1.21
p-value 0.57 0.62 0.10 0.08 0.52 0.27 0.01 0.67 0.24 0.27
Marginal effects 0.00 0.04 −0.07 0.02* −0.03 −0.04 −0.09*** 0.03 0.13 −0.01
D. Wald tests of equality of coefficients between household structures with high school-educated heads
(17) Coefficients for high school female head with OFW spouse = high school de jure female head
Chi-square 4.28 0.15 7.86 1.50 1.74 0.04 3.51 0.32 0.02 0.47
p-value 0.04 0.70 0.01 0.22 0.19 0.84 0.06 0.57 0.89 0.49
Marginal effects 0.01** 0.03 −0.13*** −0.02 −0.12 0.00 −0.08* −0.05 −0.01 0.00
(18) Coefficients for high school female head with OFW spouse = high school male head with OFW spouse
Chi-square 9.59 3.61 7.41 2.64 0.31 0.01 6.62 4.01 0.28 2.02
p-value 0.00 0.06 0.01 0.10 0.57 0.93 0.01 0.05 0.60 0.16
Marginal effects 0.02*** 0.12* −0.09*** −0.02 0.03 0.00 −0.10** −0.17** 0.05 0.00
(19) Coefficients for high school de jure female head = high school male head with OFW spouse
Chi-square 0.01 1.31 0.63 0.17 2.69 0.08 0.32 1.76 0.55 0.16
p-value 0.94 0.25 0.43 0.68 0.10 0.77 0.57 0.18 0.46 0.69
Marginal effects −0.00 0.09 0.04 −0.00 0.15 −0.00 −0.01 −0.12 0.06 −0.00
Heterogeneity in the intrahousehold allocation of international remittances 871
872 M. Pajaron

respectively, while allocations to clothing (children’s and adults’) and to alcohol and tobacco each
decrease by about 11 per cent. For de jure female heads, those with college education allocate less to
alcohol and tobacco (7%) but more to adult clothing (2%), while those with high school education
spend more on education (13%) and health (14%). For male heads with OFW spouses, the only
statistically significant result is for durable goods, for which the allocation increases by about 13 per
cent among high-school-educated male heads. For male heads with wives present, those with college
education allocate more to children’s clothing (5%) and health (6%). For those with high school
education, the results that can be seen from the marginal effects of the remittances variable in Table 5
and that test Equation (A8) imply that they tend to allocate more to education and to clothing
(children’s and adults’) at 7 per cent and 5 per cent respectively, while they allocate less to food
(1%), and alcohol and tobacco (2%).
In summary, female heads with OFW spouses, regardless of the heads’ educational attainment,
college-educated de jure female heads, and male heads with wives present who are less formally
educated appear to have a preference against making expenditures on alcohol and tobacco, showing
decreased allocations to these goods. In addition, those with high school education – female heads
with OFW spouses, de jure female heads and male heads with wives present – tend to allocate more to
education. These results are formally tested in the next subsection’s examination of heterogeneity in
allocation choices within each household structure depending on the educational attainment of the
household head.
The results showing positive and statistically significant allocations to health by less-educated de
jure female heads and college-educated male heads whose wives are present should be interpreted with
caution, because increased allocations to health do not automatically translate to improvement in the
health status of individuals. It is possible that in these households there are more infants or more
pregnant, elderly or sick members who require medical attention. Although it would be interesting to
examine the effects of remittances on children’s health outcomes (such as height–weight ratio and
other anthropometric measures, mortality rates or morbidity rates) when such data become available,
expenditures on health are still a good indicator of concern for the welfare of the household members
if there is increased attention to the health of individual members.10

4.2.2. Heterogeneity within household types across educational attainment of household heads.
Panel B of Table 6 displays the results of joint significance tests conducted after the GLM estimation
to test Equations (A9) to (A11), presented in the Online Appendix, in order to examine whether
heterogeneity exists within the three types of household structure with absent spouses (female heads
with OFW spouses, de jure female heads and male heads with OFW spouses) depending on the
educational attainment of the heads. The results of testing Equation (A12), shown in the Online
Appendix, on the other hand, to determine whether heterogeneity exists within the omitted household
type, male heads with wives present, can be seen in Table 5 through the marginal effects of the
interaction term of remittances and college education.
High-school-educated female heads with OFW spouses allocate more to food (2%) and education
(13%) but less to children’s clothing (15%) than their college-educated counterparts. High-school-
educated male heads whose wives are present, on the other hand, allocate less to alcohol and tobacco
(5%) than their college-educated counterparts. These findings lead to the conjecture that the less
formally educated household heads value the welfare of the household members more than the more
formally educated heads.

4.2.3. Heterogeneity across household types with college-educated household heads. Table 6, panel
C shows the results of testing Equations (A13) to (A18), presented in the Online Appendix, to analyse
heterogeneity in the allocations of remittances across households with college-educated heads. The
findings reveal a gender differential in the allocations for alcohol and tobacco regardless of the
presence of the spouse: the two household structures headed by males allocate more to alcohol and
tobacco than the two types of households headed by females.
Heterogeneity in the intrahousehold allocation of international remittances 873

In addition, the allocations to adult clothing and personal items are higher for de jure female heads
than for female heads with OFW spouses and male heads whose wives are present. This may imply
that whereas alcohol and tobacco are considered vices for men, clothing is a vice for women, at least
for college-educated de jure female heads (Handa, 1996).

4.2.4. Heterogeneity across household types with high school-educated household heads. Panel D of
Table 6 presents the results of post GLM estimation analysis to test Equations (A19) to (A21), shown
in the Online Appendix, which compare female heads with OFW spouses and de jure female heads,
female heads with OFW spouses and male heads with OFW spouses, and de jure female heads and
male heads with OFW spouses respectively. Table 5, on the other hand, shows the marginal effects of
the interaction terms of the three household types with absent spouses and remittances, which are the
results of testing Equations (A22) to (A24), presented in the Online Appendix, and may be interpreted
as heterogeneity in remittance allocation between these household types and the omitted household
type (male heads with wives present) conditional on high school education.
High-school-educated female heads show differences in allocation: female heads with OFW spouses
allocate 1 per cent more to food, 13 per cent less to children’s clothing and 8 per cent less to alcohol
and tobacco than de jure female heads as remittances increase.
The results also reveal that female heads with OFW spouses and with a high school education at
most have a preference towards spending on food and education, while they spend less on children’s
clothing and on alcohol and tobacco than either type of male head.

5. Conclusion
The primary goal of this article is to test how the intrahousehold allocation of remittances sent by
household members who work as OFWs differs across Philippine household types after controlling for
gender and educational attainment of heads and the absence of the spouse to reveal differences in
preferences and to account for moral hazard. It is important to incorporate migration into the intrahouse-
hold allocation framework, given that it impacts the dynamics of authority over resources, and hence their
allocation, as well as changing the demographic composition and economic status of the households.
Controlling for location, income effects and household composition, the analyses in this article offer
findings that support the claims in the existing intrahousehold allocation literature: that female heads
allocate more to goods that improve the welfare of the children and male heads allocate more to vices,
with caveats.
First, regardless of educational attainment and the presence or absence of a spouse, male heads
allocate more to alcohol and tobacco than female heads with OFW spouses do. In terms of goods that
improve the welfare of children, such as education and children’s clothing, the presence of a female
decision-maker in male-headed households matters: when the wife is present, more is spent on these
goods than is the case when the wife is absent. This finding supports previous literature that suggests
that the presence of a female authority in the household can positively influence the allocations for
children (Handa, 1994, 1996). This result may also reflect moral hazard and imperfect monitoring by
OFW wives, who may prefer but be unable to ensure that their remittances be allocated to goods that
improve the welfare of their children.
Second, this article also finds that heterogeneity in resource allocation exists within subgroups of
female heads. Among the more-educated female heads, de jure heads allocate more to adult clothing,
but among the less-educated female heads, they allocate more to children’s clothing and alcohol and
tobacco than the female heads with OFW spouses do. Third, it appears that household heads with less
formal education allocate more of remittance income to education; this is true for all female-headed
households and male-headed households with wives present. Fourth, educational attainment of the
head also matters within male-headed households with wives present in terms of allocating remittances
to alcohol and tobacco: those who have less education allocate less to these goods than their college-
educated counterparts.
874 M. Pajaron

Future research on migration and intrahousehold allocation should do more to address the endo-
geneity of household structures. Possible instrumental variables that could be explored, when data
become available, include the socioeconomic status of household heads and rainfall data before
household formation. Another aspect to be explored is the relationship of heads with other members
of the household and the roles of other members in supplementing the income of the household and in
escaping poverty. In addition, although this article is a step towards incorporating migration and
remittances into the intrahousehold allocation framework, more research is needed to study the links
between migration, household structure, poverty and children’s welfare. Better understanding of these
factors’ effects may contribute to developing ways of helping households escape poverty and
improving children’s human capital in the long run.

Notes
1. Yoong, Rabinovich, and Diepeveen (2012) provide a systematic review of the literature related to the differences in the
welfare of a household conditional on the gender of the recipient of economic resources. While some of the earlier studies
are observational and retrospective cross-sectional studies, some are quasi-experimental.
2. Moral hazard is defined as the inability of one party to monitor or observe the actions of the other party. In the case of
migration and remittances, OFWs cannot effectively and perfectly monitor the decisions of household heads in terms of how
remittances are allocated and spent.
3. Author’s computation from the 2009 Survey on Overseas Filipinos dataset.
4. Author’s computation from the 2003 Family Income and Expenditures Survey dataset.
5. FIES defines the household head as the one who manages the finances of the family; in this sense, headship indicates control
of allocation decisions.
6. The selection of a base or reference category of a categorical variable is based on the number of observations.
7. All the F-statistics of the instrumental variables fall below the threshold of 16, signifying weak instrument problems (Staiger &
Stock, 1997; Stock & Yogo, 2002). The F-statistic of excluded instrumental variables for remittances is 4.69, while for female-
headed households with OFW spouses it is 9.43, for de jure female-headed households it is 5.53, and for male-headed
households with OFW spouses it is 7.17. Instrumental variable analyses, therefore, were not pursued. In addition, lack of
data limits the candidates for valid instrumental variables and the use of other methods to address the identification issue.
8. Baum (2008) noted that using binomial distribution can be a good choice even if the dependent variable is continuous.
9. These results may be country specific, and thus care must be taken in generalising them or using them for cross-country
comparison.
10. Handa (1996) showed for Jamaica that whereas morbidity rates among children are higher in male-headed households,
female-headed households, particularly those headed by single females, allocate less to health services, which may be
attributable to the demand of female-headed households for preventive health care. The findings suggest that although
female-headed households are more vulnerable in terms of access to resources, their children are no worse off because of the
resources allocated to them and the efficient use of health inputs.

Disclosure statement
No potential conflict of interest was reported by the author.

References
Adato, M., de la Brière, B., Mindek, D., & Quisumbing, A. R. (2000). The impact of PROGRESA on women’s status and
intrahousehold relations (International Food Policy Research Institute Final Report). Retrieved from http://citeseerx.ist.psu.
edu/viewdoc/download?doi=10.1.1.225.5155&rep=rep1&type=pdf
Appleton, S. (1996). Women-headed households and household welfare: An empirical deconstruction for Uganda. World
Development, 24, 1811–1827. doi:10.1016/S0305-750X(96)00089-7
Baum, C. F. (2008). Stata tip 63: Modeling proportions. The Stata Journal, 8, 299–303.
Chant, S. H., & McIlwaine, C. (1995). Women of a lesser cost: Female labour, foreign exchange, and Philippine development.
London: Pluto Press.
Chen, J. J. (2006). Migration and imperfect monitoring: Implications for intrahousehold allocation. American Economic Review,
96, 227–231. doi:10.1257/000282806777212468
DeGraff, D. S., & Bilsborrow, R. E. (1993). Female-headed households and family welfare in rural Ecuador. Journal of
Population Economics, 6, 317–336. doi:10.1007/BF00599041
Heterogeneity in the intrahousehold allocation of international remittances 875

Duflo, E. (2003). Grandmothers and granddaughters: Old age pension and intrahousehold allocation in South Africa. World Bank
Economic Review, 17(1), 1–25. doi:10.1093/wber/lhg013
Eder, J. (2006). Gender relations and household economic planning in the rural Philippines. Journal of Southeast Asian Studies,
37, 397–413. doi:10.1017/S0022463406000701
Eviota, E. (1986). The articulation of gender and class in the Philippines. In E. Leacock & H. Safa (Eds.), Women’s work:
Development and the division of labor by gender (pp. 194–206). South Hadley, MA: Bergin and Garvey.
Guzmán, J. C., Morrison, A. R., & Sjöblom, M. (2008). The impact of remittances and gender on household expenditure
patterns: Evidence from Ghana. In A. R. Morrison, M. Schiff, & M. Sjöblom (Eds.), The international migration of women
(pp. 125–152). Washington, DC: World Bank and Palgrave Macmillan.
Handa, S. (1994). Gender, headship and intrahousehold resource allocation. World Development, 22, 1535–1547. doi:10.1016/
0305-750X(94)90036-1
Handa, S. (1996). Expenditure behavior and children’s welfare: An analysis of female headed households in Jamaica. Journal of
Development Economics, 50, 165–187. doi:10.1016/0304-3878(96)00008-9
Hoddinott, J., & Haddad, L. (1995). Does female income share influence household expenditures? Evidence from Côte d’Ivoire.
Oxford Bulletin of Economics and Statistics, 57, 77–96. doi:10.1111/obes.1995.57.issue-1
Illo, J. F. (1995). Who heads the households in the Philippines? In A. T. Torres (Ed.), The Filipino woman in focus: A book of
readings (2nd ed., pp. 235–254). Quezon City: University of the Philippines, Office of Research Coordination.
International Monetary Fund. (2010). Balance of payments statistics yearbook. Retrieved from http://data.imf.org/?sk=
7A51304B-6426-40C0-83DD-CA473CA1FD52
Israel-Sobritchea, C. (1994). Gender roles and economic change in a fishing community in Central Visayas. In I. Ushijima & C.
N. Zayas (Eds.), Fishers of the Visayas: A study of Visayan maritime communities (pp. 279–303). Quezon City: College of
Social Sciences and Philosophy Publications, University of the Philippines.
Jefremovas, V. (2000). Women are good with money: The impact of cash cropping on class relations and gender ideology in
Northern Luzon, Philippines. In A. Spring (Ed.), Women farmers and commercial ventures: Increasing food security in
developing countries (pp. 131–150). Boulder, CO: Lynne Rienner.
Joshi, S. (2004). Female household-headship in rural Bangladesh: Incidence, determinants and impact on children’s schooling
(Yale University Economic Growth Center Discussion Paper No. 894). Retrieved from http://ssrn.com/abstract=588345
Malone, L. (2007). Migrants’ remittances and investments in children’s human capital: The role of asymmetric preferences in
Mexico. University of California–Santa Cruz Center for Global, International, and Regional Studies. Retrieved from http://
www.escholarship.org/uc/item/23n6s2p3
Papke, L. E., & Wooldridge, J. M. (1996). Econometric methods for fractional response variables with an application to 401(K)
plan participation rates. Journal of Applied Econometrics, 11, 619–632. doi:10.1002/(ISSN)1099-1255
Quisumbing, A. R., & De La Brière, B. (2000). Women’s assets and intrahousehold allocation in rural Bangladesh: Testing
measures of bargaining power (International Food Policy Research Institute Food Consumption and Nutrition Division
Discussion Paper No. 86). Retrieved from http://www.ifpri.org/sites/default/files/publications/fcndp86.pdf
Quisumbing, A. R., & Maluccio, J. A. (2003). Resources at marriage and intrahousehold allocation: Evidence from Bangladesh,
Ethiopia, Indonesia and South Africa. Oxford Bulletin of Economics and Statistics, 65, 283–327. doi:10.1111/obes.2003.65.
issue-3
Rubalcava, L., Teruel, G., & Thomas, D. (2004). Spending, saving and public transfers paid to women (University of California–
Los Angeles California Center for Population Research Online Working Paper No. 024). Retrieved from http://www.
escholarship.org/uc/item/95m9f476
Senauer, B., Garcia, M., & Jacinto, E. (1988). Determinants of the intrahousehold allocation of food in the rural Philippines.
American Journal of Agricultural Economics, 70, 170–180. doi:10.2307/1241987
Skoufias, E., & McClafferty, B. (2001). Is PROGRESA working? Summary of the results of an evaluation by IFPRI
(International Food Policy Research Institute Food Consumption and Nutrition Division Discussion Paper No. 118).
Retrieved from http://ageconsearch.umn.edu/bitstream/16418/1/fc010118.pdf
Staiger, D., & Stock, J. (1997). Instrumental variables regression with weak instruments. Econometrica, 65, 557–586.
doi:10.2307/2171753
Stock, J., & Yogo, M. (2002). Testing for weak instruments in linear IV regression (National Bureau of Economic Research
Technical Working Paper No. 284).
Thomas, D. (1990). Intra-household resource allocation: An inferential approach. The Journal of Human Resources, 25, 635–
664. doi:10.2307/145670
Thomas, D., & Chen, C. L. (1994). Income shares and shares of income: Empirical tests of models of household resource
allocations (RAND Corporation Labor and Population Working Paper Series 94-08). Retrieved from http://www.rand.org/
content/dam/rand/pubs/drafts/2005/DRU705.pdf
Wooldridge, J. M. (2006). Introductory econometrics: A modern approach. Mason, OH: South-Western Cengage Learning.
World Bank. (2015). Personal remittances received. Retrieved from http://data.worldbank.org/indicator/BX.TRF.PWKR.DT.GD.ZS
Yang, D. (2008). International migration, remittances, and household investment: Evidence from Philippine migrants’ exchange
rate shocks. The Economic Journal, 118, 591–630. doi:10.1111/j.1468-0297.2008.02134.x
Yoong, J., Rabinovich, L., & Diepeveen, S. (2012). The impact of economic resource transfers to women versus men: A
systematic review (Technical report). London: EPPI-Centre, Social Science Research Unit, Institute of Education, University
of London.
Copyright of Journal of Development Studies is the property of Routledge and its content may
not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for
individual use.

Potrebbero piacerti anche