Printer Friendly

Endnotes.

(1) This draws heavily from Behrman (1997), Haddad, Hoddinott, and Alderman (1997), Strauss and Beegle (1996), and Strauss and Thomas (1995).

(2) We denote vectors in bold (e.g. x); individual elements of the vector, such as [x.sub.j] are in ordinary type.

(3) That is not to say that the identity of the income earner is unimportant in unitary models. If the price of leisure, (wage), of m and f differ then changes to the wage of either is likely to have different demand impacts for other goods due to gender specific cross price effects with wages.

(4) Perhaps the most persuasive attempt to resolve the problems of aggregation and enforcement is Becker's "rotten-kid theorem" (1974, 1981).

(5) A Pareto optimal allocation is reached when one individual within the household can only be made better off at the expense of another household member.

(6) The distribution of gains within marriage is a common application of cooperative models. However, it is possible that individuals (particularly females) may not have a choice about getting married or forming a household. One can argue that, in many contexts, the decision to marry or form a new household may be motivated by non-economic factors, such as society's views of individuals who do not marry.

(7) The Nash-bargained solution can also be reached through more complex negotiating procedures. Under quite general conditions, Harsanyi and Selten (1987) show that a sequential bargaining process converges to the Nash-bargained solution, if one exists.

(8) The discussion in the next two sections follows Thomas and Chen (1994) and Bourguinon, Browning, Chiappori, and Lechene (1993) closely.

(9) Using data from France and Canada, Bourguinon, Browning, Chiappori and Lechene (1993, 1995) find that the ratio of income effects is not unity, and so reject income pooling. However, the ratios are constant across a range of commodities, so the data are consistent with Pareto efficiency. Similar evidence for Taiwan is found in Thomas and Chen (1994).

(10) The tests are based on the structure of what they call the pseudo-Slutsky matrix, derived in the paper.

(11) This is not equivalent to a rejection of the "unitary model per se. Non zero effects are consistent with "altruism" or "dictatorial" models within the unitary framework.

(12) Manser and Brown (1980) and McElroy and Horney (1981) interpret the threat point as the utility associated with divorce, while in noncooperative models, e.g., Lundberg and Pollak (1993), the threat point could correspond to a noncooperative outcome within marriage, such as reverting to traditional gender roles.

(13) This draws heavily on Jean-Pierre Habicht's presentation during the External Advisory Committee of the USAID/WID project, "Strengthening Development Policy through Gender Analysis", May 20-21, 1999.

(14) The value of kin support is illustrated by Bangladeshi sisters' giving up their share of land inheritance in return for their brothers' support (Subramanian 1998). The assurance of their brothers' support clearly has an economic value for these women.

(15) One of her sensitivity tests uses a specification with the percentage of land owned by women, but land ownership by women in Ghana may be endogenous to marriage. In Western Ghana, Quisumbing et al. (1998) show that women obtain strong individualized land rights, equivalent to private ownership, from their husbands if they help in establishing cocoa fields. Husbands "give" their wives land to circumvent traditional inheritance practices among the matrilineal Akan.

(16) Admittedly, a potential heir could exert much power within his or her household, but the threat of disinheritance does exist.

(17) See, for example, Raos (1997) analysis of wife-beating in South India, which uses a combination of qualitative and econometric methods.

(18) In Bangladesh, for example, where related households (the bari) typically live around a common yard, land ownership and education of the head in origin households affect educational attainment of children in partitioned households (Foster 1993).

(19) We do include husband's and wife's education in the instrument set for the Bangladesh regressions.

(20) Families with at least two children are included so that sex dummies are relevant in the family fixed effects specification. The fixed effects procedure eliminates selectivity bias since family size, which affects selection into the sample, is a family-specific variable.

(21) This draws heavily on Quisumbing and de la Briere (1999).

(22) This obviously does not address issues of household formation and dissolution (Foster and Rosenzweig 1998), nor the possibility that intact couples are those where bargaining has been "successful". We do not deal with the sample selection biases introduced by analyzing only intact couples. While we attempted to construct a similar set of variables for deceased, absent, or divorced spouses by recall, these measures are less reliable than those collected for "intact" couples.

(23) This information was complimented with the data on inherited assets. When inheritance happened prior to the marriage, these assets were added to the premarriage assets if not reported in the corresponding module.

(24) A bride will typically bring stores of grain and other food items with her when she moves to her in-laws' compound.

(25) We administered the module on transfers at marriage only to wives, to avoid overloading the male respondents' questionnaire. The wife was asked about five categories of transfers: to the bride, to the groom, to the couple jointly, to the bride's family, and to the groom's family. The practice of interviewing only the wife about transfers at marriage is consistent with work by Rao (1996, 1997), who suggests that women often have better recall of these transfers, since marriage is the most important event at which assets are transferred to women. (Men, on the other hand, may receive sizable transfers at the death of a parent.)

(26) This definition is consistent with that of Bloch and Rao (1996), who define dowry as a groom-price, a payment in cash and/or kind directly made from a brides family to a grooms.

(27) This phenomenon is also largely reported in India (see Rao (1997), Bloch and Rao (1996) for references).

(28) In the first stage regressions, the F-test statistics on the exclusion restrictions for the instruments were as follows: Ln per capita expenditure =22.58; Ln per capita expenditure squared=21.02; Ln household size=13.54; Ln husband's assets at marriage=7.47; and Ln wife's assets at marriage=22.27; all with a p-value of [0.000].

(29) According to the survey design, the first and the fourth rounds were conducted at a years interval, so expenditure patterns in the fourth round are expected to be very similar to that in the first round.

(30) The retrospective survey on inheritance was patterned after a similar survey in the Philippines (Quisumbing 1994).

(31) We also included a module on transfers at marriage, but we have not analyzed these data yet. Since land is the major asset transferred at marriage in these societies, the bias due to using data on land alone is unlikely to be large.

(32) In 1991, a coalition of opposing guerilla forces, the Ethiopian People? s Revolutionary Democratic Front (EPRDF) defeated the communist-led Derg regime, which had been in power for 17 years. In addition, the 30-year war between Ethiopia and Eritrea ceased in 1993 when Ethiopia recognized Eritrean independence though it began again this year.

(33) The 1997 ERHS was undertaken by the Department of Economics, Addis Ababa University (AAU), in collaboration with IFPRI and the Center for the Study of African Economies (CSAE), Oxford University. The 1997 survey built on a panel survey conducted by AAU and CSAE in 1994/95 but these earlier rounds are not used in the present analysis.

(34) In the first stage equation for husband's assets the F-test on the exclusion restriction for the instruments is F(6,1378)=3.1 with p-value [0.005] and for wife's assets is F(6,1378)=2.6 with p-value [0.015].

(35) The 1998 re-survey examined here was directed by a consortium comprising the University of Natal, the University of Wisconsin, and the International Food Policy Research Institute.

(36) "African" here excludes the "colored", Indian (or Asian), and white populations.

(37) In the first stage equation for husbands assets the F-test on the exclusion restriction for the instruments is F(12,475)=2.4 with p-value [0.005] and is the same for wife's assets. For the first stage equation of husbands assets interacted with Indian the F-test is F(12,475)=1.9 with p-value [0.039] and for wife's F(12, 475)=1.2 with p-value[0.289].

(38) The exception is Sumatra, where the Hausman test does not suggest that fixed effects are important. In Bangladesh, we only weakly reject random effects in favor of fixed effects.
Table 1. Assets at marriage and human capital of husband and wife

 Husband Wife

 Mean Standard Mean Standard
 deviation deviation

Bangladesh

Assets at marriage (1996 Taka) 81929 145584 7064 8472
Transfers at marriage (1996
 Taka) 4053 15014 5856 11646
Years of schooling 3.18 4.03 1.68 2.78

Sumatra

Area of paddy land (hectares) 0.18 0.3 0.25 0.62
Area of forestry land
 (hectares) 0.9 1.64 0.42 1.03
Years of schooling 6.83 3.6 6.23 3.4

Ethiopia

Land & Livestock Assets at
 marriage (1997 birr) 2739 7188 461 2023
Total Assets at marriage (1997
 birr) 4194 8272 978 2424
Years of schooling 1.9 1.9 1.3 1.1

South Africa
Count of assets at marriage 1.99 1.98 0.8 1.21
Years of schooling 5.57 5 1.67 2.67

Table 2. Bangladesh: Expenditure shares as a function of assets at
marriage, 2SLS estimates (n = 1920)

Assets at marriage in 1996 Taka; regressions on rounds 2, 3, and 4;
t-statistics in bold are significant at 10% or better

 Food Health

F 13.68 2.56
p-value 0 0

 Coef t-stat Coef t-stat

Endogenous regressors
Ln per capita expenditure 34.24 1.16 -21.84 -1.70 *
Ln per capita expenditure
 squared -3.75 -1.62 1.85 1.84 *
Ln household size -2.84 -2.67 * 0.51 1.11
Ln (hus assets +1) 0.42 1.29 0.18 1.28
Ln (wife's assets +1) 0.00 0.00 -0.34 -1.63
Household demographics
Share females 20-65 -8.07 -1.69 * -1.09 -0.53
Share males 10-19 -5.02 -1.38 -0.76 -0.48
Share females 10-19 -4.81 -1.19 -0.30 -0.17
Share males 6-9 -3.64 -0.83 0.82 0.43
Share females 6-9 -3.05 -0.67 0.03 0.02
Share males 0-5 0.04 0.01 1.46 0.79
Share females 0-5 -8.93 -2.05 * 1.44 0.76
Share males 65+ -6.19 -0.81 0.77 0.23
Share females 65+ -0.87 -0.12 -8.10 -2.58 *

Site and round dummies
Saturia -3.65 -4.62 * -0.68 -1.99 *
Jessore 3.48 3.61 * 0.40 0.95
Round 3 1.26 1.57 -1.81 -5.17 *
Round 4 1.06 1.36 -1.23 -3.64 *
Constant 7.64 0.08 68.63 1.69 *

F-tests (p-values)
Husband's assets=Wife's assets 0.44 0.508 3.52 0.061 *
Overid test(chi-square,p-value) 54.72 1 37.71 1

 Children's
 Education clothing

F 17.03 11.23
p-value 0 0

 Coef t-stat Coef t-stat

Endogenous regressors
Ln per capita expenditure 8.68 1.04 1.76 0.46
Ln per capita expenditure
 squared -0.42 -0.64 -0.11 -0.36
Ln household size 0.57 1.92 * -0.09 -0.64
Ln (hus assets +1) -0.04 -0.41 0.03 0.68
Ln (wife's assets +1) 0.34 2.53 * 0.15 2.50 *
Household demographics
Share females 20-65 -2.08 -1.54 0.08 0.14
Share males 10-19 4.09 3.98 * 1.70 3.60 *
Share females 10-19 2.51 2.21 * 3.01 5.74 *
Share males 6-9 1.41 1.14 2.46 4.31 *
Share females 6-9 0.07 0.05 3.11 5.26 *
Share males 0-5 0.94 0.79 2.59 4.74 *
Share females 0-5 0.82 0.67 2.88 5.12 *
Share males 65+ 8.10 3.74 * 2 99 3.01 *
Share females 65+ 1.80 0.89 -0.29 -0.31

Site and round dummies
Saturia 0.87 3.91 * 0.42 4.13 *
Jessore -0.17 -0.63 0.30 2.39 *
Round 3 -0.58 -2.57 * 0.61 5.82 *
Round 4 -0.76 -3.46 * 0.17 1.70 *
Constant -39.51 -1.50 -8.63 -0.71

F-tests (p-values)
Husband's assets=Wife's assets 4.4 0.036 * 2.33 0.127
Overid test(chi-square,p-value) 45.74 1 33.88 1

 Cigarettes

F 5.58
p-value 0

 Coef t-stat

Endogenous regressors
Ln per capita expenditure 4.82 1.00
Ln per capita expenditure
 squared -0.44 -1.16
Ln household size -0.54 -3.11 *
Ln (hus assets +1) 0.07 1.36
Ln (wife's assets +1) -0.06 -0.79
Household demographics
Share females 20-65 -0.10 -0.12
Share males 10-19 -0 76 -1 28
Share females 10-19 -0.62 -0.93
Share males 6-9 -1.04 -1.44
Share females 6-9 -0.87 -1.17
Share males 0-5 -2 16 -3.12 *
Share females 0-5 -1.55 -2.18 *
Share males 65+ -3.15 -2.50 *
Share females 65+ -1.57 -1.33

Site and round dummies
Saturia 0.35 2.74 *
Jessore -0.27 -1.70 *
Round 3 -0.33 -2.48 *
Round 4 -0.32 -2.53 *
Constant -8.31 -0.54

F-tests (p-values)
Husband's assets=Wife's assets 1.65 0.199
Overid test(chi-square,p-value) 70.76 1

Instruments: Round 1 values: In per capita expenditure, In per capita
expenditure squared, In household size; for both husband and wife:
dummies for schooling (primary, secondary, university (husband only),
age and age squared, birth order, family background: father's
schooling, mother's schooling or literacy, parent's land, number of
siblings, number of living brothers, year of marriage (see Tables
5-8).

Note: Assets at marriage in 1996 Taka; regressions on rounds 2, 3,
and 4; t-statistics in bold are significant at 10% or better are
indicated with *.

Table 3. Sumatra expenditure shares regressions; tobit estimates
(n=114)

 Food Health

LR chi2(17) 148.12 37.01
p-value 0.00 0.0034

 Coef t-stat Coef t-stat

Ln total expenditure per capita 0.319 0.777 0.199 0.859
Ln tot exp per capita squared -0.019 -1.312 -0.007 -0.846
Ln household size -0.180 -5.739 -0.009 -0.516

Household composition (males 20-65 excluded)

Males 0-5 0.080 0.830 0.014 0.264
Females 0-5 0.109 0.970 0.003 0.049
Males 6-9 0.060 0.505 0.028 0.424
Females 6-9 0.072 0 694 0.053 0.910
Mates 10-19 0.067 0.704 0.057 1.060
Females 10-19 0.125 1.383 -0.006 -0.115
Females 20-65 -0.157 -1.315 0.061 0.898
Males 65+ -0.008 -0.092 0.170 3.418
Females 65+ -0.112 -0,805 -0.110 -1.405

Landholdings at time of marriage (hectares)

Husband's paddy land -0.073 -1.874 -0.020 -0.901
Husband's forestry land 0.012 0.881 -0.001 -0.138
Wife's paddy land 0.014 0.471 0.037 2.153
Wife's forestry land 0.014 0.669 0.001 0.062
Kerinci dummy 0.055 2.978 -0.013 -1.202
Constant -0.013 -0.004 -1.421 -0.860
Sigma 0.075 0.043

F tests (p-value):
Husband's paddy=wife's paddy 3.150 0.079 4.170 0.044
Husband's forestry land=wife's
 forestry land 0.000 0.950 0.020 0.900

Share uncensored 1 1

 Children's
 Education clothing

LR chi2(17) 69.13 103.03
p-value 0 0

 Coef t-stat Coef t-stat

Ln total expenditure per capita 0.209 1.662 0.365 2.116
Ln tot exp per capita squared -0.008 -1.694 -0.013 -2.173
Ln household size 0.016 2.150 0.004 0.410

Household composition (males 20-65 excluded)

Males 0-5 0.024 1.121 0.113 4.101
Females 0-5 -0.001 -0.050 0.108 3.409
Males 6-9 0.074 2.893 0.156 4.750
Females 6-9 0.039 1.720 0.142 4.892
Mates 10-19 0.051 2.371 0.099 3.577
Females 10-19 0.074 3.637 0.113 4.339
Females 20-65 0.032 1.189 0.044 1.279
Males 65+ -0.028 -0.960 -0.013 -0.369
Females 65+ 0.029 0.881 0.000 -0 004

Landholdings at time of marriage (hectares)

Husband's paddy land 0.011 1.269 0.015 1.390
Husband's forestry land -0.008 -2.507 -0.002 -0.614
Wife's paddy land 0.014 2.312 0.002 0.236
Wife's forestry land -0.003 -0730 -0.005 -0.793
Kerinci dummy 0.003 0.673 -0.004 -0.769
Constant -1.482 -1.674 -2.519 -2.082
Sigma 0.014 0.019

F tests (p-value):
Husband's paddy=wife's paddy 0.140 0.712 0.910 0.343
Husband's forestry land=wife's
 forestry land 0.610 0.435 0.090 0.759

Share uncensored 0.702 0.8421

 Tobacco

LR chi2(17) 29.44
p-value 0.0307

 Coef t-Stat

Ln total expenditure per capita 1.544 3.184
Ln tot exp per capita squared -0.056 -3.218
Ln household size 0.070 2.778

Household composition (males 20-65 excluded)

Males 0-5 -0.009 -0.123
Females 0-5 -0.006 -0.068
Males 6-9 0.089 1.002
Females 6-9 -0.007 -0.095
Mates 10-19 -0.048 -0.679
Females 10-19 -0.002 -0.029
Females 20-65 0.065 0.729
Males 65+ -0.065 -0.806
Females 65+ -0 051 -0 461

Landholdings at time of marriage (hectares)

Husband's paddy land 0.052 1.800
Husband's forestry land -0.011 -0.984
Wife's paddy land -0.009 -0.415
Wife's forestry land -0.002 -0.130
Kerinci dummy 0.010 0.722
Constant -10.699 -3.157
Sigma 0.053

F tests (p-value):
Husband's paddy=wife's paddy 2.800 0.097
Husband's forestry land=wife's
 forestry land 0.200 0.656

Share uncensored 0.772

Table 4. Ethiopia Expenditure Share Regressions (Two Stage Least
Squares) N=1418

 Food Education

 coeff t-stat coeff t-stat

Ln PCE -7.602 -1.2 0.820 1.1
Ln PCE 2 0.656 1.4 -0.061 -1.1
Ln HH Size 1.003 0.9 0 144 1.5
Household Composition (Males 16-50 excluded)
Females 1-5 -4.945 -1.0 -0.661 -2.1
Males 1-5 1.803 0.3 -0.902 -2.4
Females 6-15 -2.200 -0.6 0.447 1.5
Males 6-15 5.112 1.4 0.441 1.2
Females 16-50 -2.482 -0.6 1.172 1.4
Females 50 + 1.352 0.3 0.147 0.4
Males 50 + 11.482 2.4 0.860 1.1
Ethnicity (Tigray excluded)
Amhara 0.592 0.2 -0.166 -0.4
Oromo 3.421 0.9 -0.470 -1.0
South-Central -1.028 -0.3 0.188 0.4
Other/mixed 4.489 1.4 -0.250 -0.7
Not identified -4.228 -0.7 0.032 0.1
Religion (Orthodox excluded)
Muslim -0.186 -0.1 -0.042 -0.2
Other Christian 5.669 3.0 0.394 0.8
Other 0.669 0.2 -0.031 -0.1
Not identified 6.114 1.0 -0.249 -0.4
Ln Value assets at marriage
Husband -1.515 -1.5 -0.037 -0.3
Wife 2.648 1.8 -0.041 -0.3
F test overall 23.5 [0.00] 3.8 [0.00]
F test Ln PCE quadratic 2.9 [0.06] 1.3 [0.28]
F test husband = wife 4.1 [0.04] 0.0 [0.99]
Fraction Not Censored 1.00 0.26
Average Budget Share 0.74 0.01

 Health Child Clothing

 coeff t-stat coeff t-stat

Ln PCE 0.161 0.1 4.106 2.8
Ln PCE 2 -0.026 -0.2 -0.315 -3.1
Ln HH Size -0.685 -2.1 1.196 4.4
Household Composition (Males 16-50 excluded)
Females 1-5 -1.781 -1.0 -3.491 -2.8
Males 1-5 -0.944 -0.6 -3.920 -2.8
Females 6-15 0.976 0.8 0.593 0.5
Males 6-15 -1.349 -1.4 -0.486 -0.5
Females 16-50 -0.162 -0.2 -0.496 -04
Females 50 + -1.238 -1.0 -1.986 -1.7
Males 50 + -0.274 -0.2 -2.230 -2.3
Ethnicity (Tigray excluded)
Amhara -0.462 -0.8 -1.017 -1.5
Oromo 0.647 0.8 -1.254 -1.7
South-Central -0.131 -0.2 -0.818 -0.9
Other/mixed 0.082 0.1 -1.670 -2.4
Not identified 1.806 0.7 1.193 0.4
Religion (Orthodox excluded)
Muslim 0.571 0.7 -0.104 -0.2
Other Christian -0.651 -1.1 -0.399 -0.9
Other 0.031 0.0 0.745 1.0
Not identified -2.847 -1.1 -1.318 -0.4
Ln Value assets at marriage
Husband -0.350 -1.2 -0.207 -0.7
Wife 0.624 1.4 -0.228 -0.6
F test overall 4.3 [0.00] 7.3 [0.00]
F test Ln PCE quadratic 0.5 [0.61] 7.9 [0.00]
F test husband = wife 2.2 [0.14] 0.0 [0.97]
Fraction Not Censored 052 0.53
Average Budget Share 0.02 0.02

 Alcohol & tobacco

 coeff t-stat

Ln PCE 6.601 1.8
Ln PCE 2 -0.395 -1.5
Ln HH Size 0.642 0.9
Household Composition (Males 16-50 excluded)
Females 1-5 7.294 2.2
Males 1-5 3.427 1.0
Females 6-15 2.102 0.8
Males 6-15 0.929 0.4
Females 16-50 1.550 0.6
Females 50 + 4.170 1.4
Males 50 + 0.239 0.1
Ethnicity (Tigray excluded)
Amhara 1 586 1.0
Oromo 1.211 0.6
South-Central 3.336 1.6
Other/mixed 2.230 1.2
Not identified -2.905 -0.8
Religion (Orthodox excluded)
Muslim -3.178 -1.6
Other Christian -4.141 -3.2
Other 2.671 1.4
Not identified 4.761 1.4
Ln Value assets at marriage
Husband 1.543 2.4
Wife -1.807 -1.7
F test overall 6.5 [0.00]
F test Ln PCE quadratic 4.9 [0.01]
F test husband = wife 5.3 [0.02]
Fraction Not Censored 0.40
Average Budget Share 0.04

Notes: Regression also includes a constant and dummy control for all
but one of the 15 villages (not shown). Husband and wife asset
variables treated as endogenous. Instruments include parental
education and value of gifts transferred from the groom's to the
bride's family and vice versa.

Table 5: South Africa Expenditure Share Regressions (Two Stage Least
Squares) N=500

 Food Education

 coeff t-stat coeff t-stat

Ln PCE 3.266 0.3 -6.072 -1.4
Ln PCE2 -1.433 -1.5 0.551 1.5
Ln HH Size -6.393 -2.8 0.865 1.0

Household Composition (Males 16-50 excluded)
Females 1-5 -2.500 -0.3 5.806 1.5
Males 1-5 -12.313 -1.1 1.653 0.4
Females 6-15 -4.549 -0.6 4.166 1.4
Males 6-15 -1.119 -0.2 0.865 0.3
Females 16-50 4.200 0.7 3.404 1.4
Females 50 + -1.094 -0.1 0.637 0.2
Males 50 + 5.376 0.6 -0.221 -0.1

Assets at marriage
Lnfcount of husb assets at
 marriage) 7.680 0.9 -2.195 -0.7
Ln(count of wife assets at
 marriage) -12.934 -1.3 7.411 1.9

Assets at marriage interacted with Indian
Ln(count of husb assets at
 marriage) -7.512 -0.4 1.775 0.3
Ln(count of wife assets at
 marriage) 41.483 1.3 -8.455 -0.7

F test of model: F(81, 418) = 4.7 [0.00] 1.3 [0.05]
F test Ln PCE quadratic 29.3 [0.00] 1.1 [0.33]
F test husband = wife 2.2 [0.14] 3.1 [0.08]
F test Indian husb = Indian wife 0.6 [0.44] 0.0 [0.96]

Fraction Not Censored 1.00 0.88
Average Budget Share 45.88 3.87

 Health Child Clothing

 coeff t-stat coeff t-stat

Ln PCE -1.247 -0.5 1.039 0.7
Ln PCE2 0.115 0.5 -0.086 -0.6
Ln HH Size -0.630 -1.1 0 411 1.3

Household Composition (Males 16-50 excluded)
Females 1-5 -0.252 -0.1 0.501 0.4
Males 1-5 3.114 1.1 -0.302 -0.2
Females 6-15 0.913 0.5 2.109 1.9
Males 6-15 -1.606 -1.0 1.119 1.3
Females 16-50 -0.540 -0.3 -0.183 -0.2
Females 50 + -2.635 -1.1 0.325 0.2
Males 50 + 4.173 1.9 -2.903 -2.4

Assets at marriage
Lnfcount of husb assets at
 marriage) -0.613 -0.3 0.122 0.1
Ln(count of wife assets at
 marriage) 1.420 0.6 -0.678 -0.5

Assets at marriage interacted with Indian
Ln(count of husb assets at
 marriage) -2.060 -0.5 -2.182 -0.9
Ln(count of wife assets at
 marriage) -3.886 -0.5 2.974 0.6

F test of model: F(81, 418) = 1.2 [0.16] 2.1 [0.00]
F test Ln PCE quadratic 0.1 [0.88] 0.2 [0.78]
F test husband = wife 0.4 [0.56] 0.2 [0.68]
F test Indian husb = Indian wife 0.0 [0.87] 0.7 [0.39]

Fraction Not Censored 0.77 0.80
Average Budget Share 1.54 2.26

 Alcohol & Tobacco

 coeff t-stat

Ln PCE -0.213 0.0
Ln PCE2 0.036 0.1
Ln HH Size -0.048 0.0

Household Composition (Males 16-50 excluded)
Females 1-5 0.290 0.1
Males 1-5 4.965 0.9
Females 6-15 -5.461 -1.4
Males 6-15 -7.742 -2.4
Females 16-50 -7.449 -2.3
Females 50 + -3.315 -0.7
Males 50 + -5.235 -1.2

Assets at marriage
Lnfcount of husb assets at
 marriage) -0.160 0.0
Ln(count of wife assets at
 marriage) 0.655 0.1

Assets at marriage interacted with Indian
Ln(count of husb assets at
 marriage) 3.936 0.4
Ln(count of wife assets at
 marriage) -14.451 -0.9

F test of model: F(81, 418) = 1.5 [0.01]
F test Ln PCE quadratic 0.0 [0.97]
F test husband = wife 0 0 [0.91]
F test Indian husb = Indian wife 0.9 [0.42]

Fraction Not Censored 0.55
Average Budget Share 4.63

Notes: Regression also includes a constant and dummy control for all
but one of the 69 clusters (not shown). Husband and wife asset
variables, and their interactions with Indian, treated as endogenous.
Instruments include parental education, indicators of whether parents
alive at time of marriage, and the value of gifts transferred from the
groom's to the bride's family and vice versa.

Table 6A. Effects of husband's and wife's resources on children's
education, Bangladesh

A. Deviation from cohort means Coresident children ages 6-10

 OLS with robust
 SEs FE

No of observations 1210 1210
F test [p-value] 3.97 [0.000] 1.28 [0.258]
R-squared 0.055

Child characteristics coeff t-stat coeff t-stat

Daughter dummy 0.045 0.664 0.062 0.812
Age 0.003 0.161 -0.019 -0.758
Age squared 0.000 0.152 0.002 1.017
Parents' education
Father's schooling 0.019 1.838
Mother's schooling 0.040 3.023

Parents' assets at marriage
Father's assets 0.000 -0.690
Mother's assets 0.000 -0.010

Interaction terms
Daughter x father's schooling -0.011 -0.929 -0.034 -2.163
Daughter x mother's schooling -0.003 -0.183 0.025 1.116
Daugher x father's assets 0.000 0.894 0.000 -1.314
Daughter x mother's assets 0.000 -0.336 0.000 1.142
Saturia dummy 0.111 2.140
Jessore dummy 0.046 0.928
Constant -0.196 -3.175 0.025 0.419

F-tests on "main effects" (p-value)
Father's schooling=Mother's
 schooling 1.070 0.302
Father's assets=Mother's assets 0.030 0.856

F-test on equality of total effects (p-value)
Father's schooling=Mother's
 schooling 4.320 0.038
Father's assets=Mother's assets 0.450 0.501

F-test on interaction terms with daughter dummy (p-value)
Father's schooling=Mother's
 schooling 2.980 0.085
Father's assets=Mother's assets 1.760 0.185

A. Deviation from cohort means Coresident children ages 11-15

 OLS with robust
 SEs FE

No of observations 1466 1466
F test [p-value] 22.46 [0.000] 1.41 [0.197]
R-squared 0.200

Child characteristics coeff t-stat coeff t-stat

Daughter dummy 0.119 0.409 0.580 1.908
Age 0.057 0.656 0.078 1.133
Age squared -0.001 -0.533 -0.002 -1.458
Parents' education
Father's schooling 0.286 7.652
Mother's schooling 0.174 3.270

Parents' assets at marriage
Father's assets 0.000 -1.054
Mother's assets 0.000 1.009

Interaction terms
Daughter x father's schooling -0.148 -2.827 -0.072 -1.207
Daughter x mother's schooling 0.164 2.232 0.050 0.555
Daugher x father's assets 0.000 -0.009 0.000 -1.743
Daughter x mother's assets 0.000 0.507 0.000 -0.592
Saturia dummy 1.113 3.918
Jessore dummy 0.887 2.863
Constant -2.591 -3.052 -0.620 -0.851

F-tests on "main effects" (p-value)
Father's schooling=Mother's
 schooling 1.940 0.164
Father's assets=Mother's assets 1.170 0.280

F-test on equality of total effects (p-value)
Father's schooling=Mother's
 schooling 4.160 0.042
Father's assets=Mother's assets 1.390 0.239

F-test on interaction terms with daughter dummy (p-value)
Father's schooling=Mother's
 schooling 0.760 0.383
Father's assets=Mother's assets 0.070 0.095

Table 6B. Effects of husband's and wife's resources on children's
education, Bangladesh

B. Years of completed
schooling Coresident children ages 6-10

 Tobit FE

No of observations 1210 1210
F test 99.710 [0.000]
Chi-squared 803.130 [0.000]

Child characteristics coeff t-stat coeff t-stat

Daughter dummy 0.331 1.461 0.099 1.250
Age 2.740 3.707 -0.252 -9.428
Age squared -0.106 -2.398 0.040 16.028

Parents' education
Father's schooling 0.049 1.548
Mother's schooling 0.201 4.233

Parents' assets at marriage
Father's assets 0.000 -0.680
Mother's assets 0.000 -0.431

Interaction terms
Daughter x father's schooling -0.042 -0.870 -0.041 -2.558
Daughter x mother's schooling 0.007 0.094 0.045 1.923
Daugher x father's assets 0.000 0.972 0.000 -1.678
Daughter x mother's assets 0.000 0.152 0.000 0.666
Saturia dummy 0.224 1.145
Jessore dummy 0.555 2.994
Constant -15.983 -5.198 0.209 3.391

F-fesfs on "main effects" (p-value)
Father's schooling=Mother's
 schooling 4.510 0.034
Father's assets=Mother's
 assets 0.120 0.732

F-test on equality of total effects (p-value)
Father's schooling=Mother's
 schooling 5.770 0.016
Father's assets=Mother's
 assets 0.090 0.761

F-fest on interaction terms with daughter dummy (p-value)
Father's schooling=Mother's
 schooling 5.950 0.015
Father's assets=Mother's
 assets 0.830 0.363
Breusch-Pagan LM test
 (p-value) 23.600 0.000
Hausman test, FE vs RE
 (p-value) 11.530 0.117

B. Years of completed
schooling Coresident children ages 11-15

 Tobit FE

No of observations 1466 1466
F test 28.130 [0.000]
Chi-squared 486.710 [0.000]

Child characteristics coeff t-stat coeff t-stat

Daughter dummy 0.133 0.401 0.556 1.783
Age 0.792 9.115 0.804
Age squared -0.014 -7.212 -0.015 -9.591

Parents' education
Father's schooling 0.362 9.756
Mother's schooling 0.196 3.333

Parents' assets at marriage
Father's assets 0.000 -1.005
Mother's assets 0 000 1.314

Interaction terms
Daughter x father's schooling -0.189 -2.843 -0.060 -0.983
Daughter x mother's schooling 0.166 1.618 -0.019 -0.205
Daugher x father's assets 0.000 0.250 0.000 -2.077
Daughter x mother's assets 0.000 0.910 0000 0.317
Saturia dummy 0.387 1.535
Jessore dummy -0.606 -2.469
Constant -6.126 -6.602 -4.184 -5.600

F-fesfs on "main effects" (p-value)
Father's schooling=Mother's
 schooling 3.670 0.056
Father's assets=Mother's
 assets 2.060 0.152

F-test on equality of total effects (p-value)
Father's schooling=Mother's
 schooling 2.120 0 145
Father's assets=Mother's
 assets 2.620 0.106

F-fest on interaction terms with daughter dummy (p-value)
Father's schooling=Mother's
 schooling 0.080 0.774
Father's assets=Mother's
 assets 0.490 0.486
Breusch-Pagan LM test
 (p-value) 20.240 0.000
Hausman test, FE vs RE
 (p-value) 12.760 0.078

Table 7. Effects of husband's and wife's resources on children's
education, Sumatra

Levels estimates are OLS with standard errors corrected for clustering

 Deviation from cohort mean

 Levels FE

No of observations 178 178
F test 2.74 [0.0014] 0.6 [0.7963]
R-squared 0.1716

Child characteristics Coeff t-statistic Coeff t-statistic

Daughter dummy 0.783 1.060 -0.209 -0.314
Age 0.111 0.598 0.016 0.069
Age squared -0.004 -0.548 -0.001 -0.118

Parents' education
Father's schooling 0.065 0.922
Mother's schooling 0.103 0.837

Parents' land
Father's paddy -0.245 -0.294
Father's forest land 0.076 0.817
Mother's paddy 1.737 4.866
Mother's forest land -0.502 -1.473

Interaction terms
Daughter x father's
 schooling -0.021 -0.278 0.016 0.155
Daughter x mother's
 schooling -0.039 -0.381 0.029 0236
Daugher x father's
 paddy -0.686 -0.822 2201 1.361
Daughter x mother's
 paddy -1.648 -4.718 -1.512 -1.051
Daughter x father's
 forest land -0.065 -0.572 0.095 0.405
Daughter x mother's
 forest land 0.830 1.783 0.464 0.523
Kerinci dummy -0.155 -0.347
Constant -1.884 -1.372 -0.101 -0.066

F-tests on "main effects" (p-value)
Father's schooling=
 Mother's school 0.050 0.818
Father's paddy=
 Mother's paddy 5.100 0.027
Father's forest land=
 Mother's forest 2.850 0.095

F-fest on equality of total effects (p-value)
Father's schooling=
 Mother's school 0.050 0.818
Father's paddy=
 Mother's paddy 1.910 0.170
Father's forest land=
 Mother's forest 0.870 0.353

F-fesf on interaction terms with daughter dummy (p-value)
Father's schooling=
 Mother's schooling 0.000 0.948
Father's paddy=
 Mother's paddy 3.430 0.068
Father's forest land=
 Mother's forest land 0.170 0.685
Breusch-Pagan LM test
 (p-value) 24.450 0.000
Hausman test, FE vs RE
 (p-value) 6.140 0730

 Years of schooling completed

 Levels FE

No of observations 178 178
F test 63.14 [0.0000] 37.29 [0.0000]
R-squared 0.7728

Child characteristics Coeff t-statistic Coeff t-statistic

Daughter dummy 0.946 1.120 -0.326 -0.432
Age 1.597 7.857 1.549 5.792
Age squared -0.033 -4.159 -0.031 -3.233

Parents' education
Father's schooling 0.080 1.346
Mother's schooling 0.164 1.404

Parents' land
Father's paddy -0.260 -0.331
Father's forest land 0.053 0.560
Mother's paddy 1.121 2.918
Mother's forest land -0.404 -1.277

Interaction terms
Daughter x father's
 schooling -0.029 -0.392 0.006 0.049
Daughter x mother's
 schooling -0.105 -1.048 0.031 0.223
Daugher x father's
 paddy -1.024 -1.023 1.180 0.644
Daughter x mother's
 paddy -0.985 -2.646 -1.584 -0.972
Daughter x father's
 forest land -0.001 -0.009 0.075 0.282
Daughter x mother's
 forest land 0.740 1.542 0.533 0.530
Kerinci dummy 0.768 1.627
Constant -11.203 -7.017 -8.996 -5.161

F-tests on "main effects" (p-value)
Father's schooling=
 Mother's school 0.370 0.542
Father's paddy=
 Mother's paddy 2.890 0.093
Father's forest land=
 Mother's forest 2.030 0.158

F-fest on equality of total effects (p-value)
Father's schooling=
 Mother's school 0.010 0.941
Father's paddy=
 Mother's paddy 1.950 0.166
Father's forest land=
 Mother's forest 0.660 0.419

F-fesf on interaction terms with daughter dummy (p-value)
Father's schooling=
 Mother's schooling 0.010 0.911
Father's paddy=
 Mother's paddy 1.480 0.228
Father's forest land=
 Mother's forest land 0.200 0.657
Breusch-Pagan LM test
 (p-value) 29.130 0.000
Hausman test, FE vs RE
 (p-value) 8.180 0.516

Table 8: Ethiopia Education Regressions on Children Age 6-19 (Ordinary
Least Squares)

 Deviation from cohort mean

 Level HH f.e.

Child characteristics coeff t-stat coeff t-stat

(1) If female -0.388 -2.0 -0.065 -0.3
Age -0.034 -0.7 -0.047 -1.0
Age squared 0.002 0.8 0.002 0.9

Parent's characteristics
Wife's age 0.042 2.8
Wife's age squared 0.000 -2.6
Wife Education 0.039 0.4
Husband Education 0.366 6.6
Female X Wife Education 0.237 1.5 -0.113 -0.7
Female X Husband
 Education -0.163 -2.2 -0.177 -2.3

Ethnicity (Tigray excluded)
Amhara -0.404 -1.6
Oromo -0.691 -2.7
South-Central -0.422 -1.4
Other/mixed 0.091 0.3
Not identified -1.030 -1.3

Religion (Orthodox excluded)
Muslim -0.709 -4.4
Other Christian -0.073 -0.6
Other -0.130 -0.6
Not identified -0.024 0.0

Assets at marriage
Ln(count of husb assets at
 marriage) -0.010 -0.8
Ln(count of wife assets at
 marriage) -0.023 -1.2
Husb assets X Female 0.016 0.9 0.017 1.0
Wife assets X Female 0.068 3.0 0.064 2.8
F test of model 13.0 [0.00] 5.8 [0.00]
F test husband = wife 0.3 [0.61]
F test Husb assets X Female
 = wife's 2.7 [0.10] 2.1 [0.151

 Years of Schooling

 Level HH f.e.

Child characteristics coeff t-stat coeff t-stat

(1) If female -0.379 -1.9 -0.067 -0.3
Age 0.008 0.2 -0.007 -0.1
Age squared 0.007 3.6 0.008 3.8

Parent's characteristics
Wife's age 0.042 2.8
Wife's age squared 0.000 -2.6
Wife Education 0.050 0.5
Husband Education 0.372 6.7
Female X Wife Education 0.220 1.4 -0.117 -0.7
Female X Husband
 Education -0 166 -2.2 -0.180 -2.4

Ethnicity (Tigray excluded)
Amhara -0.398 -1.5
Oromo -0.672 -2.7
South-Central -0.433 -1.5
Other/mixed 0.099 0.4
Not identified -1.045 -1.4

Religion (Orthodox excluded)
Muslim -0.703 -4.3
Other Christian -0.065 -0.5
Other -0.132 -0.6
Not identified 0.385 0.2

Assets at marriage
Ln(count of husb assets at
 marriage) -0.010 -0.8
Ln(count of wife assets at
 marriage) -0.025 -1.3
Husb assets X Female 0.016 0.9 0.017 1.0
Wife assets X Female 0.073 3.2 0.068 3.0
F test of model 31.8 [0.00] 77.0 [0.00]
F test husband = wife 0.3 [0.56]
F test Husb assets X Female
 = wife's 3.1 [0.081 2.5 [0.121

Notes: Level regressions include a constant and dummy control for all
but one of the 15 villages (not shown). Household fixed effects
regressions contain a dummy for each household.

Table 9: South Africa Education Regressions on Children 16-21
(Ordinary Least Squares)

 Deviation from cohort mean

 Level HH f.e.

 coeff t-stat coeff t-stat

(1) if female -0.235 -0.5 0.975 1.5
age 1.399 0.6 -1.157 -0.5
age squared -0.037 -0.6 0.028 0.4

Wife's age 0.206 1.8
Wife's age squared -0.002 -1.8
Wife Education 0.130 2.5
Husband Education 0.024 0.5
Female X Wife Education -0.136 -1.8 -0.104 -1.1
Female X Husband Education 0.078 1.2 0.136 1.8

Assets at marriage
Ln(count of husb assets at
 marriage) -0.301 -0.9
Lnfcount of wife assets at
 marriage) 0.049 0.1

Husb assets X Indian 0.459 0.6
Wife assets X Indian -0.886 -0.8

Husb assets X Female 0.386 0.9 0.974 1.9
Wife assets X Female -0.117 -0.2 -0.219 -0.4

Husb assets X Indian X Female -0.864 -0.7 -4.387 -1.5
Wife assets X Indian X Female 3.262 1.9 10.050 2.2

F test of model excl.cluster
 dummies 1.9 [0.02] 2.5 [0.01]
F test husband = wife 0.4 [0 54]
F test Indian husb = Indian wife 2.0 [0 16]
F test Husb assets X Female =
 wife's 1.7 [0.19]
F test Indian Husb assets X
 Female = wife's 3.7 [0.06]

 Years of Schooling

 Level HH f.e.

 coeff t-stat coeff t-stat

(1) if female -0.222 -0.4 1.030 1.6
age 2.424 1.1 0.131 0.1
age squared -0.052 -0.9 0.006 0.1

Wife's age 0.207 1.8
Wife's age squared -0.002 -1.8
Wife Education 0.130 2.5
Husband Education 0.024 0.5
Female X Wife Education -0.132 -1.8 -0.110 -1.2
Female X Husband Education 0.070 1.1 0.135 1.8

Assets at marriage
Ln(count of husb assets at
 marriage) -0.284 -0.9
Lnfcount of wife assets at
 marriage) 0.040 0.1

Husb assets X Indian 0.349 0.4
Wife assets X Indian -0.899 -0.8

Husb assets X Female 0.393 1.0 1.031 2.1
Wife assets X Female -0.126 -0.3 -0.204 -0.4

Husb assets X Indian X Female -0.829 -0.7 -5.230 -1.8
Wife assets X Indian X Female 3.435 2.0 11.345 2.5

F test of model excl.cluster
 dummies 3.8 [0.00] 5.3 [0.00]
F test husband = wife 0.3 [0.57]
F test Indian husb = Indian wife 2.3 [0.13]
F test Husb assets X Female =
 wife's 1.9 [0.17]
F test Indian Husb assets X
 Female = wife's 4.9 [0.03]

Notes: Level regressions include a constant and dummy control for all
but one of the 69 clusters (not shown). Household fixed effects
regressions contain a dummy for each household.

Table 10. Summary for all countries

 Bangladesh Indonesia
 Paddy Forest

Food Share
 Wife Assets -- -- --
 Husband Assets -- NEG --
 Wife--Husband Assets -- POS POS

Education Share
 Wife Assets POS POS
 Husband Assets -- -- NEG
 Wife--Husband Assets POS -- --

Health Share
 Wife Assets -- POS --
 Husband Assets -- -- --
 Wife--Husband Assets NEG POS --

Child Clothing Share
 Wife Assets POS -- --
 Husband Assets -- -- --
 Wife--Husband Assets -- -- --

Alcohol and/or Tobacco Share
 Wife Assets -- -- --
 Husband Assets -- POS --
 Wife--Husband Assets -- NEG --

 Ethiopia South Africa
 (Africans only)

Food Share
 Wife Assets POS --
 Husband Assets -- --
 Wife--Husband Assets POS --

Education Share
 Wife Assets -- POS
 Husband Assets -- --
 Wife--Husband Assets -- POS

Health Share
 Wife Assets -- --
 Husband Assets -- --
 Wife--Husband Assets -- --

Child Clothing Share
 Wife Assets -- --
 Husband Assets -- --
 Wife--Husband Assets -- --

Alcohol and/or Tobacco Share
 Wife Assets -- --
 Husband Assets POS --
 Wife--Husband Assets NEG --
COPYRIGHT 1999 The World Bank
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1999 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Intrahousehold Allocation and Gender Relations: New Empirical Evidence
Publication:Intrahousehold Allocation and Gender Relations-New Empirical Evidence
Date:Oct 1, 1999
Words:7370
Previous Article:References.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters