Week 2
Income pooling
Data
Observational data
• Observational data- data in which some individuals possess a characteristic whose e ect we
want to measure, while others do not (if binary), or varies
◦ Use of this has lead to rejection of income pooling
◦ However, this could be due to correlated heterogeneity between choices
Thomas (1990) tried to overcome this by using unearned income, however this still poses
heterogeneity issues
• Outcomes of interest: nutrient intake, child health, survival and fertility
• Income variables: unearned incomes of wife and husband
• Estimated equation: X a Bmym BwytfMte
a
relevant
covariates
priceseducation
age etc
TEST
Emlyn yaw ieBmBw iftruethenthereispooling
• He found that the e ect of unearned income of the mother in the child’s probabilities is almost 20
fold that of the father
Schultz (1990):
• Outcomes of interest: labour supply and fertility
• Income variables: unearned incomes of wife and husband
• Main nding: income pooling is rejected, however lack of estimated equations and tests
Phipps and Burton (1998):
• Outcomes of interest: expenditures
• Income variable: incomes of wife and husband
• Estimated equation: X a BmymtbwywtBmmym tbwwywhtbmwywymtrm.ie
TEST Bm Bw BmmBmw Bmw2
PROOF
ym
Bmt2BmmymtBmwyw
gw Bwt2BwwywtBmwym
ymyw o BmBw Bmwy
ymyutoBmt2Bmmy Bmwy Bwt2Bwwy
AsBmBw 2Bmmy Bmwy2Bwwy Bmwy
2BmmBmw 2BwwBmw
Bmm Bww
youyw Bmt2BmmymtBmwyw Bwt2BwwywtBmwym
zbmmymtbmwyw
2Bmmywtsmwy abmn.com
gu Bmwlym.gs
Bmm B
• Main ndings
◦ Income is pooled for housing and expenditure, and not for other goods
◦ Higher expenditure on child care is associated with higher wife’s income, it’s not higher with
husband’ income
Income pooling
Data
Observational data
• Observational data- data in which some individuals possess a characteristic whose e ect we
want to measure, while others do not (if binary), or varies
◦ Use of this has lead to rejection of income pooling
◦ However, this could be due to correlated heterogeneity between choices
Thomas (1990) tried to overcome this by using unearned income, however this still poses
heterogeneity issues
• Outcomes of interest: nutrient intake, child health, survival and fertility
• Income variables: unearned incomes of wife and husband
• Estimated equation: X a Bmym BwytfMte
a
relevant
covariates
priceseducation
age etc
TEST
Emlyn yaw ieBmBw iftruethenthereispooling
• He found that the e ect of unearned income of the mother in the child’s probabilities is almost 20
fold that of the father
Schultz (1990):
• Outcomes of interest: labour supply and fertility
• Income variables: unearned incomes of wife and husband
• Main nding: income pooling is rejected, however lack of estimated equations and tests
Phipps and Burton (1998):
• Outcomes of interest: expenditures
• Income variable: incomes of wife and husband
• Estimated equation: X a BmymtbwywtBmmym tbwwywhtbmwywymtrm.ie
TEST Bm Bw BmmBmw Bmw2
PROOF
ym
Bmt2BmmymtBmwyw
gw Bwt2BwwywtBmwym
ymyw o BmBw Bmwy
ymyutoBmt2Bmmy Bmwy Bwt2Bwwy
AsBmBw 2Bmmy Bmwy2Bwwy Bmwy
2BmmBmw 2BwwBmw
Bmm Bww
youyw Bmt2BmmymtBmwyw Bwt2BwwywtBmwym
zbmmymtbmwyw
2Bmmywtsmwy abmn.com
gu Bmwlym.gs
Bmm B
• Main ndings
◦ Income is pooled for housing and expenditure, and not for other goods
◦ Higher expenditure on child care is associated with higher wife’s income, it’s not higher with
husband’ income
