session 2: remittances gendered determinants and impacts the impact of remittances and gender on...

17
SESSION 2: REMITTANCES GENDERED DETERMINANTS AND IMPACTS The impact of remittances and gender on household expenditure patterns: Evidence from Ghana Juan Carlos Guzmán Andrew R. Morrison Mirja Sjöblom PREM, Gender and Development Group

Upload: allan-leon-bradley

Post on 23-Dec-2015

246 views

Category:

Documents


1 download

TRANSCRIPT

SESSION 2: REMITTANCES GENDERED DETERMINANTS AND IMPACTS

The impact of remittances and gender on household expenditure patterns: Evidence from Ghana

Juan Carlos GuzmánAndrew R. MorrisonMirja Sjöblom

PREM, Gender and Development Group

MOTIVATIONMOTIVATION

In 2005, the total projected value of remittances In 2005, the total projected value of remittances exceeded $232 billion (Ratha 2005)exceeded $232 billion (Ratha 2005)

Remittances represent an important category of Remittances represent an important category of capital flows to developing countriescapital flows to developing countries

Recent research emphasizes the potential of Recent research emphasizes the potential of remittances to reduce poverty (Adams and Page, remittances to reduce poverty (Adams and Page, 2003)2003)

When men and women migrate, their ability to When men and women migrate, their ability to monitor household spending is likely to change.monitor household spending is likely to change.

International Migration of Women

Migration is likely to have important Migration is likely to have important consequences for intra-household consequences for intra-household bargaining, for the allocation of household bargaining, for the allocation of household expenditures, and for poverty alleviationexpenditures, and for poverty alleviation

But we know little about these changes…But we know little about these changes…to a great extent, the nexus between to a great extent, the nexus between remittances, bargaining and gender remittances, bargaining and gender remains unexploredremains unexplored

LITERATURE REVIEWLITERATURE REVIEW

Remittances literature:Remittances literature:

As opposed to male remitters:As opposed to male remitters: Female remitters function as an insurer of last Female remitters function as an insurer of last

resort (De la Brière resort (De la Brière et al.et al. 2002, Lee, Parish 2002, Lee, Parish et al.et al. 1994)1994)

Female remitters are more motivated by altruism Female remitters are more motivated by altruism (Vanway 2004)(Vanway 2004)

Female migrants seem to remit to pay for Female migrants seem to remit to pay for education, health, and other family-oriented education, health, and other family-oriented expenditures (De La Cruz, 1995, IOM 2005)expenditures (De La Cruz, 1995, IOM 2005)

→ → Emerging evidence suggests that female Emerging evidence suggests that female and male and male remitters have different preferences remitters have different preferences for how their for how their remittances are being spentremittances are being spent

LITERATURE REVIEWLITERATURE REVIEW

Intrahousehold expenditure literature:Intrahousehold expenditure literature:

Rejects the Rejects the unitary household modelunitary household model and suggests and suggests that:that:

There are differences in preferences between There are differences in preferences between household membershousehold members

Distribution of resources depends on individualsDistribution of resources depends on individuals’ ’ bargaining power within the householdbargaining power within the household

Men frequently have stronger bargaining power Men frequently have stronger bargaining power than womenthan women

Increases in resources controlled by women raise Increases in resources controlled by women raise expenditure allocations toward education, health expenditure allocations toward education, health and nutritionand nutrition

Do sex of household head and Do sex of household head and receipt of remittances affect receipt of remittances affect household expenditure patterns? household expenditure patterns?

Does the sex of the remitter Does the sex of the remitter matter for houeshold matter for houeshold expenditure patterns?expenditure patterns?

Two research questions:Two research questions:

DATADATA Ghana Living Standards Survey 1998/99Ghana Living Standards Survey 1998/99 5,998 households, nationally representative5,998 households, nationally representative Basic info on remitters’ characteristics: sex, Basic info on remitters’ characteristics: sex,

relationship to household, place of residencerelationship to household, place of residence Detailed expenditure dataDetailed expenditure data Good data on remittancesGood data on remittances

But no data on characteristics of non-remitting But no data on characteristics of non-remitting

migrantsmigrants

→ → not able to look at determinants of not able to look at determinants of remittances remittances or analyze whether men and or analyze whether men and women have women have differential propensities to remitdifferential propensities to remit

(Do sex of household head and receipt of remittances affect (Do sex of household head and receipt of remittances affect household expenditure patterns?)household expenditure patterns?)

Specification:Specification: Standard fractional logit model without instruments Standard fractional logit model without instruments Unit of analysis: household receiving remittancesUnit of analysis: household receiving remittances Final sample of households: 5,998 (F=2,017; M =3,981)Final sample of households: 5,998 (F=2,017; M =3,981)

Functional form: Functional form: Adjusted Working-Leser curve:Adjusted Working-Leser curve:

wwii = share of the budget devoted to good i = share of the budget devoted to good i xx= total household expenditure= total household expenditurenn= household size = household size zz= vector of household characteristics= vector of household characteristicsu= error termu= error term

ihhihih

hiiih uzn

n

xw loglog

METHODOLOGY for question #1:METHODOLOGY for question #1:

Estimating equation:Estimating equation:

ihhihih

hihhihihiiih uzn

n

xFHHRRFHHRRw loglog110

RR= household receiving remittancesFHH= female headed households

δ1= quasi difference-in-difference estimator, capturing the impact of being both female-headed household and receiving remittances

FRR-FNR = difference in mean values between female-headed households receiving remittances (FRR) and female-headed households not receiving remittances (FNR)

MRR-MNR = difference in mean values between male-headed households receiving remittances (MRR) and male-headed households not receiving remittances (MNR)

)()(1̂ MNRMRRFNRFRR

METHODOLOGY for question #2:METHODOLOGY for question #2:(Does the sex of the remitter matter for household (Does the sex of the remitter matter for household expenditure patterns?)expenditure patterns?)

Unit of analysis: remitterUnit of analysis: remitter Final sample of remitters: 4,011 (F=1,617 and Final sample of remitters: 4,011 (F=1,617 and

M=2,394)M=2,394) Dependent variables same before: ependiture sharesDependent variables same before: ependiture shares Due to the Due to the principal-agent problem, principal-agent problem, we introduce two we introduce two

new variables that measure the remitter’s degree of new variables that measure the remitter’s degree of enforcement: enforcement:

1) 1) Relationship to household head Relationship to household head (closer → better enforcement)(closer → better enforcement)

2) 2) Location of remitter Location of remitter (closer to receiving household → better enforcement)(closer to receiving household → better enforcement)

RESULTSRESULTS1. Do female headship and receiving remittances matter?1. Do female headship and receiving remittances matter?

FHHs on average have larger expenditure shares allocated to FHHs on average have larger expenditure shares allocated to food (10%) and education (40%), but smaller budget shares food (10%) and education (40%), but smaller budget shares allocated to consumer goods (-15%), housing (-8%) and other allocated to consumer goods (-15%), housing (-8%) and other goods (-8%)goods (-8%)

Quasi difference-in-difference coefficient is significant for four Quasi difference-in-difference coefficient is significant for four of six expenditure categories for FHHs receiving remittances of six expenditure categories for FHHs receiving remittances from abroad: food (-14%), consumer goods (+12%), housing from abroad: food (-14%), consumer goods (+12%), housing (+11%), other (+12%)(+11%), other (+12%)

Striking result: interaction of receiving Striking result: interaction of receiving internationalinternational remittances AND being female-headed just about offsets the remittances AND being female-headed just about offsets the impact of female headship—i.e., preferences of remitters seem impact of female headship—i.e., preferences of remitters seem quite important.quite important.

Quasi difference-in-difference estimator for internal Quasi difference-in-difference estimator for internal remittances is significant only for consumer goods (-6%)remittances is significant only for consumer goods (-6%)

RESULTSRESULTSPredicted budget shares by sex of the household head, origin of Predicted budget shares by sex of the household head, origin of remittances and type of expenditureremittances and type of expenditure

(percentage of total budget)(percentage of total budget)

Type of householdsType of households FoodFood Consumer and Consumer and durable goodsdurable goods HousingHousing HealthHealth EducationEducation OtherOther

Male-headed households not Male-headed households not receiving remittancesreceiving remittances 58.058.0 17.717.7 2.62.6 3.13.1 6.36.3 11.011.0

Male-headed households receiving Male-headed households receiving remittances from Ghanaremittances from Ghana 57.857.8 17.917.9 2.62.6 3.33.3 6.46.4 10.710.7

Male-headed households receiving Male-headed households receiving remittances from abroadremittances from abroad 56.456.4 18.718.7 2.72.7 3.33.3 7.17.1 10.810.8

Female-headed households not Female-headed households not receiving remittancesreceiving remittances 60.360.3 15.415.4 2.42.4 2.92.9 8.68.6 10.210.2

Female-headed households receiving Female-headed households receiving remittances from Ghanaremittances from Ghana 59.959.9 14.814.8 2.42.4 3.33.3 9.49.4 9.99.9

Female-headed households receiving Female-headed households receiving remittances from abroadremittances from abroad 54.954.9 18.018.0 2.82.8 3.43.4 9.39.3 11.111.1

Source: Authors' calculations.Source: Authors' calculations.Note: Percentages do not sum to 100 because of rounding.Note: Percentages do not sum to 100 because of rounding.

RESULTSRESULTS

2. Does the sex of the remitter matter?2. Does the sex of the remitter matter? Unless we account for ability to enforce Unless we account for ability to enforce

contracts, sex of remitter does contracts, sex of remitter does notnot matter. matter. Households receiving remittances from female Households receiving remittances from female

remitters in Ghana have higher percentages of remitters in Ghana have higher percentages of expenditures on food and lower percentages on expenditures on food and lower percentages on consumer and durable goods consumer and durable goods

For female remitters as a whole, expenditure For female remitters as a whole, expenditure share on food is 17% lower; this difference share on food is 17% lower; this difference accounted for by female remitters outside Ghanaaccounted for by female remitters outside Ghana

Relationship to household head seems to matter Relationship to household head seems to matter for some types of expenditure (food and health) for some types of expenditure (food and health)

RESULTSRESULTS

Remitter's characteristicsMale Remitter

Living AbroadRelationship to household head

Sibling 55.4 18.8 3.1 3.1 7.6 10.7Child 55.0 19.4 2.9 3.6 6.1 11.1Spouse 53.5 19.8 2.9 3.2 9.1 10.2

Living in GhanaRelationship to household head

Sibling 53.9 18.9 3.0 3.8 6.7 11.7Child 53.3 19.1 3.0 3.9 7.7 11.8Spouse 52.8 20.9 2.8 2.4 4.8 14.2

Female RemitterLiving Abroad

Relationship to household head

Sibling 57.9 16.9 2.8 3.3 7.3 10.4Child 57.5 17.5 2.7 3.9 5.8 10.7Spouse 56.0 17.9 2.7 3.4 8.7 9.9

Living in GhanaRelationship to household head

Sibling 59.3 16.6 2.7 3.3 6.1 9.9Child 58.7 16.8 2.6 3.4 7.0 10.0Spouse 58.2 18.5 2.5 2.1 4.3 12.1

NOTE: Percentages might not sum to 100 due to rounding.Source: Authors' calculations base on regression results presented in Table 6. Other variables set to original values of each observation.

Table 7. Predicted budget shares by remitter's sex, location and relationship to the household head, and type of expenditure: (percentage of total budget)

Food Consumer and durable goods

Housing Health Education Other

CONCLUSIONSCONCLUSIONS FHHs as a whole spend FHHs as a whole spend moremore on food, education, and on food, education, and

lessless on consumer goods and housing on consumer goods and housing

FHHs receiving remittances from abroad, however, tend FHHs receiving remittances from abroad, however, tend to behave a lot like MHHs: spend more on consumer to behave a lot like MHHs: spend more on consumer and durable goods and housing and less on food. Why? and durable goods and housing and less on food. Why? Are male remitters more successful in imposing their Are male remitters more successful in imposing their preferences? preferences?

In the analysis using remitters as the unit of analysis, In the analysis using remitters as the unit of analysis, the sex of the remitter does not have an impact if one the sex of the remitter does not have an impact if one does not control for the capacity of the remitter to does not control for the capacity of the remitter to follow-up on the intended use of the remittancesfollow-up on the intended use of the remittances

Once that is controlled for, households’ with female Once that is controlled for, households’ with female remitters seem to have a higher budget share devoted remitters seem to have a higher budget share devoted to food, but lower expenditures on consumer durables to food, but lower expenditures on consumer durables and housing. and housing.

POLICY ISSUESPOLICY ISSUES

Is there a case for:Is there a case for:

Mechanisms to increase the bargaining power Mechanisms to increase the bargaining power of women in migrant households, since of women in migrant households, since increased bargaining power may be associated increased bargaining power may be associated with greater anti-poverty expenditures with greater anti-poverty expenditures (principally food and education)?(principally food and education)?

Remittance-facilitating products which target Remittance-facilitating products which target women (both female remitters and female women (both female remitters and female remittance receivers)?remittance receivers)?

THANK YOU !