Savings and Sharing Pressure in the Extended Family Evidence from - - PowerPoint PPT Presentation

Savings and Sharing Pressure in the Extended Family Evidence from Senegal

Marie Boltz-Laemmel

Paris School of Economics

UNU-Wider Conference - September 21th, 2013

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 1 / 31

MOTIVATIONS

Starting with a quote “If you have the money, you have to give it, but if you don’t, you don’t need to. In Senegal, you don’t have the possibility to save, because the family is here, there is the pressure, there is the electricity bill of your brother that you need to pay, there are your parents to help... But the moment you start to do something, then they will let you in peace... that’s why I started to build my house on credit.” Alioune, teacher in Guinawrails, poor suburb of Dakar, December 2012.

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 2 / 31

MOTIVATIONS

Some elements of context In developing countries, esp. Sub-Saharan Africa (SSA): − limited access to financial markets, − low welfare provision, − structural vulnerability of households to shocks ... inducing two well-documented observations: − the preponderance of kinship and social networks as informal insurance providers

(risk sharing), esp. against idiosyncratic shocks,

− low saving rates achieved savings, while savings are a mean to protect against

aggregated risks.

However, few evidence on the effect of social networks on resource accumulation decisions.

→ Objective here: Investigating the role of the extended family in the decisions to

save and transfer.

Savings rates Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 3 / 31

MOTIVATIONS

The extended family in Senegal “Kin system” (Hoff and Sen, 2006), extended family or kinship network = Set of relationships acquired by blood, marriage or adoption. Family (blood) ties are not a matter of choice This “moral economy” (Scott, 1976; Platteau, 1991) entails informal redistribution In Senegal: Lineage-based society Polynuclear households strongly interlinked within the kinship network: importance of the household of origin 70 % of in- and out-transfers (in value) take place within the kin system.

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 4 / 31

LITERATURE

A strong need for commitment to save Recent interests in the economic literature on savings, with the main question: How to increase savings in developing countries? Need for commitment: commitment devices increase drastically demands for savings (Brune et al., 2011; Atkinson et al., 2010). What kind of underlying cost for savings does this need for commitment reveal? − limited attention (Karlan et al., 2011) − preference for present, procrastination behaviors (Basu, 2008; Duflo et al., 2011) − financial pressure of the kin system (Platteau, 2004)

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 5 / 31

LITERATURE

The ambivalent role of social networks Traditionally, the literature has focused on positive aspects of social networks: e.g. risk sharing, social learning, etc. However, recent work shows more balanced evidence: − Di Falco and Bulte (2012): households more exposed to social pressure spend more

• n “non-sharable” goods rather than “sharable” goods.

− Jakiela and Ozier, (2012): experiment showing that a distortion towards less visible

but less lucrative investments due to sharing norms;

− Baland et al. (2011): “Pretending to Be Poor: Borrowing to Escape Forced Solidarity

in Cameroon” (qualitative evidence + theoretical signalling model).

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 6 / 31

RESEARCH QUESTION

What are the mechanisms observed on the field concerning the role of the “family tax” on savings decisions ? How does the size of and the position in the extended family influence both saving and transfer decisions?

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 7 / 31

MECHANISMS AT PLAY

Field evidence on redistributive pressure Qualitative survey conducted in Senegal: Members of several extended families traced in different locations → better understand mechanisms at play betw. transfers and savings: Redistributing to one’s network reported as a duty, sometimes even as a pride; However, sharing pressure is also often seen as a burden for people, leading to: − Lower revenue left for savings after the family tax ⇒ decrease in savings − Or to family tax evading strategies through reduction in the observability and/or

taxability of revenues or assets

⇒ Increase in “non-sharable” or “non-observable” savings.

More strategies Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 8 / 31

SAVING TYPES

Costs and benefits Non-sharable savings: bank, shopkeeper, ROSCAs, indivisible goods

+ Asset not easily shared, b/c not observable, not directly available or not divisible. + Money not easily consummed (pref. for present) – Not directly available in case of need. – Entry costs (open a bank account), transportation costs, – Requires trust (ROSCA, shopkeeper), – Returns: low or no interest rate, highly risky for ROSCAs, livestock.

Sharable savings: money kept home

+ Directly available in case of need. – Easily captured by peers within the hh and out of the hh (through transfers) – Easily consummed (pref. for present) – No returns

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 9 / 31

MECANISMS AT PLAY

Redistributive Pressure

Figure: The heterogenous effects of a larger extended family for different positions in

the network

Position in the social network Savings Transfers Total % NS Out In Net out Salient economic difficulties – – – + – Salient economic success – – ++ – ++ Salient economic success & CS + + + – +/0

NS = Non-Sharable savings, CS = Coping Strategies aimed at lessening the pressure to transfer

Empirical strategy Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 10 / 31

DATA

“Poverty and Family Structure” in Senegal Nationally representative survey: “Poverty and Family Structure” (PSF), conducted in Senegal in 2006 and 2007 by De Vreyer, Lambert, Safir, Sylla. Sampling: 1750 hhs for 14,450 individuals, spread over 150 clusters drawn randomly from census districts. 57,1% rural, 48% male, 95 % Muslim. Household structure: − Household size: 8 individuals ; 2.4 cells /hh. − Horizontally and vertically extended household: intra and inter-generational.

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 11 / 31

DATA

What do we observe? Savings Indiv. self-reported amount of their actual saving (stock): − Less sharable savings: in formal banks and informal associations, in ROSCAs

(“tontines”)

− More sharable savings: at home or entrusted to some people

Transfers Successful requests reported for the past year (flow):

− Out-transfers sent by individuals. − In-transfers received by individuals. − Net out-transfers: out - in transfers

Strength of the family pressure

− Size of the family network: Number of siblings decomposed into : same parents,

same mother/father only and gender (brothers/sisters)

− Position in family network: economic status/ education level relative to father

(e.g.working in formal sector while father in agri), situation of parents (location, death).

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 12 / 31

DATA

Descriptive statistics on savings Socio-demographic characteristics of savers

• savers have more brothers and sisters
• they are more likely to be civil servants (7% vs. 2%)
• they earn a much higher income (792000 FCFA vs. 338 000FCFA)

Savings habits

• 56% of HH save, Women are more likely to save than men (32% vs. 17.6%)
• Among active adults, savings at home represent 6% of total savings, tontines and

saving in formal banks 21%.

Table 1 Table 2 Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 13 / 31

DATA

Descriptive statistics on transfers Transfer in general

• The vast majority of hhs receive or give transfers (83%)
• In total amount, men are the principal transfer senders and women the transfer

receivers

Out-transfers and kin network

• Among transfer senders, 82% have at least send one transfer to the family.
• 72% of total value transferred on average are sent to kin members.

Table 1 Table 2 Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 14 / 31

EMPIRICAL STRATEGY

Yih = α0 +α1Sibshipih +α2Positionih +α3Sibshipih ∗Positionih +βCih +vn +εih

Yih, outcome for individual i in household h − Savings: total, ratio non-sharable over total savings − Transfers: total out-, regular out-, total net-out transfers

Family pressure exerted on i in household h

− Sibshipih = Family size: Number of siblings, decomposed between same parents,

same mother or father only; brothers & sisters.

− Positionih, = Position in the family network: social mobility relative to father

· Economic status relative to father: working in formal sector while father in agri, · Education level relative to the father: secondary edu while father no edu.

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 15 / 31

EMPIRICAL STRATEGY

• 2

Yih = α0 +α1Sibshipih +α2Positionih +α3Sibshipih ∗Positionih +βCih +vn +εih Cih, set of controls for indiv. and household socio-economic characteristics

− Age, French and Koranic education, marital status, head or cell head, − Revenues in log, employment sector (public, agricult., informal), occupational status, − Household var: household size, relative cell size, share of inactives, − Parental background: (former) employment sector, education, death,

vn, neighborhood fixed effects εih, standard errors clustered at household level

Mechanisms Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 16 / 31

RESULTS

20-60 yr old men:

Yih = α0 +α1Sibshipih +βCih +vn +εih

TS LN(TS) NS/TS LN(Out) LN(Reg out) Out-In (1) (2) (3) (4) (5) (6)

• Nb. same-parents siblings

1.765* 0.050 0.005* 0.070* 0.064* 1.031 (1.035) (0.038) (0.003) (0.038) (0.037) (1.018)

• Nb. same-mother-only siblings

4.110** 0.151** 0.010** 0.008

• 0.011

0.868 (2.014) (0.059) (0.004) (0.067) (0.066) (1.507)

• Nb. same-father-only siblings

0.734 0.043* 0.002 0.013 0.038 0.218 (0.659) (0.023) (0.002) (0.023) (0.024) (0.599) Constant

• 23.908
• 1.612***
• 0.070*

0.256 0.800 55.100*** (16.179) (0.589) (0.042) (0.632) (0.634) (15.981) N 2479 2518 2477 2518 2518 2401 Adjusted R2 0.13 0.21 0.16 0.36 0.28 0.087

Sample: 20-to-60 year-old males. Standard errors in parentheses are clustered at hh level, neighbourhood FE incl. ∗p < 0.10,∗ ∗ p < 0.05,∗ ∗ ∗p <

0.01.

Control var incl: ln income, sector of activity, education, age, marital status, household/cell head, household size, cell identifier, relative cell size, share of inactives, parental education and sector of activity, father deceased, parents living in other places. Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 17 / 31

RESULTS

20-60 yr old women:

Yih = α0 +α1Sibshipih +βCih +vn +εih

TS LN(TS) NS/TS LN(Out) LN(Reg Out) Out-In (1) (2) (3) (4) (5) (6)

• Nb. same-parents siblings
• 0.350

0.036

• 0.001
• 0.007

0.026

• 2.538**

(0.706) (0.039) (0.002) (0.031) (0.029) (1.241)

• Nb. same-mother-only siblings
• 1.098
• 0.035
• 0.001
• 0.109*

0.004

• 0.010

(1.285) (0.065) (0.004) (0.059) (0.053) (1.636)

• Nb. same-father-only siblings

0.015 0.019 0.001 0.064*** 0.029 0.804 (0.508) (0.023) (0.001) (0.021) (0.022) (0.805) Constant

• 18.656

0.582 0.030 3.195*** 1.901*** 29.017 (13.232) (0.652) (0.032) (0.618) (0.511) (20.776) N 2754 2774 2754 2774 2774 2675 Adjusted R2 0.13 0.23 0.14 0.31 0.24 0.077

Sample: 20-to-60 year-old females. Standard errors in parentheses are clustered at household level, neighbourhood fixed effects incl. * p¡0.10, ** p¡0.05, *** p¡0.01. Control var incl: ln income, sector of activity, education, age, marital status, household/cell head, household size, cell identifier, relative cell size, share of inactives, parental education and sector of activity, father deceased, parents living in other places.

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 18 / 31

RESULTS Men:

Yih = α0 +α1Sibih +α2Positih +α3Sibih ∗Positih +βCih +vn +εih

High econ status = formal job, with social mobility = father farmer

TS LN(TS) NS/TS LN(Out) LN(Out reg) Out-In (1) (2) (3) (4) (5) (6)

• Nb. same-parents siblings SP

1.582 0.046 0.004 0.045 0.012 0.112 (1.057) (0.040) (0.003) (0.042) (0.038) (1.054)

• Nb. same-mother-only siblings SM

2.086 0.095 0.007 0.012

• 0.042

1.214 (1.982) (0.062) (0.005) (0.070) (0.069) (1.637)

• Nb. same-father-only siblings SF

1.124 0.044* 0.002 0.011 0.040 0.096 (0.697) (0.025) (0.002) (0.024) (0.025) (0.608) High econ. status w/o social mobility X SP 5.574 0.137 0.013 0.176* 0.317*** 7.641** (3.655) (0.120) (0.009) (0.106) (0.113) (3.694) High econ. status w/o social mobility X SM 14.401** 0.247 0.018

• 0.154
• 0.043
• 2.965

(6.834) (0.170) (0.013) (0.159) (0.169) (3.779) High econ. status w/o social mobility X SF

• 2.153

0.012 0.003 0.048

• 0.011

0.294 (2.305) (0.073) (0.005) (0.073) (0.074) (2.299) High econ status w/ social mobility X SP

• 10.351*
• 0.330*
• 0.024*
• 0.014

0.094

• 4.975

(6.163) (0.185) (0.013) (0.218) (0.225) (3.972) High econ status w/ social mobility X SM

• 9.647

0.283

• 0.001

0.609** 1.052*** 15.674* (10.861) (0.371) (0.028) (0.280) (0.306) (8.929) High econ status w/ social mobility X SF

• 0.894
• 0.025
• 0.008
• 0.066

0.016 1.443 (2.642) (0.097) (0.006) (0.092) (0.108) (2.027) Constant

• 28.184*
• 1.728***
• 0.081**

0.299 0.940 53.791*** (15.230) (0.569) (0.041) (0.611) (0.605) (15.377) N 2479 2518 2477 2518 2518 2401 Adjusted R2 0.13 0.21 0.16 0.36 0.29 0.090

• Sample: 20-to-60 year-old males
• S = Total Savings (FCFA), LN(S) = LN of total savings, NS/S = share of Non-Sharable in total Savings, LN(Out) = LN of out-

transfers, LN(Out-reg) = LN of regular out-tr., Out-In = net out-transfers (FCFA). Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 19 / 31

RESULTS Women:

Yih = α0 +α1Sibih +α2Positih +α3Sibih ∗Positih +βCih +vn +εih

High econ status = husband in formal job, social mobility = father farmer

TS LN(TS) NS/TS LN(Out) LN(Out reg) Out-In (1) (2) (3) (4) (5) (6)

• Nb. same-parents siblings SP
• 0.262

0.034 0.001

• 0.012

0.039

• 1.543

(0.760) (0.042) (0.002) (0.034) (0.031) (1.153)

• Nb. same-mother-only siblings SM
• 0.534

0.011 0.002

• 0.101

0.029

• 0.136

(1.279) (0.068) (0.004) (0.064) (0.057) (1.701)

• Nb. same-father-only siblings SF

0.619 0.020 0.002 0.066*** 0.026 1.502* (0.564) (0.025) (0.001) (0.023) (0.025) (0.804) High econ status w/o social mobility X SP

• 0.908

0.090

• 0.011**

0.032

• 0.112
• 6.906

(3.663) (0.118) (0.005) (0.097) (0.095) (5.044) High econ status w/o social mobility X SM

• 3.755
• 0.232
• 0.018**
• 0.026
• 0.215
• 2.397

(6.345) (0.197) (0.009) (0.181) (0.166) (6.107) High econ status w/o social mobility X SF

• 3.379**
• 0.007
• 0.002
• 0.007

0.008

• 3.058

(1.407) (0.059) (0.003) (0.052) (0.055) (2.602) High econ status w/ social mobility X SP

• 1.087
• 0.275
• 0.018**
• 0.065
• 0.053
• 2.781

(2.406) (0.196) (0.009) (0.157) (0.143) (5.966) High econ status w/ social mobility X SM

• 11.779**
• 0.400
• 0.024
• 0.241
• 0.049

4.537 (5.213) (0.512) (0.016) (0.408) (0.415) (5.782) High econ status w/ social mobility X SF

• 2.247
• 0.060
• 0.008
• 0.059

0.063

• 7.314

(2.559) (0.156) (0.005) (0.170) (0.159) (4.901) Constant

• 24.894*

0.404 0.027 3.637*** 2.131*** 20.105 (13.803) (0.656) (0.032) (0.629) (0.549) (22.005) N 2754 2774 2754 2774 2774 2675 Adjusted R2 0.13 0.23 0.14 0.31 0.24 0.080

• Sample: 20-to-60 year-old females
• S = Total Savings (FCFA), LN(S) = LN of total savings, NS/S = share of Non-Sharable in total Savings, LN(Out) = LN of
• ut-transfers, LN(Out-reg) = LN of regular out-tr., Out-In = net out-transfers (FCFA).

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 20 / 31

RESULTS

In short

Figure: The heterogenous effects of a larger extended family for different positions in

the network

Position in the social network Savings Transfers Total % of non-sharable Net out Men All 20-60 yr old males + + / High econ. status w/o social mobility: + / + Salient social mobility: – – + Women All 20-60 yr old males / /

• High econ. status of husband w/o social mobility:

– – (-) Salient social mobility: – – (+)

Mechanisms Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 21 / 31

DISCUSSION

Alternative measure of position: attending at least secondary education, relative to father w/ or w/o any education ⇒ same results. So far, suggestive evidence (though not causal) on the fact that kinship pressure as proxy by the number of siblings reduces savings for individuals who had a social mobility relative to their father. What alternative competing channels –other than redistribution through transfers– could explain the effect of siblings on savings?

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 22 / 31

CONCLUSION

This is a study on the role of informal redistribution on saving decisions. A lower capacity to save and distorted preference for some types of savings affect people’s ability to face risks and to invest, and thus to escape poverty trap. Analyzing the role of informal redistribution on ressource allocation and accumulation is crucial for the design of adequate public policies (taxation, financial inclusion). More work needs to be done, through specific data collection to assess the costs and benefits of such informal redistribution on the family networks as a whole.

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 23 / 31

Thank you

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 24 / 31

Savings in the World

return

Table: Saving rates from 1970-2010 in different regions of the world Geographic zone 1970-1990 1990-2010 Arab World 29,86 26,49 East Asia & Pacific 32,11 30,34 Europe & Central Asia 22,58 21,78 Latin America & Caribbean 22,27 20,42 Middle East & North Africa 25,45 26,71 OECD members 23,10 20,46 South Asia 16,40 23,33 Sub-Saharan Africa 21,26 15,91 Senegal 6,92 7,25 World 23,61 21,84

Source: World Bank Data

⇒ Saving rates in SSA decreased over time and become the lowest in the world.

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 25 / 31

Field-based tips for family tax evasion

Back

2 major strategies:

• 1. Reduce observability of revenues:“Never tell anyone how much you earn!”

Out of sight, out of... threat ! Migrate and start a business far from your family. Maintain the illusion of having a small business Save non-observable or exceptional income e.g.bonus/extra income. Simply don’t tell you save : use anonymous ROSCAs

• 2. Reduce taxability of revenues:

Never save at home! : the money left at the bank at 300m or at the shopkeeper next door is not subject to the practical norm. Use ROSCAs to commit yourselves publicly so people respect the fact you cannot give to honor your commitment. Finance investments by taking on a loan Favor non-sharable goods: e.g. a golden necklace or livestock easily sold in case of need!

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 26 / 31

Back

Table: Socio-economic characteristics of 20-to-60-year-old savers

Variable Obs Savers Non-savers Diff P-Val # siblings same parents 5975 4.3945 4.2910 0.103 0.1799 # siblings same mother only 5964 0.7144 0.5983 0.116 0.0086*** # siblings same father only 5961 3.6714 3.2383 0.433 0.0005*** # siblings in total 5956 8.7761 8.1295 0.647 0.0000*** # brothers in total 5964 4.4287 4.1924 0.236 0.0080*** # sisters in total 5959 4.3486 3.9371 0.411 0.0000*** Female 6173 0.6851 0.4961 0.189 0.0000*** Age 6173 37.6221 33.5217 4.100 0.0000*** Rural 6173 0.3180 0.4407

• 0.123

0.0000*** Dakar 6173 0.4634 0.3690 0.094 0.0000*** Primary edu 6173 0.2161 0.2003 0.016 0.1801 Secondary edu or more 6173 0.2192 0.1735 0.046 0.0001*** Koranic school 6173 0.1848 0.1790 0.006 0.6025 Public servant 6173 0.0701 0.0237 0.046 0.0000*** Yr Income in 1000 FCFA 5690 791.6875 338.1181 453.569 0.0000***

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 27 / 31

Descriptive statistics

Back

Table: Savings in the sample of 20 to 60 yr olds Households Active adults Women Men Diff. Total savings 56% 25.4% 32% 17.6% *** Stock in FCFA 27661 27139 28288 Savings in formal banks 17.2% 5.5% 3.6% 7.7% *** Stock in FCFA 25567 10797 43294 *** Tontines 39.5% 16% 25.6 % 4.6% *** Value of the pot in FCFA 14690 21180 6976 *** Savings at home 14.4% 5.9% 4.8% 7.2% *** Stock in FCFA 3463 1927 5304 *** “Non-sharable” savings 49.7% 21.6% 29.4% 12.2% *** Stock in FCFA 15213 8699 23077 *** N 1781 6173 3359 (54.4%) 2814 (45.6%)

* p<0.10, ** p<0.05, *** p<0.01

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 28 / 31

Back

Table: Transfers in the sample of 20 to 60 yr olds Households Active adults Women Men Diff. Any transfer 83.1% 43.5% 47.4% 38.7% *** In- &out-transfers 24.1 % 21.2 % 18.3% 23.6% *** Out-transfers 67.2% 35.3% 37.1% 33.2% *** Year amount 36071 22995 51816 *** In-transfers 59% 22.2% 28.1% 15.1% *** Year amount 49487 69051 26253 *** N 1781 6173 3359 (54.4%) 2814 (45.6%)

* p<0.10, ** p<0.05, *** p<0.01

→ Men are in value the principal transfer senders and women the transfer receivers.

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 29 / 31

Descriptive statistics

Back

Table: Out-transfers to whom?

N Mean Sd Min Max Diff Kinship At least 1 2533 .81 .39 1 Nb 2533 1.6 1.8 24 Share in value 2512 .72 .4 1 Women share 1424 .66 .42 1 *** Men share 1088 .78 .38 1 Siblings At least 1 2533 .17 .38 1 Nb 2533 23 .63 12 Share in value 2525 .1 .27 1 Women share 1433 .088 .24 1 *** Men share 1092 .12 .29 1

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 30 / 31

return

Table: Out-transfers to whom?

N Mean Sd Min Max Diff Parents At least 1 2533 .23 .42 1 Nb 2533 .3 .66 13 Share in value 2522 .18 .37 1 Women share 1432 .17 .35 1 *** Men share 1090 .21 .38 1 Children At least 1 2533 .063 .24 1 Nb 2533 .082 .36 5 Share in value 2526 .04 .18 1 Women share 1434 .03 .16 1 *** Men share 1092 .054 .21 1

Marie Boltz-Laemmel Savings and Family Solidarity in Senegal 31 / 31