The Impact of Kinship Networks on the Adoption of Risk-Mitigating Strategies in Ethiopia Salvatore Di Falco and Erwin Bulte

Jos´e Victor Cremonesi Giarola

Carlos Monge-Badilla

Universit´ e Paris 1 Panth´ eon-Sorbonne Development Economics Master

November 8, 2016

Outline Introduction The Moral Economy of Kinship Weather Shocks and Mitigating Responses in Ethiopia Data and Econometric Model Regression Result Discussion and Conclusion

1/34

Introduction Motivation

I

Ethiopia is one of the least developed countries in the world and Agriculture is the mainly activity

I

Farms in the Horn of Africa (Peninsula in Northeast Africa) are exposed to regular weather shocks

I

The role of Kinship obligations in adoption of risk mitigation strategy

I

Kinship groups differs from other groups because it’s not possible to choose to participate or not

2/34

Introduction

I

Reducing exposure to risk through different channels 1. Self-protection 2. Risk sharing

I

Kinship represents a primary principle of social organization in Sub-Saharan Africa

I

Blood relations promote altruism, what tends to ameliorate commitment problems

I

Kin networks ⇒ Sharing obligations Consequences of Sharing Obligation:

I

1. reduces idiosyncratic risk 2. possibility of free riding behavior

3/34

Introduction I

This paper focus on the adverse effects of kinship linkages

I

Adoption of risk mitigation technologies ⇒ Buffer the impact of weather shocks

I

Kinship linkages sphere ⇒ Free riding problem

I

Free riding problem ⇒ Social dilemma

I

Social dilemma in the network ⇒ Low levels of mitigation efforts emerge as an equilibrium outcome

I

Cutting back on self-protection ⇒ Not a social optimal equilibrium strategy

I

Main intend of the paper: Analyze the correlation between kinship obligations and adoption of technologies to reduce exposure to weather shocks (drought and floods)

4/34

Outline Introduction The Moral Economy of Kinship Weather Shocks and Mitigating Responses in Ethiopia Data and Econometric Model Regression Result Discussion and Conclusion

4/34

The Moral Economy of Kinship I

I

Within kinship, members may claim assistance from others if necessary Alger and Weibull model: I

I

I I

I

I

Two family members who have to decide about the optimal level of self-protection Individuals can undertake effort to reduce the risk of earning low income in case of a weather shock The effort comes at a private cost In case of a shock that affects the income of one individual, a sharing rule dictates that the other individual should provide assistance in form of transfer Individuals in a population are pairwise and randomly matched to other individuals The share of matches involving kin members increases as the number of kin members increases.

5/34

The Moral Economy of Kinship Alger and Weibull Model

I

I

Family members are altruistic, however, what happen if the sharing norm prescribes a transfer that exceeds the one that would be voluntarily provided? The level of effort varies with the level of altruism and they distinguish between two effects: I

Empathy Effect I I

I

Driven by altruism Reduce the probability of having to draw on ones kins resources

Free-rider Effect I I I

Forced sharing Ability to live off the efforts of kin Disincentive to put in effort due to the risk of sharing with members with low pay-off

6/34

The Moral Economy of Kinship I

Sharing Norm > Voluntary Transfer −→ ↓ Incentives to self-protection

I

Focus of this paper is to test the prediction that kinship ties adversely affect self-protection Two ways of self-protection:

I

I I

tree planting soil conservation

I

The model applied to the context of this paper −→ Compulsory sharing leading to more generous transfers than voluntary sharing −→ ↑ kin members will discourage self-protection against weather risk.

I

Analyze the relation between kinship links and self-protection, and explore whether the incentive effects are sufficiently large to dominate opposite effects due to, say, dilution of the norm or altruism 7/34

Outline Introduction The Moral Economy of Kinship Weather Shocks and Mitigating Responses in Ethiopia Data and Econometric Model Regression Result Discussion and Conclusion

7/34

Weather Shocks and Mitigating Responses in Ethiopia

I

Ethiopia is one of the least developed countries in the world

I

80% of workers are employed in the Agricultural sector, accounting for 45% of GDP and 85% of export revenues

I

Ethiopian agriculture heavily depends on natural rainfall (only 4-5% is irrigated)

I

Rainfall and temperature are important determinants of crop harvests

I

the amount or the temporal distribution of rainfall triggers food shortages and famine

8/34

Weather Shocks and Mitigating Responses in Ethiopia

I

2 well-known strategies of risk mitigation 1. Tree planting I

Maintain the soil moisture, conserve soil organic matter, reduce soil loss due to erosion and flooding, and provides shades for other crops

2. Implementation of soil and water conservation I

Include soil bunding, cultivation of hedges, contour ploughing, irrigation, and water harvesting activities

9/34

Outline Introduction The Moral Economy of Kinship Weather Shocks and Mitigating Responses in Ethiopia Data and Econometric Model Regression Result Discussion and Conclusion

9/34

Data and Econometric Model Data

I

Data from 2004-2005

I

Sample 1000 households (20 villages with 50 random households sample) Sample frame considers:

I

1. Percentage of cultivated and irrigated land 2. Average rainfall as well as rainfall variability 3. Vulnerability of the population I

Standard household characteristics: age, literacy, and size of the household

I

Agricultural variables: quantity of fertilizer, manure applied per unit of land, and a measure of soil quality (self-reported)

10/34

Data and Econometric Model Data

I

The Thin Plate Spline method of spatial interpolation was used to impute household specific rainfall and temperature values using latitude and longitude.

I

Kin variable is simply defined as the (self-reported) number of relatives of the household living in the same village, but not in the same household

I

Consider family members up to cousins, nephews, and nieces as kin

I

There is information about distance between the household and the village

I

There is information about how many farmers in the village are engaged in soil conservation or tree planting.

11/34

Figure: 1.Variable list and descriptive statistics

Figure: Source: Salvatore Di Falco and Erwin Bulte (2013).

12/34

Data and Econometric Model Data

I

2 non-market institutions for risk sharing are widely used by farm households in our sample(66% have access): 1. Iddirs 2. Iqqub

I

Dependent Variable: The survey aimed to analyze farmers adoption strategies in response to long run changes in climatic variables

I

90% of the sample perceived such changes in mean temperature or/and rainfall

I

The data suggest that measures were taken on 26% of the operated plots: 16% involved planting trees and 10% involved measures aimed at reducing soil and water conservation

13/34

Data and Econometric Model Econometric Strategy

I

The econometric model is: Aih = A(xh , xlh , xch , Nh ; β) + εhi

I

(1)

Where: I I I I I I I

Aih represents the i − th strategy for household h xh households characteristics xlh are land variables xch are climate variables Nh network β vector of parameters εhi is a household specific error term

14/34

Data and Econometric Model Econometric Strategy

I

There is no data on technology choices of all kin members

I

The size of one’s kinship is exogenous Concerns

I

I

I

I

Kinship may be correlated with wealth or status, and the number of local kinship members may vary with asset holdings, thus kinship may be endogenous Respondents from families with excellent soil quality may have more kin members in the family If soil quality also affects technology adoption decisions, then the kin and adoption variables are correlated but this correlation does represent a causal effect of kinship on adoption.

15/34

Data and Econometric Model Econometric Strategy

I

Measures I

I

The problem would be solved if there were complete set of control variables. There is a soil quality proxy (quantity of fertilizer and manure applied per unit of land) Deal with endogeneity problem with the following tools:

1. Control for land and manure 2. Implement an Instrumental variable probit approach to deal with the kinship endogeneity problem (Reported death of a kin member in the past 10 years) 3. Fixed effects I I

Village fixed effect To adress the issue of social status or cultural differences, include pseudo fixed effects

16/34

Data and Econometric Model Econometric Strategy

I

Run a random effects model while controlling for unobserved heterogeneity using Mundlaks approach (Mundlak, 1978; Wooldridge, 2002). This is a very general and convenient way of inserting fixed effects in a probit model.

I

It is also a useful specification in the presence of some farm households invariant variables

I

This approach exploits the existence of plot-varying exogenous variables, and models heterogeneity as a linear function of all exogenous variables.

I

Besides the possible correlation with unobservables, the kinship metric can be correlated with other observables

17/34

Data and Econometric Model Econometric Strategy

I

Therefore estimate an auxiliary regression regressing the kinship metric on all other control variables

I

Hence, after regressing kinship against all explanatory variables, we use the residuals from this regression in place of the original kinship variable to partial out the effect of other covariates

I

To analyze how kinship impacts on adoption of soil conservation measures and tree planting, first estimate two separate probit models.

18/34

Outline Introduction The Moral Economy of Kinship Weather Shocks and Mitigating Responses in Ethiopia Data and Econometric Model Regression Result Discussion and Conclusion

18/34

Regression Results I

The econometric results are divided in two different kind of estimation. I I

I

Case of soil conservation. Case of tree planting.

In the same way, using different sets of explanatory variables, each kind of model is estimated: 1. Parsimonious specification. 2. Adding variables in order to control for household characteristics and extension services. 3. Including farm inputs and soil fertility variables. 4. Using weather variables. 5. With variables to capture spillover and risk-sharing network effects. 6. IV Probit approach. 7. IV Probit pseudo FE approach.

19/34

Figure: Kinship and adaptation to climate change-the case of soil conservation

20/34

Regression Results

I

The number of kinship links is significantly correlated with a lower probability to invest in soil conservation.

I

Contrasts with earlier studies: they tend to document a positive effect of networks size on the adoption of new technologies.

I

However, these studies are based in network proxies. They include kin and non-kin.

I

Authors’ approach focus exclusively in relatives. It allows a deeper analysis identifying adverse incentives induced by forced solidarity.

21/34

Regression Results

I

Marginal effects: I I I

kinship variable: −0.005. number of adopters in the village: 0.02. availability of other risk mitigating networks: 0.29.

22/34

Figure: Kinship and adaptation to climate change-the case of tree planting

23/34

Regression Results

I I

Last table shows results qualitatively different. Most important difference: there exists a positive correlation between kinship and adoption. I

Trees serve a double function: 1. reducing exposure to weather shocks 2. enhancing tenure security by signaling outgoing use and investment.

I

So, farmers plant trees also to decrease the risk of losing one’s property to a kinship member (ex. when challenging tenure rights because of heritage).

24/34

Regression Results I

20% of the plots are titled. Author’s split the sample to distinguish between tenure secure households and tenure insecure households.

I

The kinship variable only enters significantly for the subsample of tenure insecure households.

Figure: Tenure security and tree planting

25/34

Regression Results

I

The authors implemented Wald test for exogeneity for both kind of models.

I

There is no evidence of possible endogeneity in soil conservation model. At the contrary, authors can reject the null hypothesis of exogeneity in tree planting model.

I

I

I

The standard probit estimation results in the tree planting analysis can have endogeneity bias.

However, the results for the kinship variable are qualitatively unaffected across all the different specifications.

26/34

Regression Results I

Authors tried to test the effect of access to formal credit and its effects in the kinship network.

I

The survey contains a question related to use of credit in these communities in 2004-2005.

I

In the survey around 25% of the sample reported to have used credit. This new variable could cause endogeneity problems at household level.

I

I

I

Access to formal credit may be correlated with unobservable household characteristics.

The authors used this question to build a community credit variable (exogenous to households characteristics).

27/34

Regression Results

I

They distinguish between ”credit villages” and ”non-credit villages”.

I

For that, they set a threshold at 15%: if more that 15% of the respondents in a village answer to have access to credit, that village is define as ”credit village”.

I

If not, the village is defined as ”non-credit villages”.

I

They change the threshold to 25% and 35%, and it doesn’t affect the results.

28/34

Figure: Kinship and investments in the presence of access to credit

29/34

Outline Introduction The Moral Economy of Kinship Weather Shocks and Mitigating Responses in Ethiopia Data and Econometric Model Regression Result Discussion and Conclusion

29/34

Discussion and Conclusion I

I

Traditional sharing norms in kinship networks may decrease incentives to adopt protective measures against weather shocks. The kinship variable captures the importance of sharing norms, but possibly various effects as well (altruism, social learning). I

It only allows to test if the overall effect is positive or negative.

I

Endogeneity concerns may remain.

I

The data are rather crude. For instance, kinship variable does not include kinship members living outside the village.

I

To assess if investment levels are sub-optimally low (or high) requires a comparison of all relevant marginal benefits and costs (data unavailable).

30/34

Discussion and Conclusion I

The findings are consistent with the hypothesis that incentives for selfish behavior exist, even in the ”moral economy” of kinship networks.

I

Risk sharing is not effective in the context of systemic events like droughts and floods, because it affects many or all members simultaneously.

I

Individuals have incentives to defect. The result is a social equilibrium that is Pareto dominated.

I

When Ethiopian farmers lack of property rights, to reduce exposure to losing lands, they plant trees to signal ownership. Trees double function:

I

1. they protect against temperatures increases associated with climate change 2. they protect household’s resource base against claims by others. 31/34

Discussion and Conclusion

I

Household with a greater kin network are more likely to plant trees.

I

While being part of a kin network reduces incentives for soil conservation, it increases incentives for tree planting.

I

Credit may “crowd out” some forms of network if it can supply insurance benefits at a lower cost.

I

Indeed, the negative effect of the network on adoption of risk-mitigating measures disappears once we focus on credit villages.

I

Financial expansion may dissolve poverty trap outcomes.

32/34

Thank you!

The Impact of Kinship Networks on the Adoption of Risk-Mitigating Strategies in Ethiopia Salvatore Di Falco and Erwin Bulte

Jos´e Victor Cremonesi Giarola

Carlos Monge-Badilla

Universit´ e Paris 1 Panth´ eon-Sorbonne Development Economics Master

November 8, 2016