Insuring against the weather using traditional groups
1. Insuring against the weather:
Using traditional groups to promote index-based weather
insurance in Ethiopia
Guush Berhane, Daniel Clarke,
Stefan Dercon, Ruth Vargas Hill and
Alemayehu Seyoum Taffesse
IFPRI ESSP-II
Improved evidence towards better food and agricultural
policies in Ethiopia; November 02, 2012
Hilton Hotel, Addis Ababa
1
2. Introduction
Weather risk remains a major challenge to farming in the arid
and semi-arid areas of the tropics;
With ever changing climatic conditions, agriculture has
become increasingly uncertain business!
Drought explains largest share of income variability in Ethiopia
Household level = 60 – 75%
National - strong GDP and rainfall variability
2
3. Introduction
Thin insurance possibilities. Informal insurance hampered
by risks correlated across households and villages;
Index-based weather insurance offers new possibilities;
Several experimenations, including in Ethiopia, but
demand remains invariably low; chances of scalling up
stiil very low!
Basis-risk – a key challenge
Efforts to mitigate basis risk are so far very limited;
3
4. Introduction
Question – design simple, flexible, and affordable generic
insurance policy that mitigates basis risk?
Reduce basis risk by increasing side-payments?
Institutionalization of pre-defined sharing rules needed?
Would such insurance design work? Welfare effects?
If so, what are the mechanism through which this would
work?
Can we achieve the dual goal of ‘harnessing groups to mitigate
basis risk’ and ‘make them more resilient to correlated risks’?
Approach - randomized field experiment
4
5. Weather index pilot in Ethiopia
Long run pilot—looking at group institutions takes time
first year in 2011, second year in 2012, continues …!
57 Kebeles selected around 3 weather stations in Oromia
region of Ethiopia – Shashemene, Dodota and Tibe;
Primary interest is to target risk-sharing group, so we designed
the pilot such that we can evaluate effects of our intervetions!
5
6. Pilot Design 57 Kebeles
(110 Villages)
TREATMENT CONTROL
(60 villages) (50 villages)
GROUP INDIVIDUAL
(35 villages) (25 villages)
MANDATED NON-MANDATED
(18 villages) (17 villages)
6
7. Mandated sharing-rules
What did we mandate?
Discuss and set sharing rules (or bylaws) … key features
Regular savings to a common pot;
Contribute 10% of any insurance payout in this group to this
pot;
Disburse this pot to members that experience idiosyncratic
basis risk, as loan at zero-interest;
7
8. Provision of savings
Money was contributed (to the pot) by project as “savings”
with the aim of
Examine disbursements and promote trust!
Help initiate discussions on formulation of (pre-defined)
additional sharing-rules
Disbursed to 800 Birr to both iddir villages (mandated and non-
mandated) and individual villages (16 individuals, 50 Birr each).
8
12. Insurance marketing
Very few early season (May, June and July) polices were sold
in 2011!
Discounts offered for late season (September/Meskerem) in
2011 & for all season in 2012 policies
Free insurance in Dodota and Bako Tibe;
Price discounts in Shashemene: 40%, 60%, and 80%
discounts randomly allocated across villages;
12
13. Insurance sales …2011
296 policies were sold in Shashemene (134 individuals
and 435 iddir members), about 13% of households;
13
14. Payouts …2011
September rains were poor in Shashemene – index
triggered a payout!
Insurance payout was made at the end of October in
Shashemene.
“Savings” payouts were also made at the end of October
in all three sites.
14
15. Survey & data
Baseline survey: February –March 2011:
1760 households in 110 villages
(16 households per village);
Follow up survey I: December 2011;
Follow up survey II: February-March 2012;
Follow up survey III: February-March 2013;
15
16. Baseline characteristics … households
High incidence of drought:
51% experienced drought shock in the last three years;
Very little knowledge of insurance:
10% had heard about traditional indemnity (car, life or
health) insurance;
High initial interest in index-type insurance:
87% were interested in a weather indexed insurance policy
described to them in the survey;
Indications of huge basis risk:
only 32% thought rainfall measured at the nearest weather
station can accurately measure rainfall on their plots;
16
17. Baseline characteristics … Iddirs
Key features of Iddirs:
Very prevalent in those areas (as in many parts of
Ethiopia)
92% households belong to 1-5 iddirs; only 5% did not
belong to an iddir
They are limited to ‘the village ‘…
80% span within the village
17
18. Data analysis
Compare outcomes between the control and the following
treatment groups:
Individual and iddir
Mandated and non-mandated iddirs
Run a simple ANCOVA for outcome variables of interest with
baseline data;
18
19. Results
Effects on insurance take-up:
Interventions increased insurance purchases both in individual
and mandated iddir villages, but no statistical difference in
amount purchase between the two!
Effects on access to loans and grants:
Insurance improved access to grants/loans to cover crop loss
(crowding in of risk-sharing);
Insurance increased perceived ability to finance emergencies,
but not business ventures;
Result is driven by changes in the iddir villages, particularly
changes in the mandated ones;
19
20. Results
Impact on welfare:
Only moderate effects in the short-term
Where there were payouts (Shashemene):
Those in mandated villages more likely to purchase
household durables (clothing, footwear and mobile
phones) in the 4-5 months following payouts than those in
control villages.
Livestock ownership increased in mandated villages
No effect on food consumption;
Where there were no payouts (non-Shashemene sites):
No effect on food consumption or durable purchases;
20
21. Conclusions & implications
Limits to formal and informal insurances to mitigate weather risk
Index-based insurance unable to meet individual specific risks
Iddirs unable to meet risks correlated across households and
villages!
We find evidence that there is high potential to dealing with this
problem by integrating both:
Formal insurance addressing correlated risks via the index;
informal insurance addressing individual specific risks through
strengthening of existing iddir rules; promoting more loans &
transfers.
Iddirs as retail outlets – reduce cost,& promote trust
21
22. Conclusions & implications
We find evidence that a product that integrates both
Increases household welfare (purchase of household durables)
However, for all these to work, institutionalization of new
sharing rules is required!
Policy implications:
Immense potential of ‘traditional groups’ for scaling up of
weather related insurance;
Pool iddirs beyond the village, possibily bring them under one
national – risk pooling - umbrella!
Among others, favorable national legal framework, one that
allows including international re-insurance is needed!
22
23. What next …, 2012, & beyond?
Continued with the same design, but
Add an innovative feature to the index – gap insurance –
A lot of optimism last Meher season (2012) –
1537 policies sold in Shashemene (where payouts
were made in 2011)
Payouts made in Dodota & Shashemen for May 2012
Enthusiasm of our partner (BG) MFI for scalling up as a
business model, also linked to its saving & credit products
23