Who Smoothes What Asset Smoothing_ Consumption Smoothing

Document Sample
Who Smoothes What Asset Smoothing_ Consumption Smoothing Powered By Docstoc
					Index Insurance to Enhance Productivity and
  Incomes for Small-scale Agricultural and
    Pastoral Households in Kenya & Mali

 Chris Barrett                 Michael Carter                 Andrew Mude
                                                           International Livestock
Cornell University       University of California, Davis
                                                             Research Institute



                     Ag Sector Council Seminar Series
                      Washington, D.C.11 May 2011
           The “Same Old Story” about Risk
• Commenting on the Feed the Future research strategy, Thomas
  Lumpkin (DG of CIMYT) noted that small holders are currently only
  getting 30% of the yields potentially available with already existing
  technologies
• Why is this? Risk is a big part of the story
• Risk that is high and correlated across individuals creates a number
  of development problems for small farm agriculture:
– Directly discourages investment in profitable, but costly innovations
– Undercuts the development of agricultural credit markets, forcing
  families to rely on autarchic financial strategies, increasing liquidity
  constraints and further undercutting investment
– Together these two forces undercut productivity, reduce growth and
  make people poorer than they need be given the available
  opportunities.
– Finally, risk and the absence of deep credit markets contribute to the
  inter-generational transmission of poverty, lessening the long-term
  human development impacts of even those incomes and growth that
  are achieved.
         The “Same Old Story” about Risk
• Wreckage of past agricultural credit market interventions
  proves that simply cannot legislate these markets into
  existence
• So is there another approach that might create a happier
  ending to the seemingly inevitable story about risk and
  smallholder agriculture?
• Our work begins with the idea that we can change the
  ending if we can change a key structural condition
  (uninsured, correlated risk) that underlies it
• As we will see, a number of technical & financial
  innovations open the way for the innovation of „index
  insurance contract‟ to transfer this correlated risk out of
  the system
• The goal is not to sell insurance per se, but to solve the
  development problems of low growth and human
  vulnerability that make risk matter
     Crafting a New Ending to this “Same Old Story”
• With generous funding from USAID and other donors,
  the Index Insurance Innovation Initiative (I4) is designing
  & implementing pilots focused on boosting investment
  and growth in:
   – Small scale commercial agriculture (Cotton in Mali, Coffee in
     Guatemala, Cacao in Cote d‟Ivoire)
   – Small scale food agriculture (Grains in Ethiopia; Rice in Ecuador; Maize
     in Tanzania; Livestock in Ethiopia)
• Return to discuss these growth-oriented uses of index
  insurance, but first let‟s elaborate the basic concepts
  using the example of index insurance as a productive
  social safety net
       The Genesis of Index Based Livestock Insurance
• Several years ago, DfID began to launch a cash transfer scheme
  targeted at the indigent „failed‟ pastoralist populations in Northern
  Kenya where as much as 40% of the population may live on less than
  $0.25/day
• Our question to DfID: If you
  would pay $15/indigent family
  per-month, would you pay
  $7/year to keep a vulnerable
  family from becoming indigent?
• While this appears as an aid
  value proposition for DfID, our
  bigger point is that index
  insurance can be a productive,
  value proposition for insured
  pastoralist families
• Andrew and Chris will now
  explain the operation and logic of
  this project in more detail
• Or, see the movie version:
  http://blip.tv/file/3757148
                      Piloting IBLI in Northern Kenya
• But can insurance (of any type) be sustainably offered in rangelands?
• Conventional (individual) insurance unlikely to work, especially among
  pastoralists:
    – Transactions costs
    – Moral hazard/adverse selection
• Index insurance avoids problems that make individual insurance
  unprofitable for small, remote clients:
    – No transactions costs of measuring individual losses
    – Preserves effort incentives (no moral hazard) as no single individual can influence
      index.
    – Adverse selection does not matter as payouts do not depend on the riskiness of
      those who buy the insurance
• Available on near real-time basis: faster response than conventional
  humanitarian relief
• Index insurance can, in principle, be used to create an effective safety
  net to alter poverty dynamics and help address broad-scale shocks
                       Piloting IBLI in Northern Kenya
    • Pilot represents one of the efforts to test the risk-management promise
      of Index-Based Insurance.

•    Why a Livestock Index in Northern Kenya?

      -Pastoral production is key livelihood
      facing a risk profile suitable for targeting              Marsabit
      with an index insurance product.

      - Availability of household data allows
      precise contract design.

      -Availability of potential sales delivery
      infrastructure
                           Designing the Index

• Find a reliable, objectively verifiable signal, that explains most of the
  variation in household‟s seasonal livestock mortality
    – We use functions of NDVI, a remotely sensed proxy for forage availability.


• Model a relationship between the risk to be insured (area-average
  livestock mortality) and the driving signal (NDVI):



DATA                                                             Index
• Livestock                         Response                     • Predicted
  Mortality                         Function                       Livestock
• NDVI                                                             Mortality
                                                         Testing the Index Performance

  • Performance of “Predicted Mortality Index” in predicting area-
    average livestock mortality observed in ALRMP
               – Out-of-sample prediction errors within 10% (especially in bad years)
               – Predicts historical droughts well
                                                                                                                                                                                                   Out of sample

                                  Actual Vs. Predicted Seasonal Mortality Rate - Laisamis Cluster
50%
40%
30%                                                                                                                                                                                                              Predicted
20%                                                                                                                                                                                                              Actual

10%
0%
      1982
             1983
                    1984
                           1985
                                  1986
                                         1987
                                                1988
                                                       1989
                                                              1990
                                                                     1991
                                                                            1992
                                                                                   1993
                                                                                          1994
                                                                                                 1995
                                                                                                        1996
                                                                                                               1997
                                                                                                                      1998
                                                                                                                             1999
                                                                                                                                    2000
                                                                                                                                           2001
                                                                                                                                                  2002
                                                                                                                                                         2003
                                                                                                                                                                2004
                                                                                                                                                                       2005
                                                                                                                                                                              2006
                                                                                                                                                                                     2007
                                                                                                                                                                                            2008
                                                                                                                                                                                                   2009
                                                                                                                                                                                                          2010
                                     Actual Vs. Predicted Seasonal Mortality Rate - Chalbi Cluster
50%
40%
30%                                                                                                                                                                                                              Predicted
20%                                                                                                                                                                                                              Actual

10%
0%
                                  1986




                                                                                                               1997




                                                                                                                                                                                            2008
      1982
             1983
                    1984
                           1985


                                         1987
                                                1988
                                                       1989
                                                              1990
                                                                     1991
                                                                            1992
                                                                                   1993
                                                                                          1994
                                                                                                 1995
                                                                                                        1996


                                                                                                                      1998
                                                                                                                             1999
                                                                                                                                    2000
                                                                                                                                           2001
                                                                                                                                                  2002
                                                                                                                                                         2003
                                                                                                                                                                2004
                                                                                                                                                                       2005
                                                                                                                                                                              2006
                                                                                                                                                                                     2007


                                                                                                                                                                                                   2009
                                                                                                                                                                                                          2010
                      Contract Features

• SPATIAL COVEARGE

  – How wide a geographic area can a single index-cover?
  – What is the spatial precision range of the response function?
  – At what level of resolution is the necessary data available?



                                                                   SABARET
                                                  ILLERET




                  •Two Separate NDVI-
                                                                                DUKANA

                                                                                               EL-HADI


                                                        DARADE                                            FUROLE


                  Livestock Mortality
                                                                                        BALESA
                                                                       NORTH HORR

                                                                                                 HURRI HILLS



                  Response Functions
                                                MOITE                                EL GADE
                                                               GALAS

                                                                                    KALACHA              MAIKONA
                                                             GAS

                                                        LOIYANGALANI
                                                                                                                           TURBI
                                                                      ARAPAL
                                                                   LARACHI   KURUGUM


                                                                          OLTUROT
                                                                    MT. KULAL                                      BUBISA
                                                                                         MAJENGO(MARSABIT)



                  •Five Separate Index
                                                                                       KARGI        JIRIMEQILTA
                                                                                                  HULAHULA
                                                                                                           SAGANTE
                                                                     KURUNGU                      OGUCHODIRIB GOMBO
                                                                                                         KITURUNI
                                                                                                                 SONGA
                                                                                                          KARARE JALDESA


                  Coverage Regions
                                                                                                                                   SHURA
                                                             SOUTH HORR(MARSA)HAFARE                 KAMBOYE
                                                                                            KORR
                                                                             ILLAUT(MARSABIT)
                                                                                                 LOGOLOGOGUDAS/SORIADI
                                                                                      LONYORIPICHAU
                                                                                      NGURUNIT              LAISAMIS

                                                                                              LONTOLIO
                                                                                                                   KOYA
                                                                                                 IRIRMERILLE
                     Contract Features

• TEMPORAL COVEARGE

  – Over what time span should the index cover?
  – Function of the production system being modelled?
  – Administrational and liquidity considerations?
                         Contract Features

• RISK COVERAGE AND PRICING

• Need to select an index strike point to trigger indemnity
    – Trade off: Higher Strike  Lower Risk Coverage  Lower Cost
    – Conditional or Unconditional?
    – Payoff structure: Linear, Segmented, All or Nothing, No claims bonus?




                                        Contract Cluster     Consumer Price
                                         Upper Marsabit           5.5%
                                         Lower Marsabit          3.25%
                              Implementation

• Launched in January 2010 in collaboration with commercial
  partners.

• Two sales periods of varying experience
   – Jan/Feb 2010: Sold ~2000 contracts: Premiums collected ~ $46,000: Value
     of livestock covered ~$1,200,000
   – Jan/Feb 2011: Sold ~750 contracts: Premiums collected ~ $9,500


• Key ongoing considerations/challenges:
   •   Delivery Channel
   •   Extension/Education
   •   Information Dissemination and Trust Building
   •   Regulation
                              Impact Assessment
 Site selection: 16 sites
   Confounding factor: ongoing implementation of cash transfer (HSNP)

    Encouragement design
    •Insurance education game: played among 50% sample in game site
    •Discount coupon of the first 15 TLU: (no subsidy for 40% of sample,
       10%-60% subsidies for the rest)                                                                                                    Legend
                                                          ILLERET
                                                                           SABARET                                                        MarsabitIBLI
                                                                                       DUKANA
                                                                                                                                          HSNP, IBLI Game_HSNP, No
                 IBLI Game      No IBLI                                                                EL-HADI


                                                                                                                  FUROLE
                                                                                                                                               HSNP, IBLI Game
                                                                DARADE                          BALESA

                                Game                                           NORTH HORR

                                                                                                         HURRI HILLS
                                                                                                                                               HSNP, No IBLI Game
                                                        MOITE
                                                                       GALAS
                                                                                             EL GADE                                           No HSNP, IBLI Game
    HSNP         4 sites        4 sites                              GAS
                                                                                            KALACHA              MAIKONA                       No HSNP, No IBLI Game
                                                                LOIYANGALANI
                                                                                                                                  TURBI
                                                                              ARAPAL
                                                                           LARACHI   KURUGUM



    No           5 sites        3 control                                         OLTUROT
                                                                            MT. KULAL
                                                                                             MAJENGO(MARSABIT)
                                                                                           KARGI          JIRIMEQILTA
                                                                                                                         BUBISA




    HSNP                        sites                                      KURUNGU
                                                                                                       HULAHULA
                                                                                                                 SAGANTE
                                                                                                       OGUCHODIRIB GOMBO
                                                                                                        KITURUNI
                                                                                                                SONGA
                                                                                                         KARARE JALDESA
                                                                                                                               SHURA
                                                                     SOUTH HORR(MARSA)HAFARE              KAMBOYE
                                                                                                KORR
                                                                                 ILLAUT(MARSABIT)


 Sample selection: 924 households
                                                                                                     LOGOLOGOGUDAS/SORIADI
                                                                                          LONYORIPICHAU
                                                                                              NGURUNIT              LAISAMIS

                                                                                                      LONTOLIO


    • Sample/site proportional to relative pop. sizes
                                                                                                                           KOYA
                                                                                                         IRIRMERILLE




    •For each site, random sampling stratified by livestock wealth class (L, M, H)
          IBLI and the Escape from Poverty Traps


Strong prior evidence of poverty traps
in the arid and semi-arid lands (ASAL)
of east Africa

Standard humanitarian response to
shocks/destitution: food aid.

But if transfers go only to the poor who
are already in the poverty trap, the
numbers of poor will grow. In the long-
run, today’s poor grow worse off as the
unnecessarily poor join their ranks and
compete for scarce and insufficient
transfers.
                                           Herd wealth dynamics in southern Ethiopia
                                           Source: Lybbert et al. Econ. J. 2004
       IBLI and the Escape from Poverty Traps
In theory, sustainable livestock insurance for pastoralists can:
      • Prevent downward slide of vulnerable populations
         - Enables concentrating humanitarian resources on
         those truly unable to lift themselves from poverty

      • Stabilize expectations and crowd-in investment and
        accumulation by poor populations
What we hope to learn via careful impact evaluation of IBLI

1) For whom is IBLI most attractive and effective?
- simulation-based answer: IBLI most valuable among the
vulnerable non-poor
- simulation-based and WTP survey based answer: Highly price
elastic demand for IBLI

Research objective 1: Use survey data to test these hypotheses
in quasi-experimental setting with real insurance.

Policy question: Potential for targeted subsidies of IBLI as a
productive safety net?
What we hope to learn via careful impact evaluation of IBLI

2) Does IBLI induce increased asset accumulation and escapes
from poverty? Does it reduce asset loss and falls into poverty?
How does it perform relative to cash transfers? Are there spillover
effects on the stockless poor?
-simulation-based answers: Yes on first two points. Don’t know
on latter two questions.

Research objective 1: Use survey data to test these hypotheses
in quasi-experimental setting with real insurance in a survey
designed to test IBLI and cash transfers under Kenya’s new
Hunger Safety Nets Program.

Policy question: Which instrument performs best in terms of
poverty reduction and economic growth? Are IBLI and HSNP
complements or substitutes?
          IBLI in the Face of Climate Change
Arid and semi-arid lands (ASAL) comprise ~ 2/3 of Africa,
home to ~20 mn pastoralists - extensive livestock grazing.

Pastoralist systems adapted to climate regime, but vulnerable
to drought. Rapid shift in climate could bring catastrophe.

An implication of most climate change predictions is increased
rainfall variability, so increased risk of drought.
                                               IBLI in the Face of Climate Change
Herd dynamics differ markedly between good and poor rainfall
states. As a result, herd dynamics change with drought (rainfall
<250 mm/year) risk. Halving the current risk would enhance
resilience and eliminate apparent poverty trap. By contrast,
doubling drought risk would lead to system collapse in expectation.
                                      60



                                      50             Prob. = 0.03
  Expected herd size 10 years ahead




                                                            Prob. = 0.06
                                      40
                                                                                                   Implication:
                                      30                                                           Need to alter herd
                                                                          Prob. = 0.12             dynamics to cope with
                                      20                                                           increasing drought risk.
                                                                                                   IBLI is one possible tool
                                      10
                                                                                                   to avert collapse.
                                      0
                                           0    10   20        30            40          50   60
                                                      Initial herd size
                                                                                                   Source: Barrett and Santos, 2011
      From Protection to Growth: Gueleya Nyesigi in Mali

• IBLI illustrates the development
  benefits from stabilizing livelihoods
• The the I4/PlaNet Guarantee project
  in Mali illustrates the growth potential
  of index insurance
• Key insight: Interlinking insurance
  with growth opportunity
   • Avoids the tradeoff of reduced
       variability at the cost of reduced
       average income                        • In Mali, smallholders leave
   • Instead can reduce variability            significant „money on the
       while increasing mean income            table‟ every year by planting
                                               only 1 hectare in cotton
                                             • An interlinked contract that
                                               crowds in supply & demand
                                               for credit for that 2nd hectare
                                               of cotton can create growth
                                             • Requires a high quality
                                               contract that is understood
          From Protection to Growth: Contract Design

• Credit supply in
  Mali is currently
  generously, if
  unsustainably,
  supplied through a
  parastatal
  infrastructure
• Yet small farmers
  remain reluctant to
  borrow under
  group credit
  scheme
• Working with
  farmers, I4
  researchers
  designed a novel
  contract approach     • Identified farmer “loss aversion focal point” of
                          750kg/hectare
                        • Devised a double trigger contract that radically
                          reduces basis risk while protecting against
                          moral hazard
         From Protection to Growth: Farmer Education
• Even a good contract will
  have no impact on
  growth if neither trusted
  nor understood
• Building on IBLI, the Mali
  team is employing
  games, a network of
  VIPs, and a variety of
  user-friendly educational
  material
• Outreach implemented
  by Oxfam and its local
  partners

                               • Contract on sale now
                               • Using a spatially randomized rollout, will be
                                 monitoring whether insurance crowds in
                                 greater entrepreneurial risk taking
                               • Next challenge will be when credit supply
                                 shifts following privatization
                         In Summary …

• As researchers, we are excited by the potential of index insurance to
  provide a new ending to an old story about risk
• While built on a number of technical and contractual innovations,
  these two & the other I4 projects are all implemented by commercial
  partners
• This modality opens the door to scale up of these ideas
• I4 has upcoming, industry- & government-oriented outreach events
  in East Africa and the Andean region
• While we are keen to share what we have learned, we are also
  acutely aware of the need to monitor the impacts of our work on
  human development and agricultural growth
• Stay tuned!

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:69
posted:7/22/2011
language:English
pages:24